From 526977068f919da0386e2c5463645643be142d1d Mon Sep 17 00:00:00 2001 From: MagicExists <106458387+gugugiyu@users.noreply.github.com> Date: Thu, 4 Jun 2026 21:09:43 +0700 Subject: [PATCH 01/13] ui: added single line reasoning preview (#23601) * webui: added single line reasoning preview. * patch: reduce width slightly for the previewing section * refactor: move formatter constants to the right file * feat: reimplement reasoning preview with throttled dynamic per-line rendering * chore: fix spacing Co-authored-by: Aleksander Grygier * chore: refactor to requested changes * refactor: grouped by capture pattern instead of block-level + inline * ui: fax interrupt state only trigger for 1st reasoning message * chore: make reasoning preview respects showThoughtInProgress setting * chore; newline at EOF Co-authored-by: Aleksander Grygier * fix: thread rawContent so collapsible content can handle compute preview * patch: showThoughtInProgress accidentally blocks rawContent being passed * chore: fix lint * chore: change smoke test --------- Co-authored-by: Aleksander Grygier --- .../ChatMessageAgenticContent.svelte | 20 ++++++- .../content/CollapsibleContentBlock.svelte | 52 ++++++++++++++++--- tools/ui/src/lib/constants/formatters.ts | 27 ++++++++++ tools/ui/src/lib/hooks/use-throttle.svelte.ts | 32 ++++++++++++ tools/ui/src/lib/utils/agentic.ts | 4 +- tools/ui/src/lib/utils/formatters.ts | 36 ++++++++++++- tools/ui/src/lib/utils/index.ts | 3 +- 7 files changed, 161 insertions(+), 13 deletions(-) create mode 100644 tools/ui/src/lib/hooks/use-throttle.svelte.ts diff --git a/tools/ui/src/lib/components/app/chat/ChatMessages/ChatMessageAgenticContent.svelte b/tools/ui/src/lib/components/app/chat/ChatMessages/ChatMessageAgenticContent.svelte index 3a9cc7e9356d..e21dff993ffd 100644 --- a/tools/ui/src/lib/components/app/chat/ChatMessages/ChatMessageAgenticContent.svelte +++ b/tools/ui/src/lib/components/app/chat/ChatMessages/ChatMessageAgenticContent.svelte @@ -31,7 +31,8 @@ agenticPendingPermissionRequest, agenticResolvePermission, agenticPendingContinueRequest, - agenticResolveContinue + agenticResolveContinue, + agenticLastError } from '$lib/stores/agentic.svelte'; import { config } from '$lib/stores/settings.svelte'; @@ -56,6 +57,10 @@ const showToolCallInProgress = $derived(config().showToolCallInProgress as boolean); const showThoughtInProgress = $derived(config().showThoughtInProgress as boolean); + const hasReasoningError = $derived( + isLastAssistantMessage ? !!agenticLastError(message.convId) : false + ); + let permissionDismissed = $state(false); const pendingPermission = $derived( @@ -293,11 +298,21 @@ {:else if section.type === AgenticSectionType.REASONING} + {@const reasoningSubtitle = section.wasInterrupted + ? hasReasoningError + ? 'Error' + : 'Cancelled' + : isStreaming + ? '' + : undefined} + toggleExpanded(index, section)} >
@@ -308,7 +323,7 @@ {:else if section.type === AgenticSectionType.REASONING_PENDING} {@const reasoningTitle = isStreaming ? 'Reasoning...' : 'Reasoning'} - {@const reasoningSubtitle = isStreaming ? '' : 'incomplete'} + {@const reasoningSubtitle = isStreaming ? '' : hasReasoningError ? 'Error' : 'Cancelled'} toggleExpanded(index, section)} > diff --git a/tools/ui/src/lib/components/app/content/CollapsibleContentBlock.svelte b/tools/ui/src/lib/components/app/content/CollapsibleContentBlock.svelte index b7297ab6b1aa..8bab55d19fb9 100644 --- a/tools/ui/src/lib/components/app/content/CollapsibleContentBlock.svelte +++ b/tools/ui/src/lib/components/app/content/CollapsibleContentBlock.svelte @@ -4,6 +4,9 @@ import { buttonVariants } from '$lib/components/ui/button/index.js'; import { Card } from '$lib/components/ui/card'; import { createAutoScrollController } from '$lib/hooks/use-auto-scroll.svelte'; + import { useThrottle } from '$lib/hooks/use-throttle.svelte'; + import { formatReasoningPreview } from '$lib/utils'; + import { config } from '$lib/stores/settings.svelte'; import type { Snippet } from 'svelte'; import type { Component } from 'svelte'; @@ -14,6 +17,8 @@ iconClass?: string; title: string; subtitle?: string; + preview?: string; + rawContent?: string; isStreaming?: boolean; onToggle?: () => void; children: Snippet; @@ -26,6 +31,8 @@ iconClass = 'h-4 w-4', title, subtitle, + preview, + rawContent, isStreaming = false, onToggle, children @@ -33,6 +40,20 @@ let contentContainer: HTMLDivElement | undefined = $state(); + const showThoughtInProgress = $derived(config().showThoughtInProgress as boolean); + + let previewKey = useThrottle(() => rawContent ?? preview ?? '', 500); + let displayedPreview = $state(''); + let displayedOverflow = $state(0); + + $effect(() => { + void previewKey.key; + const content = rawContent ?? preview ?? ''; + const result = formatReasoningPreview(content); + displayedPreview = result.preview; + displayedOverflow = result.overflow; + }); + const autoScroll = createAutoScrollController(); $effect(() => { @@ -58,16 +79,31 @@ class={className} > - -
- {#if IconComponent} - - {/if} + +
+
+ {#if IconComponent} + + {/if} + + {title} - {title} + {#if subtitle} + {subtitle} + {/if} +
- {#if subtitle} - {subtitle} + {#if displayedPreview && !showThoughtInProgress} +
+
+ {displayedPreview} +
+ {#if displayedOverflow > 0} + {displayedOverflow}+ chars + {/if} +
{/if}
diff --git a/tools/ui/src/lib/constants/formatters.ts b/tools/ui/src/lib/constants/formatters.ts index d6d1b883ffe9..c417faea43d9 100644 --- a/tools/ui/src/lib/constants/formatters.ts +++ b/tools/ui/src/lib/constants/formatters.ts @@ -6,3 +6,30 @@ export const MEDIUM_DURATION_THRESHOLD = 10; /** Default display value when no performance time is available */ export const DEFAULT_PERFORMANCE_TIME = '0s'; + +/** Max length before reasoning preview is truncated */ +export const MAX_PREVIEW_LENGTH = 120; + +export const STRIP_MARKDOWN_CAPTURE_PATTERNS: [RegExp, string][] = [ + [/^```(.*)/gm, '$1'], + [/(.*)```$/gm, '$1'], + [/`([^`]*)`/g, '$1'], + [/\*\*(.*?)\*\*/g, '$1'], + [/__(.*?)__/g, '$1'], + [/\*(.*?)\*/g, '$1'], + [/_(.*?)_/g, '$1'] +]; + +/* eslint-disable no-misleading-character-class */ +export const STRIP_MARKDOWN_INLINE_REGEX = new RegExp( + [ + '<[^>]*>', + '^>\\s*', + '^#{1,6}\\s+', + '^[\\s]*[-*+]\\s+', + '^[\\s]*\\d+[.)]\\s+', + '[\\u{1F600}-\\u{1F64F}\\u{1F300}-\\u{1F5FF}\\u{1F680}-\\u{1F6FF}\\u{1F1E0}-\\u{1F1FF}\\u{2600}-\\u{26FF}\\u{2700}-\\u{27BF}\\u{FE00}-\\u{FE0F}\\u{1F900}-\\u{1F9FF}\\u{1FA00}-\\u{1FA6F}\\u{1FA70}-\\u{1FAFF}\\u{200D}\\u{20E3}\\u{231A}-\\u{231B}\\u{23E9}-\\u{23F3}\\u{23F8}-\\u{23FA}\\u{25AA}-\\u{25AB}\\u{25B6}\\u{25C0}\\u{25FB}-\\u{25FE}\\u{2934}-\\u{2935}\\u{2B05}-\\u{2B07}\\u{2B1B}-\\u{2B1C}\\u{2B50}\\u{2B55}\\u{3030}\\u{303D}\\u{3297}\\u{3299}]' + ].join('|'), + 'gmu' +); +/* eslint-enable no-misleading-character-class */ diff --git a/tools/ui/src/lib/hooks/use-throttle.svelte.ts b/tools/ui/src/lib/hooks/use-throttle.svelte.ts new file mode 100644 index 000000000000..0795519787b8 --- /dev/null +++ b/tools/ui/src/lib/hooks/use-throttle.svelte.ts @@ -0,0 +1,32 @@ +/** + * Creates a reactive throttle key that increments when `getValue()` changes + * and the throttle window has elapsed since the last increment. + * + * Useful for throttling animations that should not fire on every rapid update. + * + * @param getValue - A reactive getter for the value to watch + * @param ms - Throttle window in milliseconds + * @returns A reactive number that increments when the throttled value changes + */ +export function useThrottle(getValue: () => string | undefined, ms: number) { + let key = $state(0); + let throttleEnd = $state(0); + let lastValue: string | undefined = getValue(); + + $effect(() => { + const value = getValue(); + if (value === lastValue) return; + const now = Date.now(); + if (now >= throttleEnd) { + lastValue = value; + key++; + throttleEnd = now + ms; + } + }); + + return { + get key() { + return key; + } + }; +} diff --git a/tools/ui/src/lib/utils/agentic.ts b/tools/ui/src/lib/utils/agentic.ts index 52ff35793063..d19f03434e6b 100644 --- a/tools/ui/src/lib/utils/agentic.ts +++ b/tools/ui/src/lib/utils/agentic.ts @@ -18,6 +18,7 @@ export interface AgenticSection { toolArgs?: string; toolResult?: string; toolResultExtras?: DatabaseMessageExtra[]; + wasInterrupted?: boolean; } /** @@ -51,7 +52,8 @@ function deriveSingleTurnSections( const isPending = isStreaming && !hasContentAfterReasoning; sections.push({ type: isPending ? AgenticSectionType.REASONING_PENDING : AgenticSectionType.REASONING, - content: message.reasoningContent + content: message.reasoningContent, + wasInterrupted: !isStreaming && !hasContentAfterReasoning }); } diff --git a/tools/ui/src/lib/utils/formatters.ts b/tools/ui/src/lib/utils/formatters.ts index 24a2c1c94c18..de74ee8686d1 100644 --- a/tools/ui/src/lib/utils/formatters.ts +++ b/tools/ui/src/lib/utils/formatters.ts @@ -3,7 +3,11 @@ import { SECONDS_PER_MINUTE, SECONDS_PER_HOUR, SHORT_DURATION_THRESHOLD, - MEDIUM_DURATION_THRESHOLD + MEDIUM_DURATION_THRESHOLD, + MAX_PREVIEW_LENGTH, + STRIP_MARKDOWN_INLINE_REGEX, + STRIP_MARKDOWN_CAPTURE_PATTERNS, + NEWLINE_SEPARATOR } from '$lib/constants'; /** @@ -151,3 +155,33 @@ export function formatAttachmentText( const header = extra ? `${name} (${extra})` : name; return `\n\n--- ${label}: ${header} ---\n${content}`; } + +export function formatReasoningPreview(content: string): { preview: string; overflow: number } { + if (!content) return { preview: '', overflow: 0 }; + + const lines = content.split(NEWLINE_SEPARATOR); + let lastLine = ''; + + for (let i = lines.length - 1; i >= 0; i--) { + let cleaned = lines[i].trim(); + if (!cleaned) continue; + + cleaned = cleaned.replace(STRIP_MARKDOWN_INLINE_REGEX, ''); + for (const [pattern, replacement] of STRIP_MARKDOWN_CAPTURE_PATTERNS) { + cleaned = cleaned.replace(pattern, replacement); + } + + if (cleaned.length > 0) { + lastLine = cleaned; + break; + } + } + + const fullLength = lastLine.length; + const overflow = Math.max(0, fullLength - MAX_PREVIEW_LENGTH); + if (fullLength > MAX_PREVIEW_LENGTH) { + lastLine = lastLine.slice(0, MAX_PREVIEW_LENGTH) + '...'; + } + + return { preview: lastLine, overflow }; +} diff --git a/tools/ui/src/lib/utils/index.ts b/tools/ui/src/lib/utils/index.ts index 00aa49c41765..637db8812c4d 100644 --- a/tools/ui/src/lib/utils/index.ts +++ b/tools/ui/src/lib/utils/index.ts @@ -76,7 +76,8 @@ export { formatJsonPretty, formatTime, formatPerformanceTime, - formatAttachmentText + formatAttachmentText, + formatReasoningPreview } from './formatters'; // IME utilities From 21444c822e3fcf5a5fed9a8930a5caf60c291a5e Mon Sep 17 00:00:00 2001 From: Aleksander Grygier Date: Thu, 4 Jun 2026 16:23:08 +0200 Subject: [PATCH 02/13] ui: Fixed packages (#24119) * chore(ui): pin package versions to currently installed - Update all dependencies and devDependencies to match exactly what's in package-lock.json - This ensures reproducible builds by locking to specific versions rather than semver ranges * chore: Update packages * chore: Move remaining dependencies to devDependencies * fix: Add missing `mermaid` package * chore: Update `cookie` package to `v1.1.1` * chore: Formatting * test: Update test configs --- tools/ui/package-lock.json | 1723 +++++++++-------- tools/ui/package.json | 138 +- .../stories/SidebarNavigation.stories.svelte | 22 +- tools/ui/vite.config.ts | 27 +- 4 files changed, 1063 insertions(+), 847 deletions(-) diff --git a/tools/ui/package-lock.json b/tools/ui/package-lock.json index 06f885680a6b..ffd4f6ca029a 100644 --- a/tools/ui/package-lock.json +++ b/tools/ui/package-lock.json @@ -7,77 +7,76 @@ "": { "name": "llama-ui", "version": "1.0.0", - "dependencies": { - "@modelcontextprotocol/sdk": "^1.25.1", - "highlight.js": "^11.11.1", - "mermaid": "^11.15.0", - "mode-watcher": "^1.1.0", - "pdfjs-dist": "^5.4.54", - "rehype-highlight": "^7.0.2", - "rehype-stringify": "^10.0.1", - "remark": "^15.0.1", - "remark-breaks": "^4.0.0", - "remark-gfm": "^4.0.1", - "remark-html": "^16.0.1", - "remark-rehype": "^11.1.2", - "svelte-sonner": "^1.0.5", - "unist-util-visit": "^5.0.0", - "zod": "^4.2.1" - }, "devDependencies": { - "@chromatic-com/storybook": "^5.0.0", - "@eslint/compat": "^1.2.5", - "@eslint/js": "^9.18.0", - "@internationalized/date": "^3.10.1", - "@lucide/svelte": "^0.515.0", - "@playwright/test": "^1.49.1", - "@storybook/addon-a11y": "^10.2.4", - "@storybook/addon-docs": "^10.2.4", - "@storybook/addon-svelte-csf": "^5.0.10", - "@storybook/addon-vitest": "^10.2.4", - "@storybook/sveltekit": "^10.2.4", - "@sveltejs/adapter-static": "^3.0.10", - "@sveltejs/kit": "^2.48.4", - "@sveltejs/vite-plugin-svelte": "^6.2.1", - "@tailwindcss/forms": "^0.5.9", - "@tailwindcss/typography": "^0.5.15", - "@tailwindcss/vite": "^4.0.0", + "@chromatic-com/storybook": "5.0.0", + "@eslint/compat": "1.4.1", + "@eslint/js": "9.39.2", + "@internationalized/date": "3.10.1", + "@lucide/svelte": "0.515.0", + "@modelcontextprotocol/sdk": "1.26.0", + "@playwright/test": "1.56.1", + "@storybook/addon-a11y": "10.2.4", + "@storybook/addon-docs": "10.2.4", + "@storybook/addon-svelte-csf": "5.0.10", + "@storybook/addon-vitest": "10.2.4", + "@storybook/sveltekit": "10.2.4", + "@sveltejs/adapter-static": "3.0.10", + "@sveltejs/kit": "2.60.1", + "@sveltejs/vite-plugin-svelte": "6.2.1", + "@tailwindcss/forms": "0.5.10", + "@tailwindcss/typography": "0.5.16", + "@tailwindcss/vite": "4.1.11", "@types/node": "^24", - "@vitest/browser": "^3.2.3", - "@vitest/coverage-v8": "^3.2.3", - "bits-ui": "^2.14.4", - "clsx": "^2.1.1", - "dexie": "^4.0.11", - "eslint": "^9.18.0", - "eslint-config-prettier": "^10.0.1", - "eslint-plugin-storybook": "^10.2.4", - "eslint-plugin-svelte": "^3.0.0", - "globals": "^16.0.0", - "http-server": "^14.1.1", - "mdast": "^3.0.0", - "mdsvex": "^0.12.3", - "playwright": "^1.56.1", - "prettier": "^3.4.2", - "prettier-plugin-svelte": "^3.3.3", - "prettier-plugin-tailwindcss": "^0.6.11", - "rehype-katex": "^7.0.1", - "remark-math": "^6.0.0", - "sass": "^1.93.3", - "storybook": "^10.2.4", - "svelte": "^5.38.2", - "svelte-check": "^4.0.0", - "tailwind-merge": "^3.3.1", - "tailwind-variants": "^3.2.2", - "tailwindcss": "^4.0.0", - "tw-animate-css": "^1.3.5", - "typescript": "^5.0.0", - "typescript-eslint": "^8.20.0", - "unified": "^11.0.5", - "uuid": "^13.0.0", - "vite": "^7.2.2", - "vite-plugin-devtools-json": "^0.2.0", - "vitest": "^3.2.3", - "vitest-browser-svelte": "^0.1.0" + "@vitest/browser": "4.1.8", + "@vitest/browser-playwright": "4.1.8", + "@vitest/coverage-v8": "4.1.8", + "bits-ui": "2.18.1", + "clsx": "2.1.1", + "dexie": "4.0.11", + "eslint": "9.39.2", + "eslint-config-prettier": "10.1.8", + "eslint-plugin-storybook": "10.2.4", + "eslint-plugin-svelte": "3.15.0", + "globals": "16.3.0", + "highlight.js": "11.11.1", + "http-server": "14.1.1", + "mdast": "3.0.0", + "mdsvex": "0.12.6", + "mermaid": "11.15.0", + "mode-watcher": "1.1.0", + "pdfjs-dist": "5.4.54", + "playwright": "1.56.1", + "prettier": "3.6.2", + "prettier-plugin-svelte": "3.4.0", + "prettier-plugin-tailwindcss": "0.6.14", + "rehype-highlight": "7.0.2", + "rehype-katex": "7.0.1", + "rehype-stringify": "10.0.1", + "remark": "15.0.1", + "remark-breaks": "4.0.0", + "remark-gfm": "4.0.1", + "remark-html": "16.0.1", + "remark-math": "6.0.0", + "remark-rehype": "11.1.2", + "sass": "1.93.3", + "storybook": "10.3.3", + "svelte": "5.55.7", + "svelte-check": "4.3.0", + "svelte-sonner": "1.0.5", + "tailwind-merge": "3.3.1", + "tailwind-variants": "3.2.2", + "tailwindcss": "4.1.11", + "tw-animate-css": "1.3.5", + "typescript": "5.8.3", + "typescript-eslint": "8.56.0", + "unified": "11.0.5", + "unist-util-visit": "5.0.0", + "uuid": "13.0.2", + "vite": "7.3.2", + "vite-plugin-devtools-json": "0.2.1", + "vitest": "4.1.8", + "vitest-browser-svelte": "2.1.1", + "zod": "4.2.1" } }, "node_modules/@adobe/css-tools": { @@ -105,6 +104,7 @@ "version": "1.1.0", "resolved": "https://registry.npmjs.org/@antfu/install-pkg/-/install-pkg-1.1.0.tgz", "integrity": "sha512-MGQsmw10ZyI+EJo45CdSER4zEb+p31LpDAFp2Z3gkSd1yqVZGi0Ebx++YTEMonJy4oChEMLsxZ64j8FH6sSqtQ==", + "dev": true, "license": "MIT", "dependencies": { "package-manager-detector": "^1.3.0", @@ -114,23 +114,15 @@ "url": "https://github.com/sponsors/antfu" } }, - "node_modules/@antfu/install-pkg/node_modules/tinyexec": { - "version": "1.1.2", - "resolved": "https://registry.npmjs.org/tinyexec/-/tinyexec-1.1.2.tgz", - "integrity": "sha512-dAqSqE/RabpBKI8+h26GfLq6Vb3JVXs30XYQjdMjaj/c2tS8IYYMbIzP599KtRj7c57/wYApb3QjgRgXmrCukA==", - "license": "MIT", - "engines": { - "node": ">=18" - } - }, "node_modules/@babel/code-frame": { - "version": "7.27.1", - "resolved": "https://registry.npmjs.org/@babel/code-frame/-/code-frame-7.27.1.tgz", - "integrity": "sha512-cjQ7ZlQ0Mv3b47hABuTevyTuYN4i+loJKGeV9flcCgIK37cCXRh+L1bd3iBHlynerhQ7BhCkn2BPbQUL+rGqFg==", + "version": "7.29.7", + "resolved": "https://registry.npmjs.org/@babel/code-frame/-/code-frame-7.29.7.tgz", + "integrity": "sha512-Aup7aUOfpbAUg2ROOJN6Iw5f9DMBlzu0mIkm/malLQFN/YQgO48wCj0Kxa3sEHJvPVFg7siR+qRInwXd2qhQKw==", "dev": true, "license": "MIT", + "peer": true, "dependencies": { - "@babel/helper-validator-identifier": "^7.27.1", + "@babel/helper-validator-identifier": "^7.29.7", "js-tokens": "^4.0.0", "picocolors": "^1.1.1" }, @@ -138,10 +130,18 @@ "node": ">=6.9.0" } }, + "node_modules/@babel/code-frame/node_modules/js-tokens": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-4.0.0.tgz", + "integrity": "sha512-RdJUflcE3cUzKiMqQgsCu06FPu9UdIJO0beYbPhHN4k6apgJtifcoCtT9bcxOpYBtpD2kCM6Sbzg4CausW/PKQ==", + "dev": true, + "license": "MIT", + "peer": true + }, "node_modules/@babel/helper-string-parser": { - "version": "7.27.1", - "resolved": "https://registry.npmjs.org/@babel/helper-string-parser/-/helper-string-parser-7.27.1.tgz", - "integrity": "sha512-qMlSxKbpRlAridDExk92nSobyDdpPijUq2DW6oDnUqd0iOGxmQjyqhMIihI9+zv4LPyZdRje2cavWPbCbWm3eA==", + "version": "7.29.7", + "resolved": "https://registry.npmjs.org/@babel/helper-string-parser/-/helper-string-parser-7.29.7.tgz", + "integrity": "sha512-Pb5ijPrZ89GDH8223L4UP8i6QApWxs04RbPQJTeWDV0/keR2E36MeKnyr6LYmUUvqRRI+Iv87SuF1W6ErINzYw==", "dev": true, "license": "MIT", "engines": { @@ -149,9 +149,9 @@ } }, "node_modules/@babel/helper-validator-identifier": { - "version": "7.28.5", - "resolved": "https://registry.npmjs.org/@babel/helper-validator-identifier/-/helper-validator-identifier-7.28.5.tgz", - "integrity": "sha512-qSs4ifwzKJSV39ucNjsvc6WVHs6b7S03sOh2OcHF9UHfVPqWWALUsNUVzhSBiItjRZoLHx7nIarVjqKVusUZ1Q==", + "version": "7.29.7", + "resolved": "https://registry.npmjs.org/@babel/helper-validator-identifier/-/helper-validator-identifier-7.29.7.tgz", + "integrity": "sha512-qehxGkRj55h/ff8EMaJ+cYhyaKlHIxqYDn682wQD7RNp9UujOQsHog2uS0r2vzr4pW+sXf90NeeayjcNaX3fFg==", "dev": true, "license": "MIT", "engines": { @@ -159,13 +159,13 @@ } }, "node_modules/@babel/parser": { - "version": "7.29.0", - "resolved": "https://registry.npmjs.org/@babel/parser/-/parser-7.29.0.tgz", - "integrity": "sha512-IyDgFV5GeDUVX4YdF/3CPULtVGSXXMLh1xVIgdCgxApktqnQV0r7/8Nqthg+8YLGaAtdyIlo2qIdZrbCv4+7ww==", + "version": "7.29.7", + "resolved": "https://registry.npmjs.org/@babel/parser/-/parser-7.29.7.tgz", + "integrity": "sha512-hnORnjP/1P/zFEndoeX+n+t1RwWRJiJpM/jO7FW32Kn9r5+sJB2JWOdYo4L6k78j15eCwY3Gm/7364B1EMwtNg==", "dev": true, "license": "MIT", "dependencies": { - "@babel/types": "^7.29.0" + "@babel/types": "^7.29.7" }, "bin": { "parser": "bin/babel-parser.js" @@ -175,24 +175,25 @@ } }, "node_modules/@babel/runtime": { - "version": "7.27.6", - "resolved": "https://registry.npmjs.org/@babel/runtime/-/runtime-7.27.6.tgz", - "integrity": "sha512-vbavdySgbTTrmFE+EsiqUTzlOr5bzlnJtUv9PynGCAKvfQqjIXbvFdumPM/GxMDfyuGMJaJAU6TO4zc1Jf1i8Q==", + "version": "7.29.7", + "resolved": "https://registry.npmjs.org/@babel/runtime/-/runtime-7.29.7.tgz", + "integrity": "sha512-Nq8OhGWiZIZGV6hLHoyAKLLcJihP/xFeBMGJoUrxTX2psI8dCifzLhZISFb+VWS3wFMRDmCGw5R+dOySCqPLhw==", "dev": true, "license": "MIT", + "peer": true, "engines": { "node": ">=6.9.0" } }, "node_modules/@babel/types": { - "version": "7.29.0", - "resolved": "https://registry.npmjs.org/@babel/types/-/types-7.29.0.tgz", - "integrity": "sha512-LwdZHpScM4Qz8Xw2iKSzS+cfglZzJGvofQICy7W7v4caru4EaAmyUuO6BGrbyQ2mYV11W0U8j5mBhd14dd3B0A==", + "version": "7.29.7", + "resolved": "https://registry.npmjs.org/@babel/types/-/types-7.29.7.tgz", + "integrity": "sha512-4zBIxpPzowiZpusoFkyGVwakdRJUyuH5PxQ/PrqghfdFWWasvnCdPfQXHrenDai+gyLARulZjZowCOj6fjT4pA==", "dev": true, "license": "MIT", "dependencies": { - "@babel/helper-string-parser": "^7.27.1", - "@babel/helper-validator-identifier": "^7.28.5" + "@babel/helper-string-parser": "^7.29.7", + "@babel/helper-validator-identifier": "^7.29.7" }, "engines": { "node": ">=6.9.0" @@ -208,16 +209,25 @@ "node": ">=18" } }, + "node_modules/@blazediff/core": { + "version": "1.9.1", + "resolved": "https://registry.npmjs.org/@blazediff/core/-/core-1.9.1.tgz", + "integrity": "sha512-ehg3jIkYKulZh+8om/O25vkvSsXXwC+skXmyA87FFx6A/45eqOkZsBltMw/TVteb0mloiGT8oGRTcjRAz66zaA==", + "dev": true, + "license": "MIT" + }, "node_modules/@braintree/sanitize-url": { "version": "7.1.2", "resolved": "https://registry.npmjs.org/@braintree/sanitize-url/-/sanitize-url-7.1.2.tgz", "integrity": "sha512-jigsZK+sMF/cuiB7sERuo9V7N9jx+dhmHHnQyDSVdpZwVutaBu7WvNYqMDLSgFgfB30n452TP3vjDAvFC973mA==", + "dev": true, "license": "MIT" }, "node_modules/@chevrotain/types": { "version": "11.1.2", "resolved": "https://registry.npmjs.org/@chevrotain/types/-/types-11.1.2.tgz", "integrity": "sha512-U+HFai5+zmJCkK86QsaJtoITlboZHBqrVketcO2ROv865xfCMSFpELQoz1GkX5GzME8pTa+3kbKrZHQtI0gdbw==", + "dev": true, "license": "Apache-2.0" }, "node_modules/@chromatic-com/storybook": { @@ -893,6 +903,7 @@ "version": "1.19.13", "resolved": "https://registry.npmjs.org/@hono/node-server/-/node-server-1.19.13.tgz", "integrity": "sha512-TsQLe4i2gvoTtrHje625ngThGBySOgSK3Xo2XRYOdqGN1teR8+I7vchQC46uLJi8OF62YTYA3AhSpumtkhsaKQ==", + "dev": true, "license": "MIT", "engines": { "node": ">=18.14.1" @@ -971,12 +982,14 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/@iconify/types/-/types-2.0.0.tgz", "integrity": "sha512-+wluvCrRhXrhyOmRDJ3q8mux9JkKy5SJ/v8ol2tu4FVjyYvtEzkc/3pK15ET6RKg4b4w4BmTk1+gsCUhf21Ykg==", + "dev": true, "license": "MIT" }, "node_modules/@iconify/utils": { "version": "3.1.3", "resolved": "https://registry.npmjs.org/@iconify/utils/-/utils-3.1.3.tgz", "integrity": "sha512-LPKOXPn/zV+zis1oOfGWogaXVpqUybF3ZS6SCZIsz8vg0ivVp9+fVqyYB7xq0aiST/VhUQYGO1qo6uoYSiEJqw==", + "dev": true, "license": "MIT", "dependencies": { "@antfu/install-pkg": "^1.1.0", @@ -994,24 +1007,6 @@ "@swc/helpers": "^0.5.0" } }, - "node_modules/@isaacs/cliui": { - "version": "8.0.2", - "resolved": "https://registry.npmjs.org/@isaacs/cliui/-/cliui-8.0.2.tgz", - "integrity": "sha512-O8jcjabXaleOG9DQ0+ARXWZBTfnP4WNAqzuiJK7ll44AmxGKv/J2M4TPjxjY3znBCfvBXFzucm1twdyFybFqEA==", - "dev": true, - "license": "ISC", - "dependencies": { - "string-width": "^5.1.2", - "string-width-cjs": "npm:string-width@^4.2.0", - "strip-ansi": "^7.0.1", - "strip-ansi-cjs": "npm:strip-ansi@^6.0.1", - "wrap-ansi": "^8.1.0", - "wrap-ansi-cjs": "npm:wrap-ansi@^7.0.0" - }, - "engines": { - "node": ">=12" - } - }, "node_modules/@isaacs/fs-minipass": { "version": "4.0.1", "resolved": "https://registry.npmjs.org/@isaacs/fs-minipass/-/fs-minipass-4.0.1.tgz", @@ -1025,20 +1020,11 @@ "node": ">=18.0.0" } }, - "node_modules/@istanbuljs/schema": { - "version": "0.1.3", - "resolved": "https://registry.npmjs.org/@istanbuljs/schema/-/schema-0.1.3.tgz", - "integrity": "sha512-ZXRY4jNvVgSVQ8DL3LTcakaAtXwTVUxE81hslsyD2AtoXW/wVob10HkOJ1X/pAlcI7D+2YoZKg5do8G/w6RYgA==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=8" - } - }, "node_modules/@jridgewell/gen-mapping": { "version": "0.3.12", "resolved": "https://registry.npmjs.org/@jridgewell/gen-mapping/-/gen-mapping-0.3.12.tgz", "integrity": "sha512-OuLGC46TjB5BbN1dH8JULVVZY4WTdkF7tV9Ys6wLL1rubZnCMstOhNHueU5bLCrnRuDhKPDM4g6sw4Bel5Gzqg==", + "dev": true, "license": "MIT", "dependencies": { "@jridgewell/sourcemap-codec": "^1.5.0", @@ -1049,6 +1035,7 @@ "version": "2.3.5", "resolved": "https://registry.npmjs.org/@jridgewell/remapping/-/remapping-2.3.5.tgz", "integrity": "sha512-LI9u/+laYG4Ds1TDKSJW2YPrIlcVYOwi2fUC6xB43lueCjgxV4lffOCZCtYFiH6TNOX+tQKXx97T4IKHbhyHEQ==", + "dev": true, "license": "MIT", "dependencies": { "@jridgewell/gen-mapping": "^0.3.5", @@ -1059,21 +1046,24 @@ "version": "3.1.2", "resolved": "https://registry.npmjs.org/@jridgewell/resolve-uri/-/resolve-uri-3.1.2.tgz", "integrity": "sha512-bRISgCIjP20/tbWSPWMEi54QVPRZExkuD9lJL+UIxUKtwVJA8wW1Trb1jMs1RFXo1CBTNZ/5hpC9QvmKWdopKw==", + "dev": true, "license": "MIT", "engines": { "node": ">=6.0.0" } }, "node_modules/@jridgewell/sourcemap-codec": { - "version": "1.5.4", - "resolved": "https://registry.npmjs.org/@jridgewell/sourcemap-codec/-/sourcemap-codec-1.5.4.tgz", - "integrity": "sha512-VT2+G1VQs/9oz078bLrYbecdZKs912zQlkelYpuf+SXF+QvZDYJlbx/LSx+meSAwdDFnF8FVXW92AVjjkVmgFw==", + "version": "1.5.5", + "resolved": "https://registry.npmjs.org/@jridgewell/sourcemap-codec/-/sourcemap-codec-1.5.5.tgz", + "integrity": "sha512-cYQ9310grqxueWbl+WuIUIaiUaDcj7WOq5fVhEljNVgRfOUhY9fy2zTvfoqWsnebh8Sl70VScFbICvJnLKB0Og==", + "dev": true, "license": "MIT" }, "node_modules/@jridgewell/trace-mapping": { "version": "0.3.31", "resolved": "https://registry.npmjs.org/@jridgewell/trace-mapping/-/trace-mapping-0.3.31.tgz", "integrity": "sha512-zzNR+SdQSDJzc8joaeP8QQoCQr8NuYx2dIIytl1QeBEZHJ9uW6hebsrYgbz8hJwUQao3TWCMtmfV8Nu1twOLAw==", + "dev": true, "license": "MIT", "dependencies": { "@jridgewell/resolve-uri": "^3.1.0", @@ -1112,6 +1102,7 @@ "version": "1.1.1", "resolved": "https://registry.npmjs.org/@mermaid-js/parser/-/parser-1.1.1.tgz", "integrity": "sha512-VuHdsYMK1bT6X2JbcAaWAhugTRvRBRyuZgd+c22swUeI9g/ntaxF7CY7dYarhZovofCbUNO0G7JesfmNtjYOCw==", + "dev": true, "license": "MIT", "dependencies": { "@chevrotain/types": "~11.1.1" @@ -1121,6 +1112,7 @@ "version": "1.26.0", "resolved": "https://registry.npmjs.org/@modelcontextprotocol/sdk/-/sdk-1.26.0.tgz", "integrity": "sha512-Y5RmPncpiDtTXDbLKswIJzTqu2hyBKxTNsgKqKclDbhIgg1wgtf1fRuvxgTnRfcnxtvvgbIEcqUOzZrJ6iSReg==", + "dev": true, "license": "MIT", "dependencies": { "@hono/node-server": "^1.19.9", @@ -1161,6 +1153,7 @@ "version": "8.18.0", "resolved": "https://registry.npmjs.org/ajv/-/ajv-8.18.0.tgz", "integrity": "sha512-PlXPeEWMXMZ7sPYOHqmDyCJzcfNrUr3fGNKtezX14ykXOEIvyK81d+qydx89KY5O71FKMPaQ2vBfBFI5NHR63A==", + "dev": true, "license": "MIT", "dependencies": { "fast-deep-equal": "^3.1.3", @@ -1177,12 +1170,14 @@ "version": "1.0.0", "resolved": "https://registry.npmjs.org/json-schema-traverse/-/json-schema-traverse-1.0.0.tgz", "integrity": "sha512-NM8/P9n3XjXhIZn1lLhkFaACTOURQXjWhV4BA/RnOv8xvgqtqpAX9IO4mRQxSx1Rlo4tqzeqb0sOlruaOy3dug==", + "dev": true, "license": "MIT" }, "node_modules/@napi-rs/canvas": { "version": "0.1.76", "resolved": "https://registry.npmjs.org/@napi-rs/canvas/-/canvas-0.1.76.tgz", "integrity": "sha512-YIk5okeNN53GzjvWmAyCQFE9xrLeQXzYpudX4TiLvqaz9SqXgIgxIuKPe4DKyB5nccsQMIev7JGKTzZaN5rFdw==", + "dev": true, "license": "MIT", "optional": true, "workspaces": [ @@ -1211,6 +1206,7 @@ "cpu": [ "arm64" ], + "dev": true, "license": "MIT", "optional": true, "os": [ @@ -1227,6 +1223,7 @@ "cpu": [ "arm64" ], + "dev": true, "license": "MIT", "optional": true, "os": [ @@ -1243,6 +1240,7 @@ "cpu": [ "x64" ], + "dev": true, "license": "MIT", "optional": true, "os": [ @@ -1259,6 +1257,7 @@ "cpu": [ "arm" ], + "dev": true, "license": "MIT", "optional": true, "os": [ @@ -1275,6 +1274,7 @@ "cpu": [ "arm64" ], + "dev": true, "license": "MIT", "optional": true, "os": [ @@ -1291,6 +1291,7 @@ "cpu": [ "arm64" ], + "dev": true, "license": "MIT", "optional": true, "os": [ @@ -1307,6 +1308,7 @@ "cpu": [ "riscv64" ], + "dev": true, "license": "MIT", "optional": true, "os": [ @@ -1323,6 +1325,7 @@ "cpu": [ "x64" ], + "dev": true, "license": "MIT", "optional": true, "os": [ @@ -1339,6 +1342,7 @@ "cpu": [ "x64" ], + "dev": true, "license": "MIT", "optional": true, "os": [ @@ -1355,6 +1359,7 @@ "cpu": [ "x64" ], + "dev": true, "license": "MIT", "optional": true, "os": [ @@ -1695,17 +1700,6 @@ "node": ">=0.10" } }, - "node_modules/@pkgjs/parseargs": { - "version": "0.11.0", - "resolved": "https://registry.npmjs.org/@pkgjs/parseargs/-/parseargs-0.11.0.tgz", - "integrity": "sha512-+1VkjdD0QBLPodGrJUeqarH8VAIvQODIbwh9XpP5Syisf7YoQgsJKPNFoqqLQlu+VQ/tVSshMR6loPMn8U+dPg==", - "dev": true, - "license": "MIT", - "optional": true, - "engines": { - "node": ">=14" - } - }, "node_modules/@playwright/test": { "version": "1.56.1", "resolved": "https://registry.npmjs.org/@playwright/test/-/test-1.56.1.tgz", @@ -2080,9 +2074,9 @@ ] }, "node_modules/@standard-schema/spec": { - "version": "1.0.0", - "resolved": "https://registry.npmjs.org/@standard-schema/spec/-/spec-1.0.0.tgz", - "integrity": "sha512-m2bOd0f2RT9k8QJx1JN85cZYyH1RqFBdlwtkSlf4tBDYLCiiZnv1fIIwacK6cqwXavOydf0NPToMQgpKq+dVlA==", + "version": "1.1.0", + "resolved": "https://registry.npmjs.org/@standard-schema/spec/-/spec-1.1.0.tgz", + "integrity": "sha512-l2aFy5jALhniG5HgqrD6jXLi/rUWrKvqN/qJx6yoJsgKhblVd+iqqU4RCXavm/jPityDo5TCvKMnpjKnOriy0w==", "dev": true, "license": "MIT" }, @@ -2352,6 +2346,7 @@ "version": "1.0.5", "resolved": "https://registry.npmjs.org/@sveltejs/acorn-typescript/-/acorn-typescript-1.0.5.tgz", "integrity": "sha512-IwQk4yfwLdibDlrXVE04jTZYlLnwsTT2PIOQQGNLWfjavGifnk1JD1LcZjZaBTRcxZu2FfPfNLOE04DSu9lqtQ==", + "dev": true, "license": "MIT", "peerDependencies": { "acorn": "^8.9.0" @@ -2825,19 +2820,20 @@ } }, "node_modules/@testing-library/dom": { - "version": "10.4.0", - "resolved": "https://registry.npmjs.org/@testing-library/dom/-/dom-10.4.0.tgz", - "integrity": "sha512-pemlzrSESWbdAloYml3bAJMEfNh1Z7EduzqPKprCH5S341frlpYnUEW0H72dLxa6IsYr+mPno20GiSm+h9dEdQ==", + "version": "10.4.1", + "resolved": "https://registry.npmjs.org/@testing-library/dom/-/dom-10.4.1.tgz", + "integrity": "sha512-o4PXJQidqJl82ckFaXUeoAW+XysPLauYI43Abki5hABd853iMhitooc6znOnczgbTYmEP6U6/y1ZyKAIsvMKGg==", "dev": true, "license": "MIT", + "peer": true, "dependencies": { "@babel/code-frame": "^7.10.4", "@babel/runtime": "^7.12.5", "@types/aria-query": "^5.0.1", "aria-query": "5.3.0", - "chalk": "^4.1.0", "dom-accessibility-api": "^0.5.9", "lz-string": "^1.5.0", + "picocolors": "1.1.1", "pretty-format": "^27.0.2" }, "engines": { @@ -2871,6 +2867,19 @@ "dev": true, "license": "MIT" }, + "node_modules/@testing-library/svelte-core": { + "version": "1.0.0", + "resolved": "https://registry.npmjs.org/@testing-library/svelte-core/-/svelte-core-1.0.0.tgz", + "integrity": "sha512-VkUePoLV6oOYwSUvX6ShA8KLnJqZiYMIbP2JW2t0GLWLkJxKGvuH5qrrZBV/X7cXFnLGuFQEC7RheYiZOW68KQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=16" + }, + "peerDependencies": { + "svelte": "^3 || ^4 || ^5 || ^5.0.0-next.0" + } + }, "node_modules/@testing-library/user-event": { "version": "14.6.1", "resolved": "https://registry.npmjs.org/@testing-library/user-event/-/user-event-14.6.1.tgz", @@ -2890,7 +2899,8 @@ "resolved": "https://registry.npmjs.org/@types/aria-query/-/aria-query-5.0.4.tgz", "integrity": "sha512-rfT93uj5s0PRL7EzccGMs3brplhcrghnDoV26NqKhCAS1hVo+WdNsPvE/yb6ilfr5hi2MEk6d5EWJTKdxg8jVw==", "dev": true, - "license": "MIT" + "license": "MIT", + "peer": true }, "node_modules/@types/chai": { "version": "5.2.2", @@ -2913,6 +2923,7 @@ "version": "7.4.3", "resolved": "https://registry.npmjs.org/@types/d3/-/d3-7.4.3.tgz", "integrity": "sha512-lZXZ9ckh5R8uiFVt8ogUNf+pIrK4EsWrx2Np75WvF/eTpJ0FMHNhjXk8CKEx/+gpHbNQyJWehbFaTvqmHWB3ww==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-array": "*", @@ -2951,12 +2962,14 @@ "version": "3.2.2", "resolved": "https://registry.npmjs.org/@types/d3-array/-/d3-array-3.2.2.tgz", "integrity": "sha512-hOLWVbm7uRza0BYXpIIW5pxfrKe0W+D5lrFiAEYR+pb6w3N2SwSMaJbXdUfSEv+dT4MfHBLtn5js0LAWaO6otw==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-axis": { "version": "3.0.6", "resolved": "https://registry.npmjs.org/@types/d3-axis/-/d3-axis-3.0.6.tgz", "integrity": "sha512-pYeijfZuBd87T0hGn0FO1vQ/cgLk6E1ALJjfkC0oJ8cbwkZl3TpgS8bVBLZN+2jjGgg38epgxb2zmoGtSfvgMw==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-selection": "*" @@ -2966,6 +2979,7 @@ "version": "3.0.6", "resolved": "https://registry.npmjs.org/@types/d3-brush/-/d3-brush-3.0.6.tgz", "integrity": "sha512-nH60IZNNxEcrh6L1ZSMNA28rj27ut/2ZmI3r96Zd+1jrZD++zD3LsMIjWlvg4AYrHn/Pqz4CF3veCxGjtbqt7A==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-selection": "*" @@ -2975,18 +2989,21 @@ "version": "3.0.6", "resolved": "https://registry.npmjs.org/@types/d3-chord/-/d3-chord-3.0.6.tgz", "integrity": "sha512-LFYWWd8nwfwEmTZG9PfQxd17HbNPksHBiJHaKuY1XeqscXacsS2tyoo6OdRsjf+NQYeB6XrNL3a25E3gH69lcg==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-color": { "version": "3.1.3", "resolved": "https://registry.npmjs.org/@types/d3-color/-/d3-color-3.1.3.tgz", "integrity": "sha512-iO90scth9WAbmgv7ogoq57O9YpKmFBbmoEoCHDB2xMBY0+/KVrqAaCDyCE16dUspeOvIxFFRI+0sEtqDqy2b4A==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-contour": { "version": "3.0.6", "resolved": "https://registry.npmjs.org/@types/d3-contour/-/d3-contour-3.0.6.tgz", "integrity": "sha512-BjzLgXGnCWjUSYGfH1cpdo41/hgdWETu4YxpezoztawmqsvCeep+8QGfiY6YbDvfgHz/DkjeIkkZVJavB4a3rg==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-array": "*", @@ -2997,18 +3014,21 @@ "version": "6.0.4", "resolved": "https://registry.npmjs.org/@types/d3-delaunay/-/d3-delaunay-6.0.4.tgz", "integrity": "sha512-ZMaSKu4THYCU6sV64Lhg6qjf1orxBthaC161plr5KuPHo3CNm8DTHiLw/5Eq2b6TsNP0W0iJrUOFscY6Q450Hw==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-dispatch": { "version": "3.0.7", "resolved": "https://registry.npmjs.org/@types/d3-dispatch/-/d3-dispatch-3.0.7.tgz", "integrity": "sha512-5o9OIAdKkhN1QItV2oqaE5KMIiXAvDWBDPrD85e58Qlz1c1kI/J0NcqbEG88CoTwJrYe7ntUCVfeUl2UJKbWgA==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-drag": { "version": "3.0.7", "resolved": "https://registry.npmjs.org/@types/d3-drag/-/d3-drag-3.0.7.tgz", "integrity": "sha512-HE3jVKlzU9AaMazNufooRJ5ZpWmLIoc90A37WU2JMmeq28w1FQqCZswHZ3xR+SuxYftzHq6WU6KJHvqxKzTxxQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-selection": "*" @@ -3018,18 +3038,21 @@ "version": "3.0.7", "resolved": "https://registry.npmjs.org/@types/d3-dsv/-/d3-dsv-3.0.7.tgz", "integrity": "sha512-n6QBF9/+XASqcKK6waudgL0pf/S5XHPPI8APyMLLUHd8NqouBGLsU8MgtO7NINGtPBtk9Kko/W4ea0oAspwh9g==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-ease": { "version": "3.0.2", "resolved": "https://registry.npmjs.org/@types/d3-ease/-/d3-ease-3.0.2.tgz", "integrity": "sha512-NcV1JjO5oDzoK26oMzbILE6HW7uVXOHLQvHshBUW4UMdZGfiY6v5BeQwh9a9tCzv+CeefZQHJt5SRgK154RtiA==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-fetch": { "version": "3.0.7", "resolved": "https://registry.npmjs.org/@types/d3-fetch/-/d3-fetch-3.0.7.tgz", "integrity": "sha512-fTAfNmxSb9SOWNB9IoG5c8Hg6R+AzUHDRlsXsDZsNp6sxAEOP0tkP3gKkNSO/qmHPoBFTxNrjDprVHDQDvo5aA==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-dsv": "*" @@ -3039,18 +3062,21 @@ "version": "3.0.10", "resolved": "https://registry.npmjs.org/@types/d3-force/-/d3-force-3.0.10.tgz", "integrity": "sha512-ZYeSaCF3p73RdOKcjj+swRlZfnYpK1EbaDiYICEEp5Q6sUiqFaFQ9qgoshp5CzIyyb/yD09kD9o2zEltCexlgw==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-format": { "version": "3.0.4", "resolved": "https://registry.npmjs.org/@types/d3-format/-/d3-format-3.0.4.tgz", "integrity": "sha512-fALi2aI6shfg7vM5KiR1wNJnZ7r6UuggVqtDA+xiEdPZQwy/trcQaHnwShLuLdta2rTymCNpxYTiMZX/e09F4g==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-geo": { "version": "3.1.0", "resolved": "https://registry.npmjs.org/@types/d3-geo/-/d3-geo-3.1.0.tgz", "integrity": "sha512-856sckF0oP/diXtS4jNsiQw/UuK5fQG8l/a9VVLeSouf1/PPbBE1i1W852zVwKwYCBkFJJB7nCFTbk6UMEXBOQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/geojson": "*" @@ -3060,12 +3086,14 @@ "version": "3.1.7", "resolved": "https://registry.npmjs.org/@types/d3-hierarchy/-/d3-hierarchy-3.1.7.tgz", "integrity": "sha512-tJFtNoYBtRtkNysX1Xq4sxtjK8YgoWUNpIiUee0/jHGRwqvzYxkq0hGVbbOGSz+JgFxxRu4K8nb3YpG3CMARtg==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-interpolate": { "version": "3.0.4", "resolved": "https://registry.npmjs.org/@types/d3-interpolate/-/d3-interpolate-3.0.4.tgz", "integrity": "sha512-mgLPETlrpVV1YRJIglr4Ez47g7Yxjl1lj7YKsiMCb27VJH9W8NVM6Bb9d8kkpG/uAQS5AmbA48q2IAolKKo1MA==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-color": "*" @@ -3075,30 +3103,35 @@ "version": "3.1.1", "resolved": "https://registry.npmjs.org/@types/d3-path/-/d3-path-3.1.1.tgz", "integrity": "sha512-VMZBYyQvbGmWyWVea0EHs/BwLgxc+MKi1zLDCONksozI4YJMcTt8ZEuIR4Sb1MMTE8MMW49v0IwI5+b7RmfWlg==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-polygon": { "version": "3.0.2", "resolved": "https://registry.npmjs.org/@types/d3-polygon/-/d3-polygon-3.0.2.tgz", "integrity": "sha512-ZuWOtMaHCkN9xoeEMr1ubW2nGWsp4nIql+OPQRstu4ypeZ+zk3YKqQT0CXVe/PYqrKpZAi+J9mTs05TKwjXSRA==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-quadtree": { "version": "3.0.6", "resolved": "https://registry.npmjs.org/@types/d3-quadtree/-/d3-quadtree-3.0.6.tgz", "integrity": "sha512-oUzyO1/Zm6rsxKRHA1vH0NEDG58HrT5icx/azi9MF1TWdtttWl0UIUsjEQBBh+SIkrpd21ZjEv7ptxWys1ncsg==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-random": { "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/d3-random/-/d3-random-3.0.3.tgz", "integrity": "sha512-Imagg1vJ3y76Y2ea0871wpabqp613+8/r0mCLEBfdtqC7xMSfj9idOnmBYyMoULfHePJyxMAw3nWhJxzc+LFwQ==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-scale": { "version": "4.0.9", "resolved": "https://registry.npmjs.org/@types/d3-scale/-/d3-scale-4.0.9.tgz", "integrity": "sha512-dLmtwB8zkAeO/juAMfnV+sItKjlsw2lKdZVVy6LRr0cBmegxSABiLEpGVmSJJ8O08i4+sGR6qQtb6WtuwJdvVw==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-time": "*" @@ -3108,18 +3141,21 @@ "version": "3.1.0", "resolved": "https://registry.npmjs.org/@types/d3-scale-chromatic/-/d3-scale-chromatic-3.1.0.tgz", "integrity": "sha512-iWMJgwkK7yTRmWqRB5plb1kadXyQ5Sj8V/zYlFGMUBbIPKQScw+Dku9cAAMgJG+z5GYDoMjWGLVOvjghDEFnKQ==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-selection": { "version": "3.0.11", "resolved": "https://registry.npmjs.org/@types/d3-selection/-/d3-selection-3.0.11.tgz", "integrity": "sha512-bhAXu23DJWsrI45xafYpkQ4NtcKMwWnAC/vKrd2l+nxMFuvOT3XMYTIj2opv8vq8AO5Yh7Qac/nSeP/3zjTK0w==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-shape": { "version": "3.1.8", "resolved": "https://registry.npmjs.org/@types/d3-shape/-/d3-shape-3.1.8.tgz", "integrity": "sha512-lae0iWfcDeR7qt7rA88BNiqdvPS5pFVPpo5OfjElwNaT2yyekbM0C9vK+yqBqEmHr6lDkRnYNoTBYlAgJa7a4w==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-path": "*" @@ -3129,24 +3165,28 @@ "version": "3.0.4", "resolved": "https://registry.npmjs.org/@types/d3-time/-/d3-time-3.0.4.tgz", "integrity": "sha512-yuzZug1nkAAaBlBBikKZTgzCeA+k1uy4ZFwWANOfKw5z5LRhV0gNA7gNkKm7HoK+HRN0wX3EkxGk0fpbWhmB7g==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-time-format": { "version": "4.0.3", "resolved": "https://registry.npmjs.org/@types/d3-time-format/-/d3-time-format-4.0.3.tgz", "integrity": "sha512-5xg9rC+wWL8kdDj153qZcsJ0FWiFt0J5RB6LYUNZjwSnesfblqrI/bJ1wBdJ8OQfncgbJG5+2F+qfqnqyzYxyg==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-timer": { "version": "3.0.2", "resolved": "https://registry.npmjs.org/@types/d3-timer/-/d3-timer-3.0.2.tgz", "integrity": "sha512-Ps3T8E8dZDam6fUyNiMkekK3XUsaUEik+idO9/YjPtfj2qruF8tFBXS7XhtE4iIXBLxhmLjP3SXpLhVf21I9Lw==", + "dev": true, "license": "MIT" }, "node_modules/@types/d3-transition": { "version": "3.0.9", "resolved": "https://registry.npmjs.org/@types/d3-transition/-/d3-transition-3.0.9.tgz", "integrity": "sha512-uZS5shfxzO3rGlu0cC3bjmMFKsXv+SmZZcgp0KD22ts4uGXp5EVYGzu/0YdwZeKmddhcAccYtREJKkPfXkZuCg==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-selection": "*" @@ -3156,6 +3196,7 @@ "version": "3.0.8", "resolved": "https://registry.npmjs.org/@types/d3-zoom/-/d3-zoom-3.0.8.tgz", "integrity": "sha512-iqMC4/YlFCSlO8+2Ii1GGGliCAY4XdeG748w5vQUbevlbDu0zSjH/+jojorQVBK/se0j6DUFNPBGSqD3YWYnDw==", + "dev": true, "license": "MIT", "dependencies": { "@types/d3-interpolate": "*", @@ -3166,6 +3207,7 @@ "version": "4.1.12", "resolved": "https://registry.npmjs.org/@types/debug/-/debug-4.1.12.tgz", "integrity": "sha512-vIChWdVG3LG1SMxEvI/AK+FWJthlrqlTu7fbrlywTkkaONwk/UAGaULXRlf8vkzFBLVm0zkMdCquhL5aOjhXPQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/ms": "*" @@ -3182,18 +3224,21 @@ "version": "1.0.8", "resolved": "https://registry.npmjs.org/@types/estree/-/estree-1.0.8.tgz", "integrity": "sha512-dWHzHa2WqEXI/O1E9OjrocMTKJl2mSrEolh1Iomrv6U+JuNwaHXsXx9bLu5gG7BUWFIN0skIQJQ/L1rIex4X6w==", + "dev": true, "license": "MIT" }, "node_modules/@types/geojson": { "version": "7946.0.16", "resolved": "https://registry.npmjs.org/@types/geojson/-/geojson-7946.0.16.tgz", "integrity": "sha512-6C8nqWur3j98U6+lXDfTUWIfgvZU+EumvpHKcYjujKH7woYyLj2sUmff0tRhrqM7BohUw7Pz3ZB1jj2gW9Fvmg==", + "dev": true, "license": "MIT" }, "node_modules/@types/hast": { "version": "3.0.4", "resolved": "https://registry.npmjs.org/@types/hast/-/hast-3.0.4.tgz", "integrity": "sha512-WPs+bbQw5aCj+x6laNGWLH3wviHtoCv/P3+otBhbOhJgG8qtpdAMlTCxLtsTWA7LH1Oh/bFCHsBn0TPS5m30EQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "*" @@ -3217,6 +3262,7 @@ "version": "4.0.4", "resolved": "https://registry.npmjs.org/@types/mdast/-/mdast-4.0.4.tgz", "integrity": "sha512-kGaNbPh1k7AFzgpud/gMdvIm5xuECykRR+JnWKQno9TAXVa6WIVCGTPvYGekIDL4uwCZQSYbUxNBSb1aUo79oA==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "*" @@ -3233,6 +3279,7 @@ "version": "2.1.0", "resolved": "https://registry.npmjs.org/@types/ms/-/ms-2.1.0.tgz", "integrity": "sha512-GsCCIZDE/p3i96vtEqx+7dBUGXrc7zeSK3wwPHIaRThS+9OhWIXRqzs4d6k1SVU8g91DrNRWxWUGhp5KXQb2VA==", + "dev": true, "license": "MIT" }, "node_modules/@types/node": { @@ -3246,26 +3293,28 @@ } }, "node_modules/@types/react": { - "version": "19.1.8", - "resolved": "https://registry.npmjs.org/@types/react/-/react-19.1.8.tgz", - "integrity": "sha512-AwAfQ2Wa5bCx9WP8nZL2uMZWod7J7/JSplxbTmBQ5ms6QpqNYm672H0Vu9ZVKVngQ+ii4R/byguVEUZQyeg44g==", + "version": "19.2.16", + "resolved": "https://registry.npmjs.org/@types/react/-/react-19.2.16.tgz", + "integrity": "sha512-esJiCAnl0kfpNdE69f3So4WJUXy95dLZydX0KwK46riIHDzHM7O9Vtf9xCHW0PXIqvgqNrswl522kA/5yx+F4w==", "dev": true, "license": "MIT", "peer": true, "dependencies": { - "csstype": "^3.0.2" + "csstype": "^3.2.2" } }, "node_modules/@types/trusted-types": { "version": "2.0.7", "resolved": "https://registry.npmjs.org/@types/trusted-types/-/trusted-types-2.0.7.tgz", "integrity": "sha512-ScaPdn1dQczgbl0QFTeTOmVHFULt394XJgOQNoyVhZ6r2vLnMLJfBPd53SB52T/3G36VI1/g2MZaX0cwDuXsfw==", + "dev": true, "license": "MIT" }, "node_modules/@types/unist": { "version": "2.0.11", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-2.0.11.tgz", "integrity": "sha512-CmBKiL6NNo/OqgmMn95Fk9Whlp2mtvIv+KNpQKN2F4SjvrEesubTRWGYSg+BnWZOnlCaSTU1sMpsBOzgbYhnsA==", + "dev": true, "license": "MIT" }, "node_modules/@typescript-eslint/eslint-plugin": { @@ -3418,6 +3467,7 @@ "version": "8.56.0", "resolved": "https://registry.npmjs.org/@typescript-eslint/types/-/types-8.56.0.tgz", "integrity": "sha512-DBsLPs3GsWhX5HylbP9HNG15U0bnwut55Lx12bHB9MpXxQ+R5GC8MwQe+N1UFXxAeQDvEsEDY6ZYwX03K7Z6HQ==", + "dev": true, "license": "MIT", "engines": { "node": "^18.18.0 || ^20.9.0 || >=21.1.0" @@ -3540,12 +3590,14 @@ "version": "1.3.0", "resolved": "https://registry.npmjs.org/@ungap/structured-clone/-/structured-clone-1.3.0.tgz", "integrity": "sha512-WmoN8qaIAo7WTYWbAZuG8PYEhn5fkz7dZrqTBZ7dtt//lL2Gwms1IcnQ5yHqjDfX8Ft5j4YzDM23f87zBfDe9g==", + "dev": true, "license": "ISC" }, "node_modules/@upsetjs/venn.js": { "version": "2.0.0", "resolved": "https://registry.npmjs.org/@upsetjs/venn.js/-/venn.js-2.0.0.tgz", "integrity": "sha512-WbBhLrooyePuQ1VZxrJjtLvTc4NVfpOyKx0sKqioq9bX1C1m7Jgykkn8gLrtwumBioXIqam8DLxp88Adbue6Hw==", + "dev": true, "license": "MIT", "optionalDependencies": { "d3-selection": "^3.0.0", @@ -3553,68 +3605,124 @@ } }, "node_modules/@vitest/browser": { - "version": "3.2.4", - "resolved": "https://registry.npmjs.org/@vitest/browser/-/browser-3.2.4.tgz", - "integrity": "sha512-tJxiPrWmzH8a+w9nLKlQMzAKX/7VjFs50MWgcAj7p9XQ7AQ9/35fByFYptgPELyLw+0aixTnC4pUWV+APcZ/kw==", + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/browser/-/browser-4.1.8.tgz", + "integrity": "sha512-u21VzX07HzlJYpFgkxmjEXar/tG2UqWGgyGG/46SrrPc7rSdCTPw5vuowopO9CIqF8UCUQzDFdbVnNpw6N0BfQ==", "dev": true, "license": "MIT", "dependencies": { - "@testing-library/dom": "^10.4.0", - "@testing-library/user-event": "^14.6.1", - "@vitest/mocker": "3.2.4", - "@vitest/utils": "3.2.4", - "magic-string": "^0.30.17", - "sirv": "^3.0.1", - "tinyrainbow": "^2.0.0", - "ws": "^8.18.2" + "@blazediff/core": "1.9.1", + "@vitest/mocker": "4.1.8", + "@vitest/utils": "4.1.8", + "magic-string": "^0.30.21", + "pngjs": "^7.0.0", + "sirv": "^3.0.2", + "tinyrainbow": "^3.1.0", + "ws": "^8.19.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + }, + "peerDependencies": { + "vitest": "4.1.8" + } + }, + "node_modules/@vitest/browser-playwright": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/browser-playwright/-/browser-playwright-4.1.8.tgz", + "integrity": "sha512-SR7FqgegaexEg73xvf3ArtygXegagMdXnL0EZMpxrWvvhQxvicD/E8p0ib0J91riPRtQUViyh67Xjw3NqvyhVg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/browser": "4.1.8", + "@vitest/mocker": "4.1.8", + "tinyrainbow": "^3.1.0" }, "funding": { "url": "https://opencollective.com/vitest" }, "peerDependencies": { "playwright": "*", - "vitest": "3.2.4", - "webdriverio": "^7.0.0 || ^8.0.0 || ^9.0.0" + "vitest": "4.1.8" }, "peerDependenciesMeta": { "playwright": { - "optional": true - }, - "safaridriver": { - "optional": true - }, - "webdriverio": { - "optional": true + "optional": false } } }, + "node_modules/@vitest/browser-playwright/node_modules/tinyrainbow": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/tinyrainbow/-/tinyrainbow-3.1.0.tgz", + "integrity": "sha512-Bf+ILmBgretUrdJxzXM0SgXLZ3XfiaUuOj/IKQHuTXip+05Xn+uyEYdVg0kYDipTBcLrCVyUzAPz7QmArb0mmw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/@vitest/browser/node_modules/@vitest/pretty-format": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/pretty-format/-/pretty-format-4.1.8.tgz", + "integrity": "sha512-9GasEBxpZ1VYIpqHf/0+YGg121uSNwCKOJqIrTwWP/TB7DmFCiaBpNl3aPZzoLWfWkuqhbH8vJIVobZkvdo2cA==", + "dev": true, + "license": "MIT", + "dependencies": { + "tinyrainbow": "^3.1.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/browser/node_modules/@vitest/utils": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/utils/-/utils-4.1.8.tgz", + "integrity": "sha512-uOJamYALNhfJ6iolExyQM40yIQwDqYnkKtQ5VCiSe17E33H0aQ/u+1GlRuz4LZBk6Mm3sg90G9hEbmEt37C1Zg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/pretty-format": "4.1.8", + "convert-source-map": "^2.0.0", + "tinyrainbow": "^3.1.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/browser/node_modules/tinyrainbow": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/tinyrainbow/-/tinyrainbow-3.1.0.tgz", + "integrity": "sha512-Bf+ILmBgretUrdJxzXM0SgXLZ3XfiaUuOj/IKQHuTXip+05Xn+uyEYdVg0kYDipTBcLrCVyUzAPz7QmArb0mmw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.0.0" + } + }, "node_modules/@vitest/coverage-v8": { - "version": "3.2.4", - "resolved": "https://registry.npmjs.org/@vitest/coverage-v8/-/coverage-v8-3.2.4.tgz", - "integrity": "sha512-EyF9SXU6kS5Ku/U82E259WSnvg6c8KTjppUncuNdm5QHpe17mwREHnjDzozC8x9MZ0xfBUFSaLkRv4TMA75ALQ==", + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/coverage-v8/-/coverage-v8-4.1.8.tgz", + "integrity": "sha512-lt3kovsyHwYe00wq4D1ti0Z974fWj4NLp6siqiyEufUpyFwK9Yhi7rBhac9JL5aA0zoMrJqc4vYPZRUnI7l7nw==", "dev": true, "license": "MIT", "dependencies": { - "@ampproject/remapping": "^2.3.0", "@bcoe/v8-coverage": "^1.0.2", - "ast-v8-to-istanbul": "^0.3.3", - "debug": "^4.4.1", + "@vitest/utils": "4.1.8", + "ast-v8-to-istanbul": "^1.0.0", "istanbul-lib-coverage": "^3.2.2", "istanbul-lib-report": "^3.0.1", - "istanbul-lib-source-maps": "^5.0.6", - "istanbul-reports": "^3.1.7", - "magic-string": "^0.30.17", - "magicast": "^0.3.5", - "std-env": "^3.9.0", - "test-exclude": "^7.0.1", - "tinyrainbow": "^2.0.0" + "istanbul-reports": "^3.2.0", + "magicast": "^0.5.2", + "obug": "^2.1.1", + "std-env": "^4.0.0-rc.1", + "tinyrainbow": "^3.1.0" }, "funding": { "url": "https://opencollective.com/vitest" }, "peerDependencies": { - "@vitest/browser": "3.2.4", - "vitest": "3.2.4" + "@vitest/browser": "4.1.8", + "vitest": "4.1.8" }, "peerDependenciesMeta": { "@vitest/browser": { @@ -3622,6 +3730,44 @@ } } }, + "node_modules/@vitest/coverage-v8/node_modules/@vitest/pretty-format": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/pretty-format/-/pretty-format-4.1.8.tgz", + "integrity": "sha512-9GasEBxpZ1VYIpqHf/0+YGg121uSNwCKOJqIrTwWP/TB7DmFCiaBpNl3aPZzoLWfWkuqhbH8vJIVobZkvdo2cA==", + "dev": true, + "license": "MIT", + "dependencies": { + "tinyrainbow": "^3.1.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/coverage-v8/node_modules/@vitest/utils": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/utils/-/utils-4.1.8.tgz", + "integrity": "sha512-uOJamYALNhfJ6iolExyQM40yIQwDqYnkKtQ5VCiSe17E33H0aQ/u+1GlRuz4LZBk6Mm3sg90G9hEbmEt37C1Zg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/pretty-format": "4.1.8", + "convert-source-map": "^2.0.0", + "tinyrainbow": "^3.1.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/coverage-v8/node_modules/tinyrainbow": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/tinyrainbow/-/tinyrainbow-3.1.0.tgz", + "integrity": "sha512-Bf+ILmBgretUrdJxzXM0SgXLZ3XfiaUuOj/IKQHuTXip+05Xn+uyEYdVg0kYDipTBcLrCVyUzAPz7QmArb0mmw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.0.0" + } + }, "node_modules/@vitest/expect": { "version": "3.2.4", "resolved": "https://registry.npmjs.org/@vitest/expect/-/expect-3.2.4.tgz", @@ -3640,22 +3786,22 @@ } }, "node_modules/@vitest/mocker": { - "version": "3.2.4", - "resolved": "https://registry.npmjs.org/@vitest/mocker/-/mocker-3.2.4.tgz", - "integrity": "sha512-46ryTE9RZO/rfDd7pEqFl7etuyzekzEhUbTW3BvmeO/BcCMEgq59BKhek3dXDWgAj4oMK6OZi+vRr1wPW6qjEQ==", + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/mocker/-/mocker-4.1.8.tgz", + "integrity": "sha512-LEiN/xe4OSIbKe9HQIp5OC24agGD9J5CnmMgsLohVVoOPWL9a2sBoR6VBx43jQZb7Kr1l4RCuyCJzcAa0+dojw==", "dev": true, "license": "MIT", "dependencies": { - "@vitest/spy": "3.2.4", + "@vitest/spy": "4.1.8", "estree-walker": "^3.0.3", - "magic-string": "^0.30.17" + "magic-string": "^0.30.21" }, "funding": { "url": "https://opencollective.com/vitest" }, "peerDependencies": { "msw": "^2.4.9", - "vite": "^5.0.0 || ^6.0.0 || ^7.0.0-0" + "vite": "^6.0.0 || ^7.0.0 || ^8.0.0" }, "peerDependenciesMeta": { "msw": { @@ -3666,6 +3812,16 @@ } } }, + "node_modules/@vitest/mocker/node_modules/@vitest/spy": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/spy/-/spy-4.1.8.tgz", + "integrity": "sha512-6EevtBp6OZOPF7bmz36HrGMeP3txgVSrgebWxHOafDXGkhIzfXK14f8KF6MuFfgXXUeHxmpD3BQxkV00/3s5mA==", + "dev": true, + "license": "MIT", + "funding": { + "url": "https://opencollective.com/vitest" + } + }, "node_modules/@vitest/pretty-format": { "version": "3.2.4", "resolved": "https://registry.npmjs.org/@vitest/pretty-format/-/pretty-format-3.2.4.tgz", @@ -3680,71 +3836,148 @@ } }, "node_modules/@vitest/runner": { - "version": "3.2.4", - "resolved": "https://registry.npmjs.org/@vitest/runner/-/runner-3.2.4.tgz", - "integrity": "sha512-oukfKT9Mk41LreEW09vt45f8wx7DordoWUZMYdY/cyAk7w5TWkTRCNZYF7sX7n2wB7jyGAl74OxgwhPgKaqDMQ==", + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/runner/-/runner-4.1.8.tgz", + "integrity": "sha512-EmVxeBAfMJvycdjd6Hm+RbFBbA9fKvo0Kx37hNpBYoYeavH3RNsBXWDooR1mgD52dCrxIIuP7UotpfiwOikvcg==", "dev": true, "license": "MIT", "dependencies": { - "@vitest/utils": "3.2.4", - "pathe": "^2.0.3", - "strip-literal": "^3.0.0" + "@vitest/utils": "4.1.8", + "pathe": "^2.0.3" }, "funding": { "url": "https://opencollective.com/vitest" } }, - "node_modules/@vitest/snapshot": { - "version": "3.2.4", - "resolved": "https://registry.npmjs.org/@vitest/snapshot/-/snapshot-3.2.4.tgz", - "integrity": "sha512-dEYtS7qQP2CjU27QBC5oUOxLE/v5eLkGqPE0ZKEIDGMs4vKWe7IjgLOeauHsR0D5YuuycGRO5oSRXnwnmA78fQ==", + "node_modules/@vitest/runner/node_modules/@vitest/pretty-format": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/pretty-format/-/pretty-format-4.1.8.tgz", + "integrity": "sha512-9GasEBxpZ1VYIpqHf/0+YGg121uSNwCKOJqIrTwWP/TB7DmFCiaBpNl3aPZzoLWfWkuqhbH8vJIVobZkvdo2cA==", "dev": true, "license": "MIT", "dependencies": { - "@vitest/pretty-format": "3.2.4", - "magic-string": "^0.30.17", - "pathe": "^2.0.3" + "tinyrainbow": "^3.1.0" }, "funding": { "url": "https://opencollective.com/vitest" } }, - "node_modules/@vitest/spy": { - "version": "3.2.4", - "resolved": "https://registry.npmjs.org/@vitest/spy/-/spy-3.2.4.tgz", - "integrity": "sha512-vAfasCOe6AIK70iP5UD11Ac4siNUNJ9i/9PZ3NKx07sG6sUxeag1LWdNrMWeKKYBLlzuK+Gn65Yd5nyL6ds+nw==", + "node_modules/@vitest/runner/node_modules/@vitest/utils": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/utils/-/utils-4.1.8.tgz", + "integrity": "sha512-uOJamYALNhfJ6iolExyQM40yIQwDqYnkKtQ5VCiSe17E33H0aQ/u+1GlRuz4LZBk6Mm3sg90G9hEbmEt37C1Zg==", "dev": true, "license": "MIT", "dependencies": { - "tinyspy": "^4.0.3" + "@vitest/pretty-format": "4.1.8", + "convert-source-map": "^2.0.0", + "tinyrainbow": "^3.1.0" }, "funding": { "url": "https://opencollective.com/vitest" } }, - "node_modules/@vitest/utils": { - "version": "3.2.4", - "resolved": "https://registry.npmjs.org/@vitest/utils/-/utils-3.2.4.tgz", - "integrity": "sha512-fB2V0JFrQSMsCo9HiSq3Ezpdv4iYaXRG1Sx8edX3MwxfyNn83mKiGzOcH+Fkxt4MHxr3y42fQi1oeAInqgX2QA==", + "node_modules/@vitest/runner/node_modules/tinyrainbow": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/tinyrainbow/-/tinyrainbow-3.1.0.tgz", + "integrity": "sha512-Bf+ILmBgretUrdJxzXM0SgXLZ3XfiaUuOj/IKQHuTXip+05Xn+uyEYdVg0kYDipTBcLrCVyUzAPz7QmArb0mmw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/@vitest/snapshot": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/snapshot/-/snapshot-4.1.8.tgz", + "integrity": "sha512-acfZboRmAIf05DEKcBQy33VXojFJjtUdLyo7oOmV9kebb2xdU01UknNiPuPZoJZQyO7DF0gZdTGTpeAzET9QPQ==", "dev": true, "license": "MIT", "dependencies": { - "@vitest/pretty-format": "3.2.4", - "loupe": "^3.1.4", - "tinyrainbow": "^2.0.0" + "@vitest/pretty-format": "4.1.8", + "@vitest/utils": "4.1.8", + "magic-string": "^0.30.21", + "pathe": "^2.0.3" }, "funding": { "url": "https://opencollective.com/vitest" } }, - "node_modules/accepts": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/accepts/-/accepts-2.0.0.tgz", - "integrity": "sha512-5cvg6CtKwfgdmVqY1WIiXKc3Q1bkRqGLi+2W/6ao+6Y7gu/RCwRuAhGEzh5B4KlszSuTLgZYuqFqo5bImjNKng==", + "node_modules/@vitest/snapshot/node_modules/@vitest/pretty-format": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/pretty-format/-/pretty-format-4.1.8.tgz", + "integrity": "sha512-9GasEBxpZ1VYIpqHf/0+YGg121uSNwCKOJqIrTwWP/TB7DmFCiaBpNl3aPZzoLWfWkuqhbH8vJIVobZkvdo2cA==", + "dev": true, "license": "MIT", "dependencies": { - "mime-types": "^3.0.0", - "negotiator": "^1.0.0" + "tinyrainbow": "^3.1.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/snapshot/node_modules/@vitest/utils": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/utils/-/utils-4.1.8.tgz", + "integrity": "sha512-uOJamYALNhfJ6iolExyQM40yIQwDqYnkKtQ5VCiSe17E33H0aQ/u+1GlRuz4LZBk6Mm3sg90G9hEbmEt37C1Zg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/pretty-format": "4.1.8", + "convert-source-map": "^2.0.0", + "tinyrainbow": "^3.1.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/snapshot/node_modules/tinyrainbow": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/tinyrainbow/-/tinyrainbow-3.1.0.tgz", + "integrity": "sha512-Bf+ILmBgretUrdJxzXM0SgXLZ3XfiaUuOj/IKQHuTXip+05Xn+uyEYdVg0kYDipTBcLrCVyUzAPz7QmArb0mmw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/@vitest/spy": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/@vitest/spy/-/spy-3.2.4.tgz", + "integrity": "sha512-vAfasCOe6AIK70iP5UD11Ac4siNUNJ9i/9PZ3NKx07sG6sUxeag1LWdNrMWeKKYBLlzuK+Gn65Yd5nyL6ds+nw==", + "dev": true, + "license": "MIT", + "dependencies": { + "tinyspy": "^4.0.3" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/utils": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/@vitest/utils/-/utils-3.2.4.tgz", + "integrity": "sha512-fB2V0JFrQSMsCo9HiSq3Ezpdv4iYaXRG1Sx8edX3MwxfyNn83mKiGzOcH+Fkxt4MHxr3y42fQi1oeAInqgX2QA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/pretty-format": "3.2.4", + "loupe": "^3.1.4", + "tinyrainbow": "^2.0.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/accepts": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/accepts/-/accepts-2.0.0.tgz", + "integrity": "sha512-5cvg6CtKwfgdmVqY1WIiXKc3Q1bkRqGLi+2W/6ao+6Y7gu/RCwRuAhGEzh5B4KlszSuTLgZYuqFqo5bImjNKng==", + "dev": true, + "license": "MIT", + "dependencies": { + "mime-types": "^3.0.0", + "negotiator": "^1.0.0" }, "engines": { "node": ">= 0.6" @@ -3754,6 +3987,7 @@ "version": "8.15.0", "resolved": "https://registry.npmjs.org/acorn/-/acorn-8.15.0.tgz", "integrity": "sha512-NZyJarBfL7nWwIq+FDL6Zp/yHEhePMNnnJ0y3qfieCrmNvYct8uvtiV41UvlSe6apAfk0fY1FbWx+NwfmpvtTg==", + "dev": true, "license": "MIT", "bin": { "acorn": "bin/acorn" @@ -3793,6 +4027,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/ajv-formats/-/ajv-formats-3.0.1.tgz", "integrity": "sha512-8iUql50EUR+uUcdRQ3HDqa6EVyo3docL8g5WJ3FNcWmu62IbkGUue/pEyLBW8VGKKucTPgqeks4fIU1DA4yowQ==", + "dev": true, "license": "MIT", "dependencies": { "ajv": "^8.0.0" @@ -3810,6 +4045,7 @@ "version": "8.18.0", "resolved": "https://registry.npmjs.org/ajv/-/ajv-8.18.0.tgz", "integrity": "sha512-PlXPeEWMXMZ7sPYOHqmDyCJzcfNrUr3fGNKtezX14ykXOEIvyK81d+qydx89KY5O71FKMPaQ2vBfBFI5NHR63A==", + "dev": true, "license": "MIT", "dependencies": { "fast-deep-equal": "^3.1.3", @@ -3826,6 +4062,7 @@ "version": "1.0.0", "resolved": "https://registry.npmjs.org/json-schema-traverse/-/json-schema-traverse-1.0.0.tgz", "integrity": "sha512-NM8/P9n3XjXhIZn1lLhkFaACTOURQXjWhV4BA/RnOv8xvgqtqpAX9IO4mRQxSx1Rlo4tqzeqb0sOlruaOy3dug==", + "dev": true, "license": "MIT" }, "node_modules/ansi-regex": { @@ -3834,6 +4071,7 @@ "integrity": "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ==", "dev": true, "license": "MIT", + "peer": true, "engines": { "node": ">=8" } @@ -3895,9 +4133,9 @@ } }, "node_modules/ast-v8-to-istanbul": { - "version": "0.3.11", - "resolved": "https://registry.npmjs.org/ast-v8-to-istanbul/-/ast-v8-to-istanbul-0.3.11.tgz", - "integrity": "sha512-Qya9fkoofMjCBNVdWINMjB5KZvkYfaO9/anwkWnjxibpWUxo5iHl2sOdP7/uAqaRuUYuoo8rDwnbaaKVFxoUvw==", + "version": "1.0.3", + "resolved": "https://registry.npmjs.org/ast-v8-to-istanbul/-/ast-v8-to-istanbul-1.0.3.tgz", + "integrity": "sha512-jCMQ6ZylLPudp0CDfBmQBZUsrh1/8psbmu9ibeVWKuHWD0YrH9YABwlKu5kVEFoT0GCQQW9Z/SxfuEbbkGQCRg==", "dev": true, "license": "MIT", "dependencies": { @@ -3906,13 +4144,6 @@ "js-tokens": "^10.0.0" } }, - "node_modules/ast-v8-to-istanbul/node_modules/js-tokens": { - "version": "10.0.0", - "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-10.0.0.tgz", - "integrity": "sha512-lM/UBzQmfJRo9ABXbPWemivdCW8V2G8FHaHdypQaIy523snUjog0W71ayWXTjiR+ixeMyVHN2XcpnTd/liPg/Q==", - "dev": true, - "license": "MIT" - }, "node_modules/async": { "version": "3.2.6", "resolved": "https://registry.npmjs.org/async/-/async-3.2.6.tgz", @@ -3934,6 +4165,7 @@ "version": "4.1.0", "resolved": "https://registry.npmjs.org/axobject-query/-/axobject-query-4.1.0.tgz", "integrity": "sha512-qIj0G9wZbMGNLjLmg1PT6v2mE9AH2zlnADJD/2tC6E00hgmhUOfEB6greHPAfLRSufHqROIUTkw6E+M3lH0PTQ==", + "dev": true, "license": "Apache-2.0", "engines": { "node": ">= 0.4" @@ -3943,6 +4175,7 @@ "version": "2.0.2", "resolved": "https://registry.npmjs.org/bail/-/bail-2.0.2.tgz", "integrity": "sha512-0xO6mYd7JB2YesxDKplafRpsiOzPt9V02ddPCLbY1xYGPOX24NTyN50qnUxgCPcSoYMhKpAuBTjQoRZCAkUDRw==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -4044,6 +4277,7 @@ "version": "2.2.2", "resolved": "https://registry.npmjs.org/body-parser/-/body-parser-2.2.2.tgz", "integrity": "sha512-oP5VkATKlNwcgvxi0vM0p/D3n2C3EReYVX+DNYs5TjZFn/oQt2j+4sVJtSMr18pdRr8wjTcBl6LoV+FUwzPmNA==", + "dev": true, "license": "MIT", "dependencies": { "bytes": "^3.1.2", @@ -4068,6 +4302,7 @@ "version": "0.7.2", "resolved": "https://registry.npmjs.org/iconv-lite/-/iconv-lite-0.7.2.tgz", "integrity": "sha512-im9DjEDQ55s9fL4EYzOAv0yMqmMBSZp6G0VvFyTMPKWxiSBHUj9NW/qqLmXUwXrrM7AvqSlTCfvqRb0cM8yYqw==", + "dev": true, "license": "MIT", "dependencies": { "safer-buffer": ">= 2.1.2 < 3.0.0" @@ -4125,25 +4360,17 @@ "version": "3.1.2", "resolved": "https://registry.npmjs.org/bytes/-/bytes-3.1.2.tgz", "integrity": "sha512-/Nf7TyzTx6S3yRJObOAV7956r8cr2+Oj8AC5dt8wSP3BQAoeX58NoHyCU8P8zGkNXStjTSi6fzO6F0pBdcYbEg==", - "license": "MIT", - "engines": { - "node": ">= 0.8" - } - }, - "node_modules/cac": { - "version": "6.7.14", - "resolved": "https://registry.npmjs.org/cac/-/cac-6.7.14.tgz", - "integrity": "sha512-b6Ilus+c3RrdDk+JhLKUAQfzzgLEPy6wcXqS7f/xe1EETvsDP6GORG7SFuOs6cID5YkqchW/LXZbX5bc8j7ZcQ==", "dev": true, "license": "MIT", "engines": { - "node": ">=8" + "node": ">= 0.8" } }, "node_modules/call-bind-apply-helpers": { "version": "1.0.2", "resolved": "https://registry.npmjs.org/call-bind-apply-helpers/-/call-bind-apply-helpers-1.0.2.tgz", "integrity": "sha512-Sp1ablJ0ivDkSzjcaJdxEunN5/XvksFJ2sMBFfq6x0ryhQV/2b/KwFe21cMpmHtPOSij8K99/wSfoEuTObmuMQ==", + "dev": true, "license": "MIT", "dependencies": { "es-errors": "^1.3.0", @@ -4157,6 +4384,7 @@ "version": "1.0.4", "resolved": "https://registry.npmjs.org/call-bound/-/call-bound-1.0.4.tgz", "integrity": "sha512-+ys997U96po4Kx/ABpBCqhA9EuxJaQWDQg7295H4hBphv3IZg0boBKuwYpt4YXp6MZ5AmZQnU/tyMTlRpaSejg==", + "dev": true, "license": "MIT", "dependencies": { "call-bind-apply-helpers": "^1.0.2", @@ -4183,6 +4411,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/ccount/-/ccount-2.0.1.tgz", "integrity": "sha512-eyrF0jiFpY+3drT6383f1qhkbGsLSifNAjA61IUjZjmLCWjItY6LB9ft9YhoDgwfmclB2zhu51Lc7+95b8NRAg==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -4227,6 +4456,7 @@ "version": "2.0.2", "resolved": "https://registry.npmjs.org/character-entities/-/character-entities-2.0.2.tgz", "integrity": "sha512-shx7oQ0Awen/BRIdkjkvz54PnEEI/EjwXDSIZp86/KKdbafHh1Df/RYGBhn4hbe2+uKC9FnT5UCEdyPz3ai9hQ==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -4237,6 +4467,7 @@ "version": "2.1.0", "resolved": "https://registry.npmjs.org/character-entities-html4/-/character-entities-html4-2.1.0.tgz", "integrity": "sha512-1v7fgQRj6hnSwFpq1Eu0ynr/CDEw0rXo2B61qXrLNdHZmPKgb7fqS1a2JwF0rISo9q77jDI8VMEHoApn8qDoZA==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -4247,6 +4478,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/character-entities-legacy/-/character-entities-legacy-3.0.0.tgz", "integrity": "sha512-RpPp0asT/6ufRm//AJVwpViZbGM/MkjQFxJccQRHmISF/22NBtsHqAWmL+/pmkPWoIUJdWyeVleTl1wydHATVQ==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -4317,6 +4549,7 @@ "version": "2.1.1", "resolved": "https://registry.npmjs.org/clsx/-/clsx-2.1.1.tgz", "integrity": "sha512-eYm0QWBtUrBWZWG0d386OGAw16Z995PiOVo2B7bjWSbHedGl5e0ZWaq65kOGgUSNesEIDkB9ISbTg/JK9dhCZA==", + "dev": true, "license": "MIT", "engines": { "node": ">=6" @@ -4346,6 +4579,7 @@ "version": "2.0.3", "resolved": "https://registry.npmjs.org/comma-separated-tokens/-/comma-separated-tokens-2.0.3.tgz", "integrity": "sha512-Fu4hJdvzeylCfQPp9SGWidpzrMs7tTrlu6Vb8XGaRGck8QSNZJJp538Wrb60Lax4fPwR64ViY468OIUTbRlGZg==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -4356,6 +4590,7 @@ "version": "8.3.0", "resolved": "https://registry.npmjs.org/commander/-/commander-8.3.0.tgz", "integrity": "sha512-OkTL9umf+He2DZkUq8f8J9of7yL6RJKI24dVITBmNfZBmri9zYZQrKkuXiKhyfPSu8tUhnVBB1iKXevvnlR4Ww==", + "dev": true, "license": "MIT", "engines": { "node": ">= 12" @@ -4372,6 +4607,7 @@ "version": "1.0.1", "resolved": "https://registry.npmjs.org/content-disposition/-/content-disposition-1.0.1.tgz", "integrity": "sha512-oIXISMynqSqm241k6kcQ5UwttDILMK4BiurCfGEREw6+X9jkkpEe5T9FZaApyLGGOnFuyMWZpdolTXMtvEJ08Q==", + "dev": true, "license": "MIT", "engines": { "node": ">=18" @@ -4385,25 +4621,38 @@ "version": "1.0.5", "resolved": "https://registry.npmjs.org/content-type/-/content-type-1.0.5.tgz", "integrity": "sha512-nTjqfcBFEipKdXCv4YDQWCfmcLZKm81ldF0pAopTvyrFGVbcR6P/VAAd5G7N+0tTr8QqiU0tFadD6FK4NtJwOA==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.6" } }, + "node_modules/convert-source-map": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/convert-source-map/-/convert-source-map-2.0.0.tgz", + "integrity": "sha512-Kvp459HrV2FEJ1CAsi1Ku+MY3kasH19TFykTz2xWmMeq6bk2NU3XXvfJ+Q61m0xktWwt+1HSYf3JZsTms3aRJg==", + "dev": true, + "license": "MIT" + }, "node_modules/cookie": { - "version": "0.6.0", - "resolved": "https://registry.npmjs.org/cookie/-/cookie-0.6.0.tgz", - "integrity": "sha512-U71cyTamuh1CRNCfpGY6to28lxvNwPG4Guz/EVjgf3Jmzv0vlDp1atT9eS5dDjMYHucpHbWns6Lwf3BKz6svdw==", + "version": "1.1.1", + "resolved": "https://registry.npmjs.org/cookie/-/cookie-1.1.1.tgz", + "integrity": "sha512-ei8Aos7ja0weRpFzJnEA9UHJ/7XQmqglbRwnf2ATjcB9Wq874VKH9kfjjirM6UhU2/E5fFYadylyhFldcqSidQ==", "dev": true, "license": "MIT", "engines": { - "node": ">= 0.6" + "node": ">=18" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/express" } }, "node_modules/cookie-signature": { "version": "1.2.2", "resolved": "https://registry.npmjs.org/cookie-signature/-/cookie-signature-1.2.2.tgz", "integrity": "sha512-D76uU73ulSXrD1UXF4KE2TMxVVwhsnCgfAyTg9k8P6KGZjlXKrOLe4dJQKI3Bxi5wjesZoFXJWElNWBjPZMbhg==", + "dev": true, "license": "MIT", "engines": { "node": ">=6.6.0" @@ -4413,6 +4662,7 @@ "version": "2.8.5", "resolved": "https://registry.npmjs.org/cors/-/cors-2.8.5.tgz", "integrity": "sha512-KIHbLJqu73RGr/hnbrO9uBeixNGuvSQjul/jdFvS/KFSIH1hWVd1ng7zOHx+YrEfInLG7q4n6GHQ9cDtxv/P6g==", + "dev": true, "license": "MIT", "dependencies": { "object-assign": "^4", @@ -4436,6 +4686,7 @@ "version": "1.0.3", "resolved": "https://registry.npmjs.org/cose-base/-/cose-base-1.0.3.tgz", "integrity": "sha512-s9whTXInMSgAp/NVXVNuVxVKzGH2qck3aQlVHxDCdAEPgtMKwc4Wq6/QKhgdEdgbLSi9rBTAcPoRa6JpiG4ksg==", + "dev": true, "license": "MIT", "dependencies": { "layout-base": "^1.0.0" @@ -4445,6 +4696,7 @@ "version": "7.0.6", "resolved": "https://registry.npmjs.org/cross-spawn/-/cross-spawn-7.0.6.tgz", "integrity": "sha512-uV2QOWP2nWzsy2aMp8aRibhi9dlzF5Hgh5SHaB9OiTGEyDTiJJyx0uy51QXdyWbtAHNua4XJzUKca3OzKUd3vA==", + "dev": true, "license": "MIT", "dependencies": { "path-key": "^3.1.0", @@ -4476,17 +4728,18 @@ } }, "node_modules/csstype": { - "version": "3.1.3", - "resolved": "https://registry.npmjs.org/csstype/-/csstype-3.1.3.tgz", - "integrity": "sha512-M1uQkMl8rQK/szD0LNhtqxIPLpimGm8sOBwU7lLnCpSbTyY3yeU1Vc7l4KT5zT4s/yOxHH5O7tIuuLOCnLADRw==", + "version": "3.2.3", + "resolved": "https://registry.npmjs.org/csstype/-/csstype-3.2.3.tgz", + "integrity": "sha512-z1HGKcYy2xA8AGQfwrn0PAy+PB7X/GSj3UVJW9qKyn43xWa+gl5nXmU4qqLMRzWVLFC8KusUX8T/0kCiOYpAIQ==", "dev": true, "license": "MIT", "peer": true }, "node_modules/cytoscape": { - "version": "3.33.4", - "resolved": "https://registry.npmjs.org/cytoscape/-/cytoscape-3.33.4.tgz", - "integrity": "sha512-HIN5Pmd9MrX9BkV7tDwnOcEJCSFvCpc8X97h3f508J6I5FsqAY65wKOCvgH2CuP42CaahWaz4tuh32SOOIH7ww==", + "version": "3.34.0", + "resolved": "https://registry.npmjs.org/cytoscape/-/cytoscape-3.34.0.tgz", + "integrity": "sha512-62rNSrioXw93uliKFBwjukeQyeWwH2PqDrTac31r2P6464u3AUvTk0xS4LVvT251g7IgkFunrI48ZEZGjywSOg==", + "dev": true, "license": "MIT", "engines": { "node": ">=0.10" @@ -4496,6 +4749,7 @@ "version": "4.1.0", "resolved": "https://registry.npmjs.org/cytoscape-cose-bilkent/-/cytoscape-cose-bilkent-4.1.0.tgz", "integrity": "sha512-wgQlVIUJF13Quxiv5e1gstZ08rnZj2XaLHGoFMYXz7SkNfCDOOteKBE6SYRfA9WxxI/iBc3ajfDoc6hb/MRAHQ==", + "dev": true, "license": "MIT", "dependencies": { "cose-base": "^1.0.0" @@ -4508,6 +4762,7 @@ "version": "2.2.0", "resolved": "https://registry.npmjs.org/cytoscape-fcose/-/cytoscape-fcose-2.2.0.tgz", "integrity": "sha512-ki1/VuRIHFCzxWNrsshHYPs6L7TvLu3DL+TyIGEsRcvVERmxokbf5Gdk7mFxZnTdiGtnA4cfSmjZJMviqSuZrQ==", + "dev": true, "license": "MIT", "dependencies": { "cose-base": "^2.2.0" @@ -4520,6 +4775,7 @@ "version": "2.2.0", "resolved": "https://registry.npmjs.org/cose-base/-/cose-base-2.2.0.tgz", "integrity": "sha512-AzlgcsCbUMymkADOJtQm3wO9S3ltPfYOFD5033keQn9NJzIbtnZj+UdBJe7DYml/8TdbtHJW3j58SOnKhWY/5g==", + "dev": true, "license": "MIT", "dependencies": { "layout-base": "^2.0.0" @@ -4529,12 +4785,14 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/layout-base/-/layout-base-2.0.1.tgz", "integrity": "sha512-dp3s92+uNI1hWIpPGH3jK2kxE2lMjdXdr+DH8ynZHpd6PUlH6x6cbuXnoMmiNumznqaNO31xu9e79F0uuZ0JFg==", + "dev": true, "license": "MIT" }, "node_modules/d3": { "version": "7.9.0", "resolved": "https://registry.npmjs.org/d3/-/d3-7.9.0.tgz", "integrity": "sha512-e1U46jVP+w7Iut8Jt8ri1YsPOvFpg46k+K8TpCb0P+zjCkjkPnV7WzfDJzMHy1LnA+wj5pLT1wjO901gLXeEhA==", + "dev": true, "license": "ISC", "dependencies": { "d3-array": "3", @@ -4576,6 +4834,7 @@ "version": "3.2.4", "resolved": "https://registry.npmjs.org/d3-array/-/d3-array-3.2.4.tgz", "integrity": "sha512-tdQAmyA18i4J7wprpYq8ClcxZy3SC31QMeByyCFyRt7BVHdREQZ5lpzoe5mFEYZUWe+oq8HBvk9JjpibyEV4Jg==", + "dev": true, "license": "ISC", "dependencies": { "internmap": "1 - 2" @@ -4588,6 +4847,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/d3-axis/-/d3-axis-3.0.0.tgz", "integrity": "sha512-IH5tgjV4jE/GhHkRV0HiVYPDtvfjHQlQfJHs0usq7M30XcSBvOotpmH1IgkcXsO/5gEQZD43B//fc7SRT5S+xw==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4597,6 +4857,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/d3-brush/-/d3-brush-3.0.0.tgz", "integrity": "sha512-ALnjWlVYkXsVIGlOsuWH1+3udkYFI48Ljihfnh8FZPF2QS9o+PzGLBslO0PjzVoHLZ2KCVgAM8NVkXPJB2aNnQ==", + "dev": true, "license": "ISC", "dependencies": { "d3-dispatch": "1 - 3", @@ -4613,6 +4874,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-chord/-/d3-chord-3.0.1.tgz", "integrity": "sha512-VE5S6TNa+j8msksl7HwjxMHDM2yNK3XCkusIlpX5kwauBfXuyLAtNg9jCp/iHH61tgI4sb6R/EIMWCqEIdjT/g==", + "dev": true, "license": "ISC", "dependencies": { "d3-path": "1 - 3" @@ -4625,6 +4887,7 @@ "version": "3.1.0", "resolved": "https://registry.npmjs.org/d3-color/-/d3-color-3.1.0.tgz", "integrity": "sha512-zg/chbXyeBtMQ1LbD/WSoW2DpC3I0mpmPdW+ynRTj/x2DAWYrIY7qeZIHidozwV24m4iavr15lNwIwLxRmOxhA==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4634,6 +4897,7 @@ "version": "4.0.2", "resolved": "https://registry.npmjs.org/d3-contour/-/d3-contour-4.0.2.tgz", "integrity": "sha512-4EzFTRIikzs47RGmdxbeUvLWtGedDUNkTcmzoeyg4sP/dvCexO47AaQL7VKy/gul85TOxw+IBgA8US2xwbToNA==", + "dev": true, "license": "ISC", "dependencies": { "d3-array": "^3.2.0" @@ -4646,6 +4910,7 @@ "version": "6.0.4", "resolved": "https://registry.npmjs.org/d3-delaunay/-/d3-delaunay-6.0.4.tgz", "integrity": "sha512-mdjtIZ1XLAM8bm/hx3WwjfHt6Sggek7qH043O8KEjDXN40xi3vx/6pYSVTwLjEgiXQTbvaouWKynLBiUZ6SK6A==", + "dev": true, "license": "ISC", "dependencies": { "delaunator": "5" @@ -4658,6 +4923,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-dispatch/-/d3-dispatch-3.0.1.tgz", "integrity": "sha512-rzUyPU/S7rwUflMyLc1ETDeBj0NRuHKKAcvukozwhshr6g6c5d8zh4c2gQjY2bZ0dXeGLWc1PF174P2tVvKhfg==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4667,6 +4933,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/d3-drag/-/d3-drag-3.0.0.tgz", "integrity": "sha512-pWbUJLdETVA8lQNJecMxoXfH6x+mO2UQo8rSmZ+QqxcbyA3hfeprFgIT//HW2nlHChWeIIMwS2Fq+gEARkhTkg==", + "dev": true, "license": "ISC", "dependencies": { "d3-dispatch": "1 - 3", @@ -4680,6 +4947,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-dsv/-/d3-dsv-3.0.1.tgz", "integrity": "sha512-UG6OvdI5afDIFP9w4G0mNq50dSOsXHJaRE8arAS5o9ApWnIElp8GZw1Dun8vP8OyHOZ/QJUKUJwxiiCCnUwm+Q==", + "dev": true, "license": "ISC", "dependencies": { "commander": "7", @@ -4705,6 +4973,7 @@ "version": "7.2.0", "resolved": "https://registry.npmjs.org/commander/-/commander-7.2.0.tgz", "integrity": "sha512-QrWXB+ZQSVPmIWIhtEO9H+gwHaMGYiF5ChvoJ+K9ZGHG/sVsa6yiesAD1GC/x46sET00Xlwo1u49RVVVzvcSkw==", + "dev": true, "license": "MIT", "engines": { "node": ">= 10" @@ -4714,6 +4983,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-ease/-/d3-ease-3.0.1.tgz", "integrity": "sha512-wR/XK3D3XcLIZwpbvQwQ5fK+8Ykds1ip7A2Txe0yxncXSdq1L9skcG7blcedkOX+ZcgxGAmLX1FrRGbADwzi0w==", + "dev": true, "license": "BSD-3-Clause", "engines": { "node": ">=12" @@ -4723,6 +4993,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-fetch/-/d3-fetch-3.0.1.tgz", "integrity": "sha512-kpkQIM20n3oLVBKGg6oHrUchHM3xODkTzjMoj7aWQFq5QEM+R6E4WkzT5+tojDY7yjez8KgCBRoj4aEr99Fdqw==", + "dev": true, "license": "ISC", "dependencies": { "d3-dsv": "1 - 3" @@ -4735,6 +5006,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/d3-force/-/d3-force-3.0.0.tgz", "integrity": "sha512-zxV/SsA+U4yte8051P4ECydjD/S+qeYtnaIyAs9tgHCqfguma/aAQDjo85A9Z6EKhBirHRJHXIgJUlffT4wdLg==", + "dev": true, "license": "ISC", "dependencies": { "d3-dispatch": "1 - 3", @@ -4749,6 +5021,7 @@ "version": "3.1.2", "resolved": "https://registry.npmjs.org/d3-format/-/d3-format-3.1.2.tgz", "integrity": "sha512-AJDdYOdnyRDV5b6ArilzCPPwc1ejkHcoyFarqlPqT7zRYjhavcT3uSrqcMvsgh2CgoPbK3RCwyHaVyxYcP2Arg==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4758,6 +5031,7 @@ "version": "3.1.1", "resolved": "https://registry.npmjs.org/d3-geo/-/d3-geo-3.1.1.tgz", "integrity": "sha512-637ln3gXKXOwhalDzinUgY83KzNWZRKbYubaG+fGVuc/dxO64RRljtCTnf5ecMyE1RIdtqpkVcq0IbtU2S8j2Q==", + "dev": true, "license": "ISC", "dependencies": { "d3-array": "2.5.0 - 3" @@ -4770,6 +5044,7 @@ "version": "3.1.2", "resolved": "https://registry.npmjs.org/d3-hierarchy/-/d3-hierarchy-3.1.2.tgz", "integrity": "sha512-FX/9frcub54beBdugHjDCdikxThEqjnR93Qt7PvQTOHxyiNCAlvMrHhclk3cD5VeAaq9fxmfRp+CnWw9rEMBuA==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4779,6 +5054,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-interpolate/-/d3-interpolate-3.0.1.tgz", "integrity": "sha512-3bYs1rOD33uo8aqJfKP3JWPAibgw8Zm2+L9vBKEHJ2Rg+viTR7o5Mmv5mZcieN+FRYaAOWX5SJATX6k1PWz72g==", + "dev": true, "license": "ISC", "dependencies": { "d3-color": "1 - 3" @@ -4791,6 +5067,7 @@ "version": "3.1.0", "resolved": "https://registry.npmjs.org/d3-path/-/d3-path-3.1.0.tgz", "integrity": "sha512-p3KP5HCf/bvjBSSKuXid6Zqijx7wIfNW+J/maPs+iwR35at5JCbLUT0LzF1cnjbCHWhqzQTIN2Jpe8pRebIEFQ==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4800,6 +5077,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-polygon/-/d3-polygon-3.0.1.tgz", "integrity": "sha512-3vbA7vXYwfe1SYhED++fPUQlWSYTTGmFmQiany/gdbiWgU/iEyQzyymwL9SkJjFFuCS4902BSzewVGsHHmHtXg==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4809,6 +5087,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-quadtree/-/d3-quadtree-3.0.1.tgz", "integrity": "sha512-04xDrxQTDTCFwP5H6hRhsRcb9xxv2RzkcsygFzmkSIOJy3PeRJP7sNk3VRIbKXcog561P9oU0/rVH6vDROAgUw==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4818,6 +5097,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-random/-/d3-random-3.0.1.tgz", "integrity": "sha512-FXMe9GfxTxqd5D6jFsQ+DJ8BJS4E/fT5mqqdjovykEB2oFbTMDVdg1MGFxfQW+FBOGoB++k8swBrgwSHT1cUXQ==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4827,6 +5107,7 @@ "version": "0.12.3", "resolved": "https://registry.npmjs.org/d3-sankey/-/d3-sankey-0.12.3.tgz", "integrity": "sha512-nQhsBRmM19Ax5xEIPLMY9ZmJ/cDvd1BG3UVvt5h3WRxKg5zGRbvnteTyWAbzeSvlh3tW7ZEmq4VwR5mB3tutmQ==", + "dev": true, "license": "BSD-3-Clause", "dependencies": { "d3-array": "1 - 2", @@ -4837,6 +5118,7 @@ "version": "2.12.1", "resolved": "https://registry.npmjs.org/d3-array/-/d3-array-2.12.1.tgz", "integrity": "sha512-B0ErZK/66mHtEsR1TkPEEkwdy+WDesimkM5gpZr5Dsg54BiTA5RXtYW5qTLIAcekaS9xfZrzBLF/OAkB3Qn1YQ==", + "dev": true, "license": "BSD-3-Clause", "dependencies": { "internmap": "^1.0.0" @@ -4846,12 +5128,14 @@ "version": "1.0.9", "resolved": "https://registry.npmjs.org/d3-path/-/d3-path-1.0.9.tgz", "integrity": "sha512-VLaYcn81dtHVTjEHd8B+pbe9yHWpXKZUC87PzoFmsFrJqgFwDe/qxfp5MlfsfM1V5E/iVt0MmEbWQ7FVIXh/bg==", + "dev": true, "license": "BSD-3-Clause" }, "node_modules/d3-sankey/node_modules/d3-shape": { "version": "1.3.7", "resolved": "https://registry.npmjs.org/d3-shape/-/d3-shape-1.3.7.tgz", "integrity": "sha512-EUkvKjqPFUAZyOlhY5gzCxCeI0Aep04LwIRpsZ/mLFelJiUfnK56jo5JMDSE7yyP2kLSb6LtF+S5chMk7uqPqw==", + "dev": true, "license": "BSD-3-Clause", "dependencies": { "d3-path": "1" @@ -4861,12 +5145,14 @@ "version": "1.0.1", "resolved": "https://registry.npmjs.org/internmap/-/internmap-1.0.1.tgz", "integrity": "sha512-lDB5YccMydFBtasVtxnZ3MRBHuaoE8GKsppq+EchKL2U4nK/DmEpPHNH8MZe5HkMtpSiTSOZwfN0tzYjO/lJEw==", + "dev": true, "license": "ISC" }, "node_modules/d3-scale": { "version": "4.0.2", "resolved": "https://registry.npmjs.org/d3-scale/-/d3-scale-4.0.2.tgz", "integrity": "sha512-GZW464g1SH7ag3Y7hXjf8RoUuAFIqklOAq3MRl4OaWabTFJY9PN/E1YklhXLh+OQ3fM9yS2nOkCoS+WLZ6kvxQ==", + "dev": true, "license": "ISC", "dependencies": { "d3-array": "2.10.0 - 3", @@ -4883,6 +5169,7 @@ "version": "3.1.0", "resolved": "https://registry.npmjs.org/d3-scale-chromatic/-/d3-scale-chromatic-3.1.0.tgz", "integrity": "sha512-A3s5PWiZ9YCXFye1o246KoscMWqf8BsD9eRiJ3He7C9OBaxKhAd5TFCdEx/7VbKtxxTsu//1mMJFrEt572cEyQ==", + "dev": true, "license": "ISC", "dependencies": { "d3-color": "1 - 3", @@ -4896,6 +5183,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/d3-selection/-/d3-selection-3.0.0.tgz", "integrity": "sha512-fmTRWbNMmsmWq6xJV8D19U/gw/bwrHfNXxrIN+HfZgnzqTHp9jOmKMhsTUjXOJnZOdZY9Q28y4yebKzqDKlxlQ==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4905,6 +5193,7 @@ "version": "3.2.0", "resolved": "https://registry.npmjs.org/d3-shape/-/d3-shape-3.2.0.tgz", "integrity": "sha512-SaLBuwGm3MOViRq2ABk3eLoxwZELpH6zhl3FbAoJ7Vm1gofKx6El1Ib5z23NUEhF9AsGl7y+dzLe5Cw2AArGTA==", + "dev": true, "license": "ISC", "dependencies": { "d3-path": "^3.1.0" @@ -4917,6 +5206,7 @@ "version": "3.1.0", "resolved": "https://registry.npmjs.org/d3-time/-/d3-time-3.1.0.tgz", "integrity": "sha512-VqKjzBLejbSMT4IgbmVgDjpkYrNWUYJnbCGo874u7MMKIWsILRX+OpX/gTk8MqjpT1A/c6HY2dCA77ZN0lkQ2Q==", + "dev": true, "license": "ISC", "dependencies": { "d3-array": "2 - 3" @@ -4929,6 +5219,7 @@ "version": "4.1.0", "resolved": "https://registry.npmjs.org/d3-time-format/-/d3-time-format-4.1.0.tgz", "integrity": "sha512-dJxPBlzC7NugB2PDLwo9Q8JiTR3M3e4/XANkreKSUxF8vvXKqm1Yfq4Q5dl8budlunRVlUUaDUgFt7eA8D6NLg==", + "dev": true, "license": "ISC", "dependencies": { "d3-time": "1 - 3" @@ -4941,6 +5232,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-timer/-/d3-timer-3.0.1.tgz", "integrity": "sha512-ndfJ/JxxMd3nw31uyKoY2naivF+r29V+Lc0svZxe1JvvIRmi8hUsrMvdOwgS1o6uBHmiz91geQ0ylPP0aj1VUA==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -4950,6 +5242,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/d3-transition/-/d3-transition-3.0.1.tgz", "integrity": "sha512-ApKvfjsSR6tg06xrL434C0WydLr7JewBB3V+/39RMHsaXTOG0zmt/OAXeng5M5LBm0ojmxJrpomQVZ1aPvBL4w==", + "dev": true, "license": "ISC", "dependencies": { "d3-color": "1 - 3", @@ -4969,6 +5262,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/d3-zoom/-/d3-zoom-3.0.0.tgz", "integrity": "sha512-b8AmV3kfQaqWAuacbPuNbL6vahnOJflOhexLzMMNLga62+/nh0JzvJ0aO/5a5MVgUFGS7Hu1P9P03o3fJkDCyw==", + "dev": true, "license": "ISC", "dependencies": { "d3-dispatch": "1 - 3", @@ -4985,6 +5279,7 @@ "version": "7.0.14", "resolved": "https://registry.npmjs.org/dagre-d3-es/-/dagre-d3-es-7.0.14.tgz", "integrity": "sha512-P4rFMVq9ESWqmOgK+dlXvOtLwYg0i7u0HBGJER0LZDJT2VHIPAMZ/riPxqJceWMStH5+E61QxFra9kIS3AqdMg==", + "dev": true, "license": "MIT", "dependencies": { "d3": "^7.9.0", @@ -4992,15 +5287,17 @@ } }, "node_modules/dayjs": { - "version": "1.11.20", - "resolved": "https://registry.npmjs.org/dayjs/-/dayjs-1.11.20.tgz", - "integrity": "sha512-YbwwqR/uYpeoP4pu043q+LTDLFBLApUP6VxRihdfNTqu4ubqMlGDLd6ErXhEgsyvY0K6nCs7nggYumAN+9uEuQ==", + "version": "1.11.21", + "resolved": "https://registry.npmjs.org/dayjs/-/dayjs-1.11.21.tgz", + "integrity": "sha512-98IT+HOahAisibz/yjKbzuOBwYcjJ7BCLPzARyHiyEBmRz4fatF+KPJszEHXsGYjUG234aH/cOjW1wwTbKUZlA==", + "dev": true, "license": "MIT" }, "node_modules/debug": { "version": "4.4.3", "resolved": "https://registry.npmjs.org/debug/-/debug-4.4.3.tgz", "integrity": "sha512-RGwwWnwQvkVfavKVt22FGLw+xYSdzARwm0ru6DhTVA3umU5hZc28V3kO4stgYryrTlLpuvgI9GiijltAjNbcqA==", + "dev": true, "license": "MIT", "dependencies": { "ms": "^2.1.3" @@ -5018,6 +5315,7 @@ "version": "1.2.0", "resolved": "https://registry.npmjs.org/decode-named-character-reference/-/decode-named-character-reference-1.2.0.tgz", "integrity": "sha512-c6fcElNV6ShtZXmsgNgFFV5tVX2PaV4g+MOAkb8eXHvn6sryJBrZa9r0zV6+dtTyoCKxtDy5tyQ5ZwQuidtd+Q==", + "dev": true, "license": "MIT", "dependencies": { "character-entities": "^2.0.0" @@ -5123,6 +5421,7 @@ "version": "5.1.0", "resolved": "https://registry.npmjs.org/delaunator/-/delaunator-5.1.0.tgz", "integrity": "sha512-AGrQ4QSgssa1NGmWmLPqN5NY2KajF5MqxetNEO+o0n3ZwZZeTmt7bBnvzHWrmkZFxGgr4HdyFgelzgi06otLuQ==", + "dev": true, "license": "ISC", "dependencies": { "robust-predicates": "^3.0.2" @@ -5132,6 +5431,7 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/depd/-/depd-2.0.0.tgz", "integrity": "sha512-g7nH6P6dyDioJogAAGprGpCtVImJhpPk/roCzdb3fIh61/s/nPsfR6onyMwkCAR/OlC3yBC0lESvUoQEAssIrw==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.8" @@ -5141,6 +5441,7 @@ "version": "2.0.3", "resolved": "https://registry.npmjs.org/dequal/-/dequal-2.0.3.tgz", "integrity": "sha512-0je+qPKHEMohvfRTCEo3CrPG6cAzAYgmzKyxRiYSSDkS6eGJdyVJm7WaYA5ECaAD9wLB2T4EEeymA5aFVcYXCA==", + "dev": true, "license": "MIT", "engines": { "node": ">=6" @@ -5160,12 +5461,14 @@ "version": "5.8.1", "resolved": "https://registry.npmjs.org/devalue/-/devalue-5.8.1.tgz", "integrity": "sha512-4CXDYRBGqN+57wVJkuXBYmpAVUSg3L6JAQa/DFqm238G73E1wuyc/JhGQJzN7vUf/CMphYau2zXbfWzDR5aTEw==", + "dev": true, "license": "MIT" }, "node_modules/devlop": { "version": "1.1.0", "resolved": "https://registry.npmjs.org/devlop/-/devlop-1.1.0.tgz", "integrity": "sha512-RWmIqhcFf1lRYBvNmr7qTNuyCt/7/ns2jbpp1+PalgE/rDQcBT0fioSMUpJ93irlUhC5hrg4cYqe6U+0ImW0rA==", + "dev": true, "license": "MIT", "dependencies": { "dequal": "^2.0.0" @@ -5187,12 +5490,14 @@ "resolved": "https://registry.npmjs.org/dom-accessibility-api/-/dom-accessibility-api-0.5.16.tgz", "integrity": "sha512-X7BJ2yElsnOJ30pZF4uIIDfBEVgF4XEBxL9Bxhy6dnrm5hkzqmsWHGTiHqRiITNhMyFLyAiWndIJP7Z1NTteDg==", "dev": true, - "license": "MIT" + "license": "MIT", + "peer": true }, "node_modules/dompurify": { - "version": "3.4.5", - "resolved": "https://registry.npmjs.org/dompurify/-/dompurify-3.4.5.tgz", - "integrity": "sha512-OrwIBKsdNSVEeubdJ1HBv/wNENRM9ytAVCv7YXt//A3vPdVMNuACRqK9mXCGCBW2ln7BT/A4X0jXHo2Gu89miA==", + "version": "3.4.8", + "resolved": "https://registry.npmjs.org/dompurify/-/dompurify-3.4.8.tgz", + "integrity": "sha512-yb1cEmaOum7wFvOCSQxyfgVlv5D47Rc30iZWoMpbDIWTnJ6grDDQyu2KFJzB2k7u0pMuJcQ1zphH//fFnw2tjQ==", + "dev": true, "license": "(MPL-2.0 OR Apache-2.0)", "optionalDependencies": { "@types/trusted-types": "^2.0.7" @@ -5202,6 +5507,7 @@ "version": "1.0.1", "resolved": "https://registry.npmjs.org/dunder-proto/-/dunder-proto-1.0.1.tgz", "integrity": "sha512-KIN/nDJBQRcXw0MLVhZE9iQHmG68qAVIBg9CqmUYjmQIhgij9U5MFvrqkUL5FbtyyzZuOeOt0zdeRe4UY7ct+A==", + "dev": true, "license": "MIT", "dependencies": { "call-bind-apply-helpers": "^1.0.1", @@ -5212,23 +5518,10 @@ "node": ">= 0.4" } }, - "node_modules/eastasianwidth": { - "version": "0.2.0", - "resolved": "https://registry.npmjs.org/eastasianwidth/-/eastasianwidth-0.2.0.tgz", - "integrity": "sha512-I88TYZWc9XiYHRQ4/3c5rjjfgkjhLyW2luGIheGERbNQ6OY7yTybanSpDXZa8y7VUP9YmDcYa+eyq4ca7iLqWA==", - "dev": true, - "license": "MIT" - }, "node_modules/ee-first": { "version": "1.1.1", "resolved": "https://registry.npmjs.org/ee-first/-/ee-first-1.1.1.tgz", "integrity": "sha512-WMwm9LhRUo+WUaRN+vRuETqG89IgZphVSNkdFgeb6sS/E4OrDIN7t48CAewSHXc6C8lefD8KKfr5vY61brQlow==", - "license": "MIT" - }, - "node_modules/emoji-regex": { - "version": "9.2.2", - "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-9.2.2.tgz", - "integrity": "sha512-L18DaJsXSUk2+42pv8mLs5jJT2hqFkFE4j21wOmgbUqsZ2hL72NsUU785g9RXgo3s0ZNgVl42TiHp3ZtOv/Vyg==", "dev": true, "license": "MIT" }, @@ -5236,6 +5529,7 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/encodeurl/-/encodeurl-2.0.0.tgz", "integrity": "sha512-Q0n9HRi4m6JuGIV1eFlmvJB7ZEVxu93IrMyiMsGC0lrMJMWzRgx6WGquyfQgZVb31vhGgXnfmPNNXmxnOkRBrg==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.8" @@ -5272,6 +5566,7 @@ "version": "1.0.1", "resolved": "https://registry.npmjs.org/es-define-property/-/es-define-property-1.0.1.tgz", "integrity": "sha512-e3nRfgfUZ4rNGL232gUgX06QNyyez04KdjFrF+LTRoOXmrOgFKDg4BCdsjW8EnT69eqdYGmRpJwiPVYNrCaW3g==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.4" @@ -5281,15 +5576,16 @@ "version": "1.3.0", "resolved": "https://registry.npmjs.org/es-errors/-/es-errors-1.3.0.tgz", "integrity": "sha512-Zf5H2Kxt2xjTvbJvP2ZWLEICxA6j+hAmMzIlypy4xcBg1vKVnx89Wy0GbS+kf5cwCVFFzdCFh2XSCFNULS6csw==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.4" } }, "node_modules/es-module-lexer": { - "version": "1.7.0", - "resolved": "https://registry.npmjs.org/es-module-lexer/-/es-module-lexer-1.7.0.tgz", - "integrity": "sha512-jEQoCwk8hyb2AZziIOLhDqpm5+2ww5uIE6lkO/6jcOCusfk6LhMHpXXfBLXTZ7Ydyt0j4VoUQv6uGNYbdW+kBA==", + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/es-module-lexer/-/es-module-lexer-2.1.0.tgz", + "integrity": "sha512-n27zTYMjYu1aj4MjCWzSP7G9r75utsaoc8m61weK+W8JMBGGQybd43GstCXZ3WNmSFtGT9wi59qQTW6mhTR5LQ==", "dev": true, "license": "MIT" }, @@ -5297,6 +5593,7 @@ "version": "1.1.1", "resolved": "https://registry.npmjs.org/es-object-atoms/-/es-object-atoms-1.1.1.tgz", "integrity": "sha512-FGgH2h8zKNim9ljj7dankFPcICIK9Cp5bm+c2gQSYePhpaG5+esrLODihIorn+Pe6FGJzWhXQotPv73jTaldXA==", + "dev": true, "license": "MIT", "dependencies": { "es-errors": "^1.3.0" @@ -5309,6 +5606,7 @@ "version": "1.46.1", "resolved": "https://registry.npmjs.org/es-toolkit/-/es-toolkit-1.46.1.tgz", "integrity": "sha512-5eNtXOs3tbfxXOj04tjjseeWkRWaoCjdEI+96DgwzZoe6c9juL49pXlzAFTI72aWC9Y8p7168g6XIKjh7k6pyQ==", + "dev": true, "license": "MIT", "workspaces": [ "docs", @@ -5361,6 +5659,7 @@ "version": "1.0.3", "resolved": "https://registry.npmjs.org/escape-html/-/escape-html-1.0.3.tgz", "integrity": "sha512-NiSupZ4OeuGwr68lGIeym/ksIZMJodUGOSCZ/FSnTxcrekbvqrgdUxlJOMpijaKZVjAJrWrGs/6Jy8OMuyj9ow==", + "dev": true, "license": "MIT" }, "node_modules/escape-string-regexp": { @@ -5534,6 +5833,7 @@ "version": "1.2.2", "resolved": "https://registry.npmjs.org/esm-env/-/esm-env-1.2.2.tgz", "integrity": "sha512-Epxrv+Nr/CaL4ZcFGPJIYLWFom+YeV1DqMLHJoEd9SYRxNbaFruBwfEX/kkHUJf55j2+TUbmDcmuilbP1TmXHA==", + "dev": true, "license": "MIT" }, "node_modules/espree": { @@ -5638,6 +5938,7 @@ "version": "1.8.1", "resolved": "https://registry.npmjs.org/etag/-/etag-1.8.1.tgz", "integrity": "sha512-aIL5Fx7mawVa300al2BnEE4iNvo1qETxLrPI/o05L7z6go7fCw1J6EQmbK4FmJ2AS7kgVF/KEZWufBfdClMcPg==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.6" @@ -5654,6 +5955,7 @@ "version": "3.0.7", "resolved": "https://registry.npmjs.org/eventsource/-/eventsource-3.0.7.tgz", "integrity": "sha512-CRT1WTyuQoD771GW56XEZFQ/ZoSfWid1alKGDYMmkt2yl8UXrVR4pspqWNEcqKvVIzg6PAltWjxcSSPrboA4iA==", + "dev": true, "license": "MIT", "dependencies": { "eventsource-parser": "^3.0.1" @@ -5666,15 +5968,16 @@ "version": "3.0.6", "resolved": "https://registry.npmjs.org/eventsource-parser/-/eventsource-parser-3.0.6.tgz", "integrity": "sha512-Vo1ab+QXPzZ4tCa8SwIHJFaSzy4R6SHf7BY79rFBDf0idraZWAkYrDjDj8uWaSm3S2TK+hJ7/t1CEmZ7jXw+pg==", + "dev": true, "license": "MIT", "engines": { "node": ">=18.0.0" } }, "node_modules/expect-type": { - "version": "1.2.2", - "resolved": "https://registry.npmjs.org/expect-type/-/expect-type-1.2.2.tgz", - "integrity": "sha512-JhFGDVJ7tmDJItKhYgJCGLOWjuK9vPxiXoUFLwLDc99NlmklilbiQJwoctZtt13+xMw91MCk/REan6MWHqDjyA==", + "version": "1.3.0", + "resolved": "https://registry.npmjs.org/expect-type/-/expect-type-1.3.0.tgz", + "integrity": "sha512-knvyeauYhqjOYvQ66MznSMs83wmHrCycNEN6Ao+2AeYEfxUIkuiVxdEa1qlGEPK+We3n0THiDciYSsCcgW/DoA==", "dev": true, "license": "Apache-2.0", "engines": { @@ -5685,6 +5988,7 @@ "version": "5.2.1", "resolved": "https://registry.npmjs.org/express/-/express-5.2.1.tgz", "integrity": "sha512-hIS4idWWai69NezIdRt2xFVofaF4j+6INOpJlVOLDO8zXGpUVEVzIYk12UUi2JzjEzWL3IOAxcTubgz9Po0yXw==", + "dev": true, "license": "MIT", "dependencies": { "accepts": "^2.0.0", @@ -5728,6 +6032,7 @@ "version": "8.5.2", "resolved": "https://registry.npmjs.org/express-rate-limit/-/express-rate-limit-8.5.2.tgz", "integrity": "sha512-5Kb34ipNX694DH48vN9irak1Qx30nb0PLYHXfJgw4YEjiC3ZEmZJhwOp+VfiCYwFzvFTdB9QkArYS5kXa2cx2A==", + "dev": true, "license": "MIT", "dependencies": { "ip-address": "^10.2.0" @@ -5742,25 +6047,18 @@ "express": ">= 4.11" } }, - "node_modules/express/node_modules/cookie": { - "version": "0.7.2", - "resolved": "https://registry.npmjs.org/cookie/-/cookie-0.7.2.tgz", - "integrity": "sha512-yki5XnKuf750l50uGTllt6kKILY4nQ1eNIQatoXEByZ5dWgnKqbnqmTrBE5B4N7lrMJKQ2ytWMiTO2o0v6Ew/w==", - "license": "MIT", - "engines": { - "node": ">= 0.6" - } - }, "node_modules/extend": { "version": "3.0.2", "resolved": "https://registry.npmjs.org/extend/-/extend-3.0.2.tgz", "integrity": "sha512-fjquC59cD7CyW6urNXK0FBufkZcoiGG80wTuPujX590cB5Ttln20E2UB4S/WARVqhXffZl2LNgS+gQdPIIim/g==", + "dev": true, "license": "MIT" }, "node_modules/fast-deep-equal": { "version": "3.1.3", "resolved": "https://registry.npmjs.org/fast-deep-equal/-/fast-deep-equal-3.1.3.tgz", "integrity": "sha512-f3qQ9oQy9j2AhBe/H9VC91wLmKBCCU/gDOnKNAYG5hswO7BLKj09Hc5HYNz9cGI++xlpDCIgDaitVs03ATR84Q==", + "dev": true, "license": "MIT" }, "node_modules/fast-json-stable-stringify": { @@ -5781,6 +6079,7 @@ "version": "3.1.2", "resolved": "https://registry.npmjs.org/fast-uri/-/fast-uri-3.1.2.tgz", "integrity": "sha512-rVjf7ArG3LTk+FS6Yw81V1DLuZl1bRbNrev6Tmd/9RaroeeRRJhAt7jg/6YFxbvAQXUCavSoZhPPj6oOx+5KjQ==", + "dev": true, "funding": [ { "type": "github", @@ -5852,6 +6151,7 @@ "version": "2.1.1", "resolved": "https://registry.npmjs.org/finalhandler/-/finalhandler-2.1.1.tgz", "integrity": "sha512-S8KoZgRZN+a5rNwqTxlZZePjT/4cnm0ROV70LedRHZ0p8u9fRID0hJUZQpkKLzro8LfmC8sx23bY6tVNxv8pQA==", + "dev": true, "license": "MIT", "dependencies": { "debug": "^4.4.0", @@ -5928,27 +6228,11 @@ } } }, - "node_modules/foreground-child": { - "version": "3.3.1", - "resolved": "https://registry.npmjs.org/foreground-child/-/foreground-child-3.3.1.tgz", - "integrity": "sha512-gIXjKqtFuWEgzFRJA9WCQeSJLZDjgJUOMCMzxtvFq/37KojM1BFGufqsCy0r4qSQmYLsZYMeyRqzIWOMup03sw==", - "dev": true, - "license": "ISC", - "dependencies": { - "cross-spawn": "^7.0.6", - "signal-exit": "^4.0.1" - }, - "engines": { - "node": ">=14" - }, - "funding": { - "url": "https://github.com/sponsors/isaacs" - } - }, "node_modules/forwarded": { "version": "0.2.0", "resolved": "https://registry.npmjs.org/forwarded/-/forwarded-0.2.0.tgz", "integrity": "sha512-buRG0fpBtRHSTCOASe6hD258tEubFoRLb4ZNA6NxMVHNw2gOcwHo9wyablzMzOA5z9xA9L1KNjk/Nt6MT9aYow==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.6" @@ -5958,6 +6242,7 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/fresh/-/fresh-2.0.0.tgz", "integrity": "sha512-Rx/WycZ60HOaqLKAi6cHRKKI7zxWbJ31MhntmtwMoaTeF7XFH9hhBp8vITaMidfljRQ6eYWCKkaTK+ykVJHP2A==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.8" @@ -5982,6 +6267,7 @@ "version": "1.1.2", "resolved": "https://registry.npmjs.org/function-bind/-/function-bind-1.1.2.tgz", "integrity": "sha512-7XHNxH7qX9xG5mIwxkhumTox/MIRNcOgDrxWsMt2pAr23WHp6MrRlN7FBSFpCpr+oVO0F744iUgR82nJMfG2SA==", + "dev": true, "license": "MIT", "funding": { "url": "https://github.com/sponsors/ljharb" @@ -5991,6 +6277,7 @@ "version": "1.3.0", "resolved": "https://registry.npmjs.org/get-intrinsic/-/get-intrinsic-1.3.0.tgz", "integrity": "sha512-9fSjSaos/fRIVIp+xSJlE6lfwhES7LNtKaCBIamHsjr2na1BiABJPo0mOjjz8GJDURarmCPGqaiVg5mfjb98CQ==", + "dev": true, "license": "MIT", "dependencies": { "call-bind-apply-helpers": "^1.0.2", @@ -6015,6 +6302,7 @@ "version": "1.0.1", "resolved": "https://registry.npmjs.org/get-proto/-/get-proto-1.0.1.tgz", "integrity": "sha512-sTSfBjoXBp89JvIKIefqw7U2CCebsc74kiY6awiGogKtoSGbgjYE/G/+l9sF3MWFPNc9IcoOC4ODfKHfxFmp0g==", + "dev": true, "license": "MIT", "dependencies": { "dunder-proto": "^1.0.1", @@ -6024,27 +6312,6 @@ "node": ">= 0.4" } }, - "node_modules/glob": { - "version": "10.5.0", - "resolved": "https://registry.npmjs.org/glob/-/glob-10.5.0.tgz", - "integrity": "sha512-DfXN8DfhJ7NH3Oe7cFmu3NCu1wKbkReJ8TorzSAFbSKrlNaQSKfIzqYqVY8zlbs2NLBbWpRiU52GX2PbaBVNkg==", - "dev": true, - "license": "ISC", - "dependencies": { - "foreground-child": "^3.1.0", - "jackspeak": "^3.1.2", - "minimatch": "^9.0.4", - "minipass": "^7.1.2", - "package-json-from-dist": "^1.0.0", - "path-scurry": "^1.11.1" - }, - "bin": { - "glob": "dist/esm/bin.mjs" - }, - "funding": { - "url": "https://github.com/sponsors/isaacs" - } - }, "node_modules/glob-parent": { "version": "6.0.2", "resolved": "https://registry.npmjs.org/glob-parent/-/glob-parent-6.0.2.tgz", @@ -6058,32 +6325,6 @@ "node": ">=10.13.0" } }, - "node_modules/glob/node_modules/brace-expansion": { - "version": "2.0.3", - "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-2.0.3.tgz", - "integrity": "sha512-MCV/fYJEbqx68aE58kv2cA/kiky1G8vux3OR6/jbS+jIMe/6fJWa0DTzJU7dqijOWYwHi1t29FlfYI9uytqlpA==", - "dev": true, - "license": "MIT", - "dependencies": { - "balanced-match": "^1.0.0" - } - }, - "node_modules/glob/node_modules/minimatch": { - "version": "9.0.9", - "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-9.0.9.tgz", - "integrity": "sha512-OBwBN9AL4dqmETlpS2zasx+vTeWclWzkblfZk7KTA5j3jeOONz/tRCnZomUyvNg83wL5Zv9Ss6HMJXAgL8R2Yg==", - "dev": true, - "license": "ISC", - "dependencies": { - "brace-expansion": "^2.0.2" - }, - "engines": { - "node": ">=16 || 14 >=14.17" - }, - "funding": { - "url": "https://github.com/sponsors/isaacs" - } - }, "node_modules/globals": { "version": "16.3.0", "resolved": "https://registry.npmjs.org/globals/-/globals-16.3.0.tgz", @@ -6101,6 +6342,7 @@ "version": "1.2.0", "resolved": "https://registry.npmjs.org/gopd/-/gopd-1.2.0.tgz", "integrity": "sha512-ZUKRh6/kUFoAiTAtTYPZJ3hw9wNxx+BIBOijnlG9PnrJsCcSjs1wyyD6vJpaYtgnzDrKYRSqf3OO6Rfa93xsRg==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.4" @@ -6120,6 +6362,7 @@ "version": "0.5.2", "resolved": "https://registry.npmjs.org/hachure-fill/-/hachure-fill-0.5.2.tgz", "integrity": "sha512-3GKBOn+m2LX9iq+JC1064cSFprJY4jL1jCXTcpnfER5HYE2l/4EfWSGzkPa/ZDBmYI0ZOEj5VHV/eKnPGkHuOg==", + "dev": true, "license": "MIT" }, "node_modules/has-flag": { @@ -6136,6 +6379,7 @@ "version": "1.1.0", "resolved": "https://registry.npmjs.org/has-symbols/-/has-symbols-1.1.0.tgz", "integrity": "sha512-1cDNdwJ2Jaohmb3sg4OmKaMBwuC48sYni5HUw2DvsC8LjGTLK9h+eb1X6RyuOHe4hT0ULCW68iomhjUoKUqlPQ==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.4" @@ -6148,6 +6392,7 @@ "version": "2.0.2", "resolved": "https://registry.npmjs.org/hasown/-/hasown-2.0.2.tgz", "integrity": "sha512-0hJU9SCPvmMzIBdZFqNPXWa6dqh7WdH0cII9y+CyS8rG3nL48Bclra9HmKhVVUHyPWNH5Y7xDwAB7bfgSjkUMQ==", + "dev": true, "license": "MIT", "dependencies": { "function-bind": "^1.1.2" @@ -6276,6 +6521,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/hast-util-is-element/-/hast-util-is-element-3.0.0.tgz", "integrity": "sha512-Val9mnv2IWpLbNPqc/pUem+a7Ipj2aHacCwgNfTiK0vJKl0LF+4Ba4+v1oPHFpf3bLYmreq0/l3Gud9S5OH42g==", + "dev": true, "license": "MIT", "dependencies": { "@types/hast": "^3.0.0" @@ -6303,6 +6549,7 @@ "version": "5.0.2", "resolved": "https://registry.npmjs.org/hast-util-sanitize/-/hast-util-sanitize-5.0.2.tgz", "integrity": "sha512-3yTWghByc50aGS7JlGhk61SPenfE/p1oaFeNwkOOyrscaOkMGrcW9+Cy/QAIOBpZxP1yqDIzFMR0+Np0i0+usg==", + "dev": true, "license": "MIT", "dependencies": { "@types/hast": "^3.0.0", @@ -6318,6 +6565,7 @@ "version": "9.0.5", "resolved": "https://registry.npmjs.org/hast-util-to-html/-/hast-util-to-html-9.0.5.tgz", "integrity": "sha512-OguPdidb+fbHQSU4Q4ZiLKnzWo8Wwsf5bZfbvu7//a9oTYoqD/fWpe96NuHkoS9h0ccGOTe0C4NGXdtS0iObOw==", + "dev": true, "license": "MIT", "dependencies": { "@types/hast": "^3.0.0", @@ -6341,12 +6589,14 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/hast-util-to-text": { "version": "4.0.2", "resolved": "https://registry.npmjs.org/hast-util-to-text/-/hast-util-to-text-4.0.2.tgz", "integrity": "sha512-KK6y/BN8lbaq654j7JgBydev7wuNMcID54lkRav1P0CaE1e47P72AWWPiGKXTJU271ooYzcvTAn/Zt0REnvc7A==", + "dev": true, "license": "MIT", "dependencies": { "@types/hast": "^3.0.0", @@ -6363,12 +6613,14 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/hast-util-whitespace": { "version": "3.0.0", "resolved": "https://registry.npmjs.org/hast-util-whitespace/-/hast-util-whitespace-3.0.0.tgz", "integrity": "sha512-88JUN06ipLwsnv+dVn+OIYOvAuvBMy/Qoi6O7mQHxdPXpjy+Cd6xRkWwux7DKO+4sYILtLBRIKgsdpS2gQc7qw==", + "dev": true, "license": "MIT", "dependencies": { "@types/hast": "^3.0.0" @@ -6410,6 +6662,7 @@ "version": "11.11.1", "resolved": "https://registry.npmjs.org/highlight.js/-/highlight.js-11.11.1.tgz", "integrity": "sha512-Xwwo44whKBVCYoliBQwaPvtd/2tYFkRQtXDWj1nackaV2JPXx3L0+Jvd8/qCJ2p+ML0/XVkJ2q+Mr+UVdpJK5w==", + "dev": true, "license": "BSD-3-Clause", "engines": { "node": ">=12.0.0" @@ -6419,6 +6672,7 @@ "version": "4.12.19", "resolved": "https://registry.npmjs.org/hono/-/hono-4.12.19.tgz", "integrity": "sha512-xa3eYXYXx68XTT4hZ7dRzsXBhaq85ToSrlUJNoR0gwz/1Ap/CNwX47wfvV7pc/xWhjKVVkLT7zBJy8chhNguqQ==", + "dev": true, "license": "MIT", "engines": { "node": ">=16.9.0" @@ -6448,6 +6702,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/html-void-elements/-/html-void-elements-3.0.0.tgz", "integrity": "sha512-bEqo66MRXsUGxWHV5IP0PUiAWwoEjba4VCzg0LjFJBpchPaTfyfCKTG6bc5F8ucKec3q5y6qOdGyYTSBEvhCrg==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -6458,6 +6713,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/http-errors/-/http-errors-2.0.1.tgz", "integrity": "sha512-4FbRdAX+bSdmo4AUFuS0WNiPz8NgFt+r8ThgNWmlrjQjt1Q7ZR9+zTlce2859x4KSXrwIsaeTqDoKQmtP8pLmQ==", + "dev": true, "license": "MIT", "dependencies": { "depd": "~2.0.0", @@ -6521,6 +6777,7 @@ "version": "0.6.3", "resolved": "https://registry.npmjs.org/iconv-lite/-/iconv-lite-0.6.3.tgz", "integrity": "sha512-4fCk79wshMdzMp2rH06qWrJE4iolqLhCUH+OiuIgU++RB0+94NlDL81atO7GX55uUKueo0txHNtvEyI6D7WdMw==", + "dev": true, "license": "MIT", "dependencies": { "safer-buffer": ">= 2.1.2 < 3.0.0" @@ -6567,6 +6824,7 @@ "version": "4.2.0", "resolved": "https://registry.npmjs.org/import-meta-resolve/-/import-meta-resolve-4.2.0.tgz", "integrity": "sha512-Iqv2fzaTQN28s/FwZAoFq0ZSs/7hMAHJVX+w8PZl3cY19Pxk6jFFalxQoIfW2826i/fDLXv8IiEZRIT0lDuWcg==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -6597,18 +6855,21 @@ "version": "2.0.4", "resolved": "https://registry.npmjs.org/inherits/-/inherits-2.0.4.tgz", "integrity": "sha512-k/vGaX4/Yla3WzyMCvTQOXYeIHvqOKtnqBduzTHpzpQZzAskKMhZ2K+EnBiSM9zGSoIFeMpXKxa4dYeZIQqewQ==", + "dev": true, "license": "ISC" }, "node_modules/inline-style-parser": { "version": "0.2.4", "resolved": "https://registry.npmjs.org/inline-style-parser/-/inline-style-parser-0.2.4.tgz", "integrity": "sha512-0aO8FkhNZlj/ZIbNi7Lxxr12obT7cL1moPfE4tg1LkX7LlLfC6DeX4l2ZEud1ukP9jNQyNnfzQVqwbwmAATY4Q==", + "dev": true, "license": "MIT" }, "node_modules/internmap": { "version": "2.0.3", "resolved": "https://registry.npmjs.org/internmap/-/internmap-2.0.3.tgz", "integrity": "sha512-5Hh7Y1wQbvY5ooGgPbDaL5iYLAPzMTUrjMulskHLH6wnv/A+1q5rgEaiuqEjB+oxGXIVZs1FF+R/KPN3ZSQYYg==", + "dev": true, "license": "ISC", "engines": { "node": ">=12" @@ -6618,6 +6879,7 @@ "version": "10.2.0", "resolved": "https://registry.npmjs.org/ip-address/-/ip-address-10.2.0.tgz", "integrity": "sha512-/+S6j4E9AHvW9SWMSEY9Xfy66O5PWvVEJ08O0y5JGyEKQpojb0K0GKpz/v5HJ/G0vi3D2sjGK78119oXZeE0qA==", + "dev": true, "license": "MIT", "engines": { "node": ">= 12" @@ -6627,6 +6889,7 @@ "version": "1.9.1", "resolved": "https://registry.npmjs.org/ipaddr.js/-/ipaddr.js-1.9.1.tgz", "integrity": "sha512-0KI/607xoxSToH7GjN1FfSbLoU0+btTicjsQSWQlh/hZykN8KpmMf7uYwPW3R+akZ6R/w18ZlXSHBYXiYUPO3g==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.10" @@ -6658,16 +6921,6 @@ "node": ">=0.10.0" } }, - "node_modules/is-fullwidth-code-point": { - "version": "3.0.0", - "resolved": "https://registry.npmjs.org/is-fullwidth-code-point/-/is-fullwidth-code-point-3.0.0.tgz", - "integrity": "sha512-zymm5+u+sCsSWyD9qNaejV3DFvhCKclKdizYaJUuHA83RLjb7nSuGnddCHGv0hk+KY7BMAlsWeK4Ueg6EV6XQg==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=8" - } - }, "node_modules/is-glob": { "version": "4.0.3", "resolved": "https://registry.npmjs.org/is-glob/-/is-glob-4.0.3.tgz", @@ -6715,6 +6968,7 @@ "version": "4.1.0", "resolved": "https://registry.npmjs.org/is-plain-obj/-/is-plain-obj-4.1.0.tgz", "integrity": "sha512-+Pgi+vMuUNkJyExiMBt5IlFoMyKnr5zhJ4Uspz58WOhBF5QoIZkFyNHIbBAtHwzVAgk5RtndVNsDRN61/mmDqg==", + "dev": true, "license": "MIT", "engines": { "node": ">=12" @@ -6727,6 +6981,7 @@ "version": "4.0.0", "resolved": "https://registry.npmjs.org/is-promise/-/is-promise-4.0.0.tgz", "integrity": "sha512-hvpoI6korhJMnej285dSg6nu1+e6uxs7zG3BYAm5byqDsgJNWwxzM6z6iZiAgQR4TJ30JmBTOwqZUw3WlyH3AQ==", + "dev": true, "license": "MIT" }, "node_modules/is-wsl": { @@ -6749,6 +7004,7 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/isexe/-/isexe-2.0.0.tgz", "integrity": "sha512-RHxMLp9lnKHGHRng9QFhRCMbYAcVpn69smSGcq3f36xjgVVWThj4qqLbTLlq7Ssj8B+fIQ1EuCEGI2lKsyQeIw==", + "dev": true, "license": "ISC" }, "node_modules/istanbul-lib-coverage": { @@ -6776,21 +7032,6 @@ "node": ">=10" } }, - "node_modules/istanbul-lib-source-maps": { - "version": "5.0.6", - "resolved": "https://registry.npmjs.org/istanbul-lib-source-maps/-/istanbul-lib-source-maps-5.0.6.tgz", - "integrity": "sha512-yg2d+Em4KizZC5niWhQaIomgf5WlL4vOOjZ5xGCmF8SnPE/mDWWXgvRExdcpCgh9lLRRa1/fSYp2ymmbJ1pI+A==", - "dev": true, - "license": "BSD-3-Clause", - "dependencies": { - "@jridgewell/trace-mapping": "^0.3.23", - "debug": "^4.1.1", - "istanbul-lib-coverage": "^3.0.0" - }, - "engines": { - "node": ">=10" - } - }, "node_modules/istanbul-reports": { "version": "3.2.0", "resolved": "https://registry.npmjs.org/istanbul-reports/-/istanbul-reports-3.2.0.tgz", @@ -6805,22 +7046,6 @@ "node": ">=8" } }, - "node_modules/jackspeak": { - "version": "3.4.3", - "resolved": "https://registry.npmjs.org/jackspeak/-/jackspeak-3.4.3.tgz", - "integrity": "sha512-OGlZQpz2yfahA/Rd1Y8Cd9SIEsqvXkLVoSw/cgwhnhFMDbsQFeZYoJJ7bIZBS9BcamUW96asq/npPWugM+RQBw==", - "dev": true, - "license": "BlueOak-1.0.0", - "dependencies": { - "@isaacs/cliui": "^8.0.2" - }, - "funding": { - "url": "https://github.com/sponsors/isaacs" - }, - "optionalDependencies": { - "@pkgjs/parseargs": "^0.11.0" - } - }, "node_modules/jiti": { "version": "2.4.2", "resolved": "https://registry.npmjs.org/jiti/-/jiti-2.4.2.tgz", @@ -6835,15 +7060,16 @@ "version": "6.1.3", "resolved": "https://registry.npmjs.org/jose/-/jose-6.1.3.tgz", "integrity": "sha512-0TpaTfihd4QMNwrz/ob2Bp7X04yuxJkjRGi4aKmOqwhov54i6u79oCv7T+C7lo70MKH6BesI3vscD1yb/yzKXQ==", + "dev": true, "license": "MIT", "funding": { "url": "https://github.com/sponsors/panva" } }, "node_modules/js-tokens": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-4.0.0.tgz", - "integrity": "sha512-RdJUflcE3cUzKiMqQgsCu06FPu9UdIJO0beYbPhHN4k6apgJtifcoCtT9bcxOpYBtpD2kCM6Sbzg4CausW/PKQ==", + "version": "10.0.0", + "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-10.0.0.tgz", + "integrity": "sha512-lM/UBzQmfJRo9ABXbPWemivdCW8V2G8FHaHdypQaIy523snUjog0W71ayWXTjiR+ixeMyVHN2XcpnTd/liPg/Q==", "dev": true, "license": "MIT" }, @@ -6878,6 +7104,7 @@ "version": "8.0.2", "resolved": "https://registry.npmjs.org/json-schema-typed/-/json-schema-typed-8.0.2.tgz", "integrity": "sha512-fQhoXdcvc3V28x7C7BMs4P5+kNlgUURe2jmUT1T//oBRMDrqy1QPelJimwZGo7Hg9VPV3EQV5Bnq4hbFy2vetA==", + "dev": true, "license": "BSD-2-Clause" }, "node_modules/json-stable-stringify-without-jsonify": { @@ -6904,6 +7131,7 @@ "version": "0.16.47", "resolved": "https://registry.npmjs.org/katex/-/katex-0.16.47.tgz", "integrity": "sha512-Eeo8Ys1doU1z+x8AZsPpQu+p/QcZBI5PeOo7QGQdy2x2m0MU/hYagBbGOmXwr5KVbEfVuWv9LpnQWeehogurjg==", + "dev": true, "funding": [ "https://opencollective.com/katex", "https://github.com/sponsors/katex" @@ -6929,7 +7157,8 @@ "node_modules/khroma": { "version": "2.1.0", "resolved": "https://registry.npmjs.org/khroma/-/khroma-2.1.0.tgz", - "integrity": "sha512-Ls993zuzfayK269Svk9hzpeGUKob/sIgZzyHYdjQoAdQetRKpOLj+k/QQQ/6Qi0Yz65mlROrfd+Ev+1+7dz9Kw==" + "integrity": "sha512-Ls993zuzfayK269Svk9hzpeGUKob/sIgZzyHYdjQoAdQetRKpOLj+k/QQQ/6Qi0Yz65mlROrfd+Ev+1+7dz9Kw==", + "dev": true }, "node_modules/kleur": { "version": "4.1.5", @@ -6952,6 +7181,7 @@ "version": "1.0.2", "resolved": "https://registry.npmjs.org/layout-base/-/layout-base-1.0.2.tgz", "integrity": "sha512-8h2oVEZNktL4BH2JCOI90iD1yXwL6iNW7KcCKT2QZgQJR2vbqDsldCTPRU9NifTCqHZci57XvQQ15YTu+sTYPg==", + "dev": true, "license": "MIT" }, "node_modules/levn": { @@ -7221,6 +7451,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/locate-character/-/locate-character-3.0.0.tgz", "integrity": "sha512-SW13ws7BjaeJ6p7Q6CO2nchbYEc3X3J6WrmTTDto7yMPqVSZTUyY5Tjbid+Ab8gLnATtygYtiDIJGQRRn2ZOiA==", + "dev": true, "license": "MIT" }, "node_modules/locate-path": { @@ -7243,6 +7474,7 @@ "version": "4.18.1", "resolved": "https://registry.npmjs.org/lodash-es/-/lodash-es-4.18.1.tgz", "integrity": "sha512-J8xewKD/Gk22OZbhpOVSwcs60zhd95ESDwezOFuA3/099925PdHJ7OFHNTGtajL3AlZkykD32HykiMo+BIBI8A==", + "dev": true, "license": "MIT" }, "node_modules/lodash.castarray": { @@ -7270,6 +7502,7 @@ "version": "3.1.0", "resolved": "https://registry.npmjs.org/longest-streak/-/longest-streak-3.1.0.tgz", "integrity": "sha512-9Ri+o0JYgehTaVBBDoMqIl8GXtbWg711O3srftcHhZ0dqnETqLaoIK0x17fUw9rFSlK/0NlsKe0Ahhyl5pXE2g==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -7287,6 +7520,7 @@ "version": "3.3.0", "resolved": "https://registry.npmjs.org/lowlight/-/lowlight-3.3.0.tgz", "integrity": "sha512-0JNhgFoPvP6U6lE/UdVsSq99tn6DhjjpAj5MxG49ewd2mOBVtwWYIT8ClyABhq198aXXODMU6Ox8DrGy/CpTZQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/hast": "^3.0.0", @@ -7298,13 +7532,6 @@ "url": "https://github.com/sponsors/wooorm" } }, - "node_modules/lru-cache": { - "version": "10.4.3", - "resolved": "https://registry.npmjs.org/lru-cache/-/lru-cache-10.4.3.tgz", - "integrity": "sha512-JNAzZcXrCt42VGLuYz0zfAzDfAvJWW6AfYlDBQyDV5DClI2m5sAmK+OIO7s59XfsRsWHp02jAJrRadPRGTt6SQ==", - "dev": true, - "license": "ISC" - }, "node_modules/lz-string": { "version": "1.5.0", "resolved": "https://registry.npmjs.org/lz-string/-/lz-string-1.5.0.tgz", @@ -7316,24 +7543,25 @@ } }, "node_modules/magic-string": { - "version": "0.30.17", - "resolved": "https://registry.npmjs.org/magic-string/-/magic-string-0.30.17.tgz", - "integrity": "sha512-sNPKHvyjVf7gyjwS4xGTaW/mCnF8wnjtifKBEhxfZ7E/S8tQ0rssrwGNn6q8JH/ohItJfSQp9mBtQYuTlH5QnA==", + "version": "0.30.21", + "resolved": "https://registry.npmjs.org/magic-string/-/magic-string-0.30.21.tgz", + "integrity": "sha512-vd2F4YUyEXKGcLHoq+TEyCjxueSeHnFxyyjNp80yg0XV4vUhnDer/lvvlqM/arB5bXQN5K2/3oinyCRyx8T2CQ==", + "dev": true, "license": "MIT", "dependencies": { - "@jridgewell/sourcemap-codec": "^1.5.0" + "@jridgewell/sourcemap-codec": "^1.5.5" } }, "node_modules/magicast": { - "version": "0.3.5", - "resolved": "https://registry.npmjs.org/magicast/-/magicast-0.3.5.tgz", - "integrity": "sha512-L0WhttDl+2BOsybvEOLK7fW3UA0OQ0IQ2d6Zl2x/a6vVRs3bAY0ECOSHHeL5jD+SbOpOCUEi0y1DgHEn9Qn1AQ==", + "version": "0.5.3", + "resolved": "https://registry.npmjs.org/magicast/-/magicast-0.5.3.tgz", + "integrity": "sha512-pVKE4UdSQ7DvHzivsCIFx2BJn1mHG6KsyrFcaxFx6tONdneEuThrDx0Cj3AMg58KyN4pzYT+LHOotxDQDjNvkw==", "dev": true, "license": "MIT", "dependencies": { - "@babel/parser": "^7.25.4", - "@babel/types": "^7.25.4", - "source-map-js": "^1.2.0" + "@babel/parser": "^7.29.3", + "@babel/types": "^7.29.0", + "source-map-js": "^1.2.1" } }, "node_modules/make-dir": { @@ -7356,6 +7584,7 @@ "version": "3.0.4", "resolved": "https://registry.npmjs.org/markdown-table/-/markdown-table-3.0.4.tgz", "integrity": "sha512-wiYz4+JrLyb/DqW2hkFJxP7Vd7JuTDm77fvbM8VfEQdmSMqcImWeeRbHwZjBjIFki/VaMK2BhFi7oUUZeM5bqw==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -7366,6 +7595,7 @@ "version": "16.4.2", "resolved": "https://registry.npmjs.org/marked/-/marked-16.4.2.tgz", "integrity": "sha512-TI3V8YYWvkVf3KJe1dRkpnjs68JUPyEa5vjKrp1XEEJUAOaQc+Qj+L1qWbPd0SJuAdQkFU0h73sXXqwDYxsiDA==", + "dev": true, "license": "MIT", "bin": { "marked": "bin/marked.js" @@ -7378,6 +7608,7 @@ "version": "1.1.0", "resolved": "https://registry.npmjs.org/math-intrinsics/-/math-intrinsics-1.1.0.tgz", "integrity": "sha512-/IXtbwEk5HTPyEwyKX6hGkYXxM9nbj64B+ilVJnC/R6B0pH5G4V3b0pVbL7DBj4tkhBAppbQUlf6F6Xl9LHu1g==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.4" @@ -7394,6 +7625,7 @@ "version": "3.0.2", "resolved": "https://registry.npmjs.org/mdast-util-find-and-replace/-/mdast-util-find-and-replace-3.0.2.tgz", "integrity": "sha512-Tmd1Vg/m3Xz43afeNxDIhWRtFZgM2VLyaf4vSTYwudTyeuTneoL3qtWMA5jeLyz/O1vDJmmV4QuScFCA2tBPwg==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -7410,6 +7642,7 @@ "version": "5.0.0", "resolved": "https://registry.npmjs.org/escape-string-regexp/-/escape-string-regexp-5.0.0.tgz", "integrity": "sha512-/veY75JbMK4j1yjvuUxuVsiS/hr/4iHs9FTT6cgTexxdE0Ly/glccBAkloH/DofkjRbZU3bnoj38mOmhkZ0lHw==", + "dev": true, "license": "MIT", "engines": { "node": ">=12" @@ -7422,6 +7655,7 @@ "version": "2.0.2", "resolved": "https://registry.npmjs.org/mdast-util-from-markdown/-/mdast-util-from-markdown-2.0.2.tgz", "integrity": "sha512-uZhTV/8NBuw0WHkPTrCqDOl0zVe1BIng5ZtHoDk49ME1qqcjYmmLmOf0gELgcRMxN4w2iuIeVso5/6QymSrgmA==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -7446,12 +7680,14 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/mdast-util-from-markdown/node_modules/unist-util-stringify-position": { "version": "4.0.0", "resolved": "https://registry.npmjs.org/unist-util-stringify-position/-/unist-util-stringify-position-4.0.0.tgz", "integrity": "sha512-0ASV06AAoKCDkS2+xw5RXJywruurpbC4JZSm7nr7MOt1ojAzvyyaO+UxZf18j8FCF6kmzCZKcAgN/yu2gm2XgQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "^3.0.0" @@ -7465,6 +7701,7 @@ "version": "3.1.0", "resolved": "https://registry.npmjs.org/mdast-util-gfm/-/mdast-util-gfm-3.1.0.tgz", "integrity": "sha512-0ulfdQOM3ysHhCJ1p06l0b0VKlhU0wuQs3thxZQagjcjPrlFRqY215uZGHHJan9GEAXd9MbfPjFJz+qMkVR6zQ==", + "dev": true, "license": "MIT", "dependencies": { "mdast-util-from-markdown": "^2.0.0", @@ -7484,6 +7721,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/mdast-util-gfm-autolink-literal/-/mdast-util-gfm-autolink-literal-2.0.1.tgz", "integrity": "sha512-5HVP2MKaP6L+G6YaxPNjuL0BPrq9orG3TsrZ9YXbA3vDw/ACI4MEsnoDpn6ZNm7GnZgtAcONJyPhOP8tNJQavQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -7501,6 +7739,7 @@ "version": "2.1.0", "resolved": "https://registry.npmjs.org/mdast-util-gfm-footnote/-/mdast-util-gfm-footnote-2.1.0.tgz", "integrity": "sha512-sqpDWlsHn7Ac9GNZQMeUzPQSMzR6Wv0WKRNvQRg0KqHh02fpTz69Qc1QSseNX29bhz1ROIyNyxExfawVKTm1GQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -7518,6 +7757,7 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/mdast-util-gfm-strikethrough/-/mdast-util-gfm-strikethrough-2.0.0.tgz", "integrity": "sha512-mKKb915TF+OC5ptj5bJ7WFRPdYtuHv0yTRxK2tJvi+BDqbkiG7h7u/9SI89nRAYcmap2xHQL9D+QG/6wSrTtXg==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -7533,6 +7773,7 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/mdast-util-gfm-table/-/mdast-util-gfm-table-2.0.0.tgz", "integrity": "sha512-78UEvebzz/rJIxLvE7ZtDd/vIQ0RHv+3Mh5DR96p7cS7HsBhYIICDBCu8csTNWNO6tBWfqXPWekRuj2FNOGOZg==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -7550,6 +7791,7 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/mdast-util-gfm-task-list-item/-/mdast-util-gfm-task-list-item-2.0.0.tgz", "integrity": "sha512-IrtvNvjxC1o06taBAVJznEnkiHxLFTzgonUdy8hzFVeDun0uTjxxrRGVaNFqkU1wJR3RBPEfsxmU6jDWPofrTQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -7586,6 +7828,7 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/mdast-util-newline-to-break/-/mdast-util-newline-to-break-2.0.0.tgz", "integrity": "sha512-MbgeFca0hLYIEx/2zGsszCSEJJ1JSCdiY5xQxRcLDDGa8EPvlLPupJ4DSajbMPAnC0je8jfb9TiUATnxxrHUog==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -7600,6 +7843,7 @@ "version": "4.1.0", "resolved": "https://registry.npmjs.org/mdast-util-phrasing/-/mdast-util-phrasing-4.1.0.tgz", "integrity": "sha512-TqICwyvJJpBwvGAMZjj4J2n0X8QWp21b9l0o7eXyVJ25YNWYbJDVIyD1bZXE6WtV6RmKJVYmQAKWa0zWOABz2w==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -7614,6 +7858,7 @@ "version": "13.2.1", "resolved": "https://registry.npmjs.org/mdast-util-to-hast/-/mdast-util-to-hast-13.2.1.tgz", "integrity": "sha512-cctsq2wp5vTsLIcaymblUriiTcZd0CwWtCbLvrOzYCDZoWyMNV8sZ7krj09FSnsiJi3WVsHLM4k6Dq/yaPyCXA==", + "dev": true, "license": "MIT", "dependencies": { "@types/hast": "^3.0.0", @@ -7635,6 +7880,7 @@ "version": "2.1.2", "resolved": "https://registry.npmjs.org/mdast-util-to-markdown/-/mdast-util-to-markdown-2.1.2.tgz", "integrity": "sha512-xj68wMTvGXVOKonmog6LwyJKrYXZPvlwabaryTjLh9LuvovB/KAH+kvi8Gjj+7rJjsFi23nkUxRQv1KqSroMqA==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -7656,12 +7902,14 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/mdast-util-to-string": { "version": "4.0.0", "resolved": "https://registry.npmjs.org/mdast-util-to-string/-/mdast-util-to-string-4.0.0.tgz", "integrity": "sha512-0H44vDimn51F0YwvxSJSm0eCDOJTRlmN0R1yBh4HLj9wiV1Dn0QoXGbvFAWj2hSItVTlCmBF1hqKlIyUBVFLPg==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0" @@ -7735,6 +7983,7 @@ "version": "1.1.0", "resolved": "https://registry.npmjs.org/media-typer/-/media-typer-1.1.0.tgz", "integrity": "sha512-aisnrDP4GNe06UcKFnV5bfMNPBUw4jsLGaWwWfnH3v02GnBuXX2MCVn5RbrWo0j3pczUilYblq7fQ7Nw2t5XKw==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.8" @@ -7744,6 +7993,7 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/merge-descriptors/-/merge-descriptors-2.0.0.tgz", "integrity": "sha512-Snk314V5ayFLhp3fkUREub6WtjBfPdCPY1Ln8/8munuLuiYhsABgBVWsozAG+MWMbVEvcdcpbi9R7ww22l9Q3g==", + "dev": true, "license": "MIT", "engines": { "node": ">=18" @@ -7756,6 +8006,7 @@ "version": "11.15.0", "resolved": "https://registry.npmjs.org/mermaid/-/mermaid-11.15.0.tgz", "integrity": "sha512-pTMbcf3rWdtLiYGpmoTjHEpeY8seiy6sR+9nD7LOs8KfUbHE4lOUAprTRqRAcWSQ6MQpdX+YEsxShtGsINtPtw==", + "dev": true, "license": "MIT", "dependencies": { "@braintree/sanitize-url": "^7.1.1", @@ -7785,6 +8036,7 @@ "version": "4.0.2", "resolved": "https://registry.npmjs.org/micromark/-/micromark-4.0.2.tgz", "integrity": "sha512-zpe98Q6kvavpCr1NPVSCMebCKfD7CA2NqZ+rykeNhONIJBpc1tFKt9hucLGwha3jNTNI8lHpctWJWoimVF4PfA==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -7820,6 +8072,7 @@ "version": "2.0.3", "resolved": "https://registry.npmjs.org/micromark-core-commonmark/-/micromark-core-commonmark-2.0.3.tgz", "integrity": "sha512-RDBrHEMSxVFLg6xvnXmb1Ayr2WzLAWjeSATAoxwKYJV94TeNavgoIdA0a9ytzDSVzBy2YKFK+emCPOEibLeCrg==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -7854,6 +8107,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/micromark-extension-gfm/-/micromark-extension-gfm-3.0.0.tgz", "integrity": "sha512-vsKArQsicm7t0z2GugkCKtZehqUm31oeGBV/KVSorWSy8ZlNAv7ytjFhvaryUiCUJYqs+NoE6AFhpQvBTM6Q4w==", + "dev": true, "license": "MIT", "dependencies": { "micromark-extension-gfm-autolink-literal": "^2.0.0", @@ -7874,6 +8128,7 @@ "version": "2.1.0", "resolved": "https://registry.npmjs.org/micromark-extension-gfm-autolink-literal/-/micromark-extension-gfm-autolink-literal-2.1.0.tgz", "integrity": "sha512-oOg7knzhicgQ3t4QCjCWgTmfNhvQbDDnJeVu9v81r7NltNCVmhPy1fJRX27pISafdjL+SVc4d3l48Gb6pbRypw==", + "dev": true, "license": "MIT", "dependencies": { "micromark-util-character": "^2.0.0", @@ -7890,6 +8145,7 @@ "version": "2.1.0", "resolved": "https://registry.npmjs.org/micromark-extension-gfm-footnote/-/micromark-extension-gfm-footnote-2.1.0.tgz", "integrity": "sha512-/yPhxI1ntnDNsiHtzLKYnE3vf9JZ6cAisqVDauhp4CEHxlb4uoOTxOCJ+9s51bIB8U1N1FJ1RXOKTIlD5B/gqw==", + "dev": true, "license": "MIT", "dependencies": { "devlop": "^1.0.0", @@ -7910,6 +8166,7 @@ "version": "2.1.0", "resolved": "https://registry.npmjs.org/micromark-extension-gfm-strikethrough/-/micromark-extension-gfm-strikethrough-2.1.0.tgz", "integrity": "sha512-ADVjpOOkjz1hhkZLlBiYA9cR2Anf8F4HqZUO6e5eDcPQd0Txw5fxLzzxnEkSkfnD0wziSGiv7sYhk/ktvbf1uw==", + "dev": true, "license": "MIT", "dependencies": { "devlop": "^1.0.0", @@ -7928,6 +8185,7 @@ "version": "2.1.1", "resolved": "https://registry.npmjs.org/micromark-extension-gfm-table/-/micromark-extension-gfm-table-2.1.1.tgz", "integrity": "sha512-t2OU/dXXioARrC6yWfJ4hqB7rct14e8f7m0cbI5hUmDyyIlwv5vEtooptH8INkbLzOatzKuVbQmAYcbWoyz6Dg==", + "dev": true, "license": "MIT", "dependencies": { "devlop": "^1.0.0", @@ -7945,6 +8203,7 @@ "version": "2.0.0", "resolved": "https://registry.npmjs.org/micromark-extension-gfm-tagfilter/-/micromark-extension-gfm-tagfilter-2.0.0.tgz", "integrity": "sha512-xHlTOmuCSotIA8TW1mDIM6X2O1SiX5P9IuDtqGonFhEK0qgRI4yeC6vMxEV2dgyr2TiD+2PQ10o+cOhdVAcwfg==", + "dev": true, "license": "MIT", "dependencies": { "micromark-util-types": "^2.0.0" @@ -7958,6 +8217,7 @@ "version": "2.1.0", "resolved": "https://registry.npmjs.org/micromark-extension-gfm-task-list-item/-/micromark-extension-gfm-task-list-item-2.1.0.tgz", "integrity": "sha512-qIBZhqxqI6fjLDYFTBIa4eivDMnP+OZqsNwmQ3xNLE4Cxwc+zfQEfbs6tzAo2Hjq+bh6q5F+Z8/cksrLFYWQQw==", + "dev": true, "license": "MIT", "dependencies": { "devlop": "^1.0.0", @@ -7995,6 +8255,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-factory-destination/-/micromark-factory-destination-2.0.1.tgz", "integrity": "sha512-Xe6rDdJlkmbFRExpTOmRj9N3MaWmbAgdpSrBQvCFqhezUn4AHqJHbaEnfbVYYiexVSs//tqOdY/DxhjdCiJnIA==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8016,6 +8277,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-factory-label/-/micromark-factory-label-2.0.1.tgz", "integrity": "sha512-VFMekyQExqIW7xIChcXn4ok29YE3rnuyveW3wZQWWqF4Nv9Wk5rgJ99KzPvHjkmPXF93FXIbBp6YdW3t71/7Vg==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8038,6 +8300,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-factory-space/-/micromark-factory-space-2.0.1.tgz", "integrity": "sha512-zRkxjtBxxLd2Sc0d+fbnEunsTj46SWXgXciZmHq0kDYGnck/ZSGj9/wULTV95uoeYiK5hRXP2mJ98Uo4cq/LQg==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8058,6 +8321,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-factory-title/-/micromark-factory-title-2.0.1.tgz", "integrity": "sha512-5bZ+3CjhAd9eChYTHsjy6TGxpOFSKgKKJPJxr293jTbfry2KDoWkhBb6TcPVB4NmzaPhMs1Frm9AZH7OD4Cjzw==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8080,6 +8344,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-factory-whitespace/-/micromark-factory-whitespace-2.0.1.tgz", "integrity": "sha512-Ob0nuZ3PKt/n0hORHyvoD9uZhr+Za8sFoP+OnMcnWK5lngSzALgQYKMr9RJVOWLqQYuyn6ulqGWSXdwf6F80lQ==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8102,6 +8367,7 @@ "version": "2.1.1", "resolved": "https://registry.npmjs.org/micromark-util-character/-/micromark-util-character-2.1.1.tgz", "integrity": "sha512-wv8tdUTJ3thSFFFJKtpYKOYiGP2+v96Hvk4Tu8KpCAsTMs6yi+nVmGh1syvSCsaxz45J6Jbw+9DD6g97+NV67Q==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8122,6 +8388,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-util-chunked/-/micromark-util-chunked-2.0.1.tgz", "integrity": "sha512-QUNFEOPELfmvv+4xiNg2sRYeS/P84pTW0TCgP5zc9FpXetHY0ab7SxKyAQCNCc1eK0459uoLI1y5oO5Vc1dbhA==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8141,6 +8408,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-util-classify-character/-/micromark-util-classify-character-2.0.1.tgz", "integrity": "sha512-K0kHzM6afW/MbeWYWLjoHQv1sgg2Q9EccHEDzSkxiP/EaagNzCm7T/WMKZ3rjMbvIpvBiZgwR3dKMygtA4mG1Q==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8162,6 +8430,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-util-combine-extensions/-/micromark-util-combine-extensions-2.0.1.tgz", "integrity": "sha512-OnAnH8Ujmy59JcyZw8JSbK9cGpdVY44NKgSM7E9Eh7DiLS2E9RNQf0dONaGDzEG9yjEl5hcqeIsj4hfRkLH/Bg==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8182,6 +8451,7 @@ "version": "2.0.2", "resolved": "https://registry.npmjs.org/micromark-util-decode-numeric-character-reference/-/micromark-util-decode-numeric-character-reference-2.0.2.tgz", "integrity": "sha512-ccUbYk6CwVdkmCQMyr64dXz42EfHGkPQlBj5p7YVGzq8I7CtjXZJrubAYezf7Rp+bjPseiROqe7G6foFd+lEuw==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8201,6 +8471,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-util-decode-string/-/micromark-util-decode-string-2.0.1.tgz", "integrity": "sha512-nDV/77Fj6eH1ynwscYTOsbK7rR//Uj0bZXBwJZRfaLEJ1iGBR6kIfNmlNqaqJf649EP0F3NWNdeJi03elllNUQ==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8223,6 +8494,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-util-encode/-/micromark-util-encode-2.0.1.tgz", "integrity": "sha512-c3cVx2y4KqUnwopcO9b/SCdo2O67LwJJ/UyqGfbigahfegL9myoEFoDYZgkT7f36T0bLrM9hZTAaAyH+PCAXjw==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8239,6 +8511,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-util-html-tag-name/-/micromark-util-html-tag-name-2.0.1.tgz", "integrity": "sha512-2cNEiYDhCWKI+Gs9T0Tiysk136SnR13hhO8yW6BGNyhOC4qYFnwF1nKfD3HFAIXA5c45RrIG1ub11GiXeYd1xA==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8255,6 +8528,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-util-normalize-identifier/-/micromark-util-normalize-identifier-2.0.1.tgz", "integrity": "sha512-sxPqmo70LyARJs0w2UclACPUUEqltCkJ6PhKdMIDuJ3gSf/Q+/GIe3WKl0Ijb/GyH9lOpUkRAO2wp0GVkLvS9Q==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8274,6 +8548,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-util-resolve-all/-/micromark-util-resolve-all-2.0.1.tgz", "integrity": "sha512-VdQyxFWFT2/FGJgwQnJYbe1jjQoNTS4RjglmSjTUlpUMa95Htx9NHeYW4rGDJzbjvCsl9eLjMQwGeElsqmzcHg==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8293,6 +8568,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-util-sanitize-uri/-/micromark-util-sanitize-uri-2.0.1.tgz", "integrity": "sha512-9N9IomZ/YuGGZZmQec1MbgxtlgougxTodVwDzzEouPKo3qFWvymFHWcnDi2vzV1ff6kas9ucW+o3yzJK9YB1AQ==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8314,6 +8590,7 @@ "version": "2.1.0", "resolved": "https://registry.npmjs.org/micromark-util-subtokenize/-/micromark-util-subtokenize-2.1.0.tgz", "integrity": "sha512-XQLu552iSctvnEcgXw6+Sx75GflAPNED1qx7eBJ+wydBb2KCbRZe+NwvIEEMM83uml1+2WSXpBAcp9IUCgCYWA==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8336,6 +8613,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/micromark-util-symbol/-/micromark-util-symbol-2.0.1.tgz", "integrity": "sha512-vs5t8Apaud9N28kgCrRUdEed4UJ+wWNvicHLPxCa9ENlYuAY31M0ETy5y1vA33YoNPDFTghEbnh6efaE8h4x0Q==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8352,6 +8630,7 @@ "version": "2.0.2", "resolved": "https://registry.npmjs.org/micromark-util-types/-/micromark-util-types-2.0.2.tgz", "integrity": "sha512-Yw0ECSpJoViF1qTU4DC6NwtC4aWGt1EkzaQB8KPPyCRR8z9TWeV0HbEFGTO+ZY1wB22zmxnJqhPyTpOVCpeHTA==", + "dev": true, "funding": [ { "type": "GitHub Sponsors", @@ -8410,6 +8689,7 @@ "version": "1.54.0", "resolved": "https://registry.npmjs.org/mime-db/-/mime-db-1.54.0.tgz", "integrity": "sha512-aU5EJuIN2WDemCcAp2vFBfp/m4EAhWJnUNSSw0ixs7/kXbd6Pg64EmwJkNdFhB8aWt1sH2CTXrLxo/iAGV3oPQ==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.6" @@ -8419,6 +8699,7 @@ "version": "3.0.2", "resolved": "https://registry.npmjs.org/mime-types/-/mime-types-3.0.2.tgz", "integrity": "sha512-Lbgzdk0h4juoQ9fCKXW4by0UJqj+nOOrI9MJ1sSj4nI8aI2eo1qmvQEie4VD1glsS250n15LsWsYtCugiStS5A==", + "dev": true, "license": "MIT", "dependencies": { "mime-db": "^1.54.0" @@ -8501,6 +8782,7 @@ "version": "1.1.0", "resolved": "https://registry.npmjs.org/mode-watcher/-/mode-watcher-1.1.0.tgz", "integrity": "sha512-mUT9RRGPDYenk59qJauN1rhsIMKBmWA3xMF+uRwE8MW/tjhaDSCCARqkSuDTq8vr4/2KcAxIGVjACxTjdk5C3g==", + "dev": true, "license": "MIT", "dependencies": { "runed": "^0.25.0", @@ -8534,6 +8816,7 @@ "version": "2.1.3", "resolved": "https://registry.npmjs.org/ms/-/ms-2.1.3.tgz", "integrity": "sha512-6FlzubTLZG3J2a/NVCAleEhjzq5oxgHyaCU9yYXvcLsvoVaHJq/s5xXI6/XXP6tz7R9xAOtHnSO/tXtF3WRTlA==", + "dev": true, "license": "MIT" }, "node_modules/nanoid": { @@ -8566,6 +8849,7 @@ "version": "1.0.0", "resolved": "https://registry.npmjs.org/negotiator/-/negotiator-1.0.0.tgz", "integrity": "sha512-8Ofs/AUQh8MaEcrlq5xOX0CQ9ypTF5dl78mjlMNfOK08fzpgTHQRQPBxcPlEtIw0yRpws+Zo/3r+5WRby7u3Gg==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.6" @@ -8583,6 +8867,7 @@ "version": "4.1.1", "resolved": "https://registry.npmjs.org/object-assign/-/object-assign-4.1.1.tgz", "integrity": "sha512-rJgTQnkUnH1sFw8yT6VSU3zD3sWmu6sZhIseY8VX+GRu3P6F7Fu+JNDoXfklElbLJSnc3FUQHVe4cU5hj+BcUg==", + "dev": true, "license": "MIT", "engines": { "node": ">=0.10.0" @@ -8592,6 +8877,7 @@ "version": "1.13.4", "resolved": "https://registry.npmjs.org/object-inspect/-/object-inspect-1.13.4.tgz", "integrity": "sha512-W67iLl4J2EXEGTbfeHCffrjDfitvLANg0UlX3wFUUSTx92KXRFegMHUVgSqE+wvhAbi4WqjGg9czysTV2Epbew==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.4" @@ -8600,10 +8886,25 @@ "url": "https://github.com/sponsors/ljharb" } }, + "node_modules/obug": { + "version": "2.1.2", + "resolved": "https://registry.npmjs.org/obug/-/obug-2.1.2.tgz", + "integrity": "sha512-AWGB9WFcRXOQs48Z/udjI5ZcZMHXwX8XPByNpOydgcGsDLIzjGizhoMWJyKAWze7AVW/2W1i+/gPX4YtKe5cyg==", + "dev": true, + "funding": [ + "https://github.com/sponsors/sxzz", + "https://opencollective.com/debug" + ], + "license": "MIT", + "engines": { + "node": ">=12.20.0" + } + }, "node_modules/on-finished": { "version": "2.4.1", "resolved": "https://registry.npmjs.org/on-finished/-/on-finished-2.4.1.tgz", "integrity": "sha512-oVlzkg3ENAhCk2zdv7IJwd/QUD4z2RxRwpkcGY8psCVcCYZNq4wYnVWALHM+brtuJjePWiYF/ClmuDr8Ch5+kg==", + "dev": true, "license": "MIT", "dependencies": { "ee-first": "1.1.1" @@ -8616,6 +8917,7 @@ "version": "1.4.0", "resolved": "https://registry.npmjs.org/once/-/once-1.4.0.tgz", "integrity": "sha512-lNaJgI+2Q5URQBkccEKHTQOPaXdUxnZZElQTZY0MFUAuaEqe1E+Nyvgdz/aIyNi6Z9MzO5dv1H8n58/GELp3+w==", + "dev": true, "license": "ISC", "dependencies": { "wrappy": "1" @@ -8700,17 +9002,11 @@ "url": "https://github.com/sponsors/sindresorhus" } }, - "node_modules/package-json-from-dist": { - "version": "1.0.1", - "resolved": "https://registry.npmjs.org/package-json-from-dist/-/package-json-from-dist-1.0.1.tgz", - "integrity": "sha512-UEZIS3/by4OC8vL3P2dTXRETpebLI2NiI5vIrjaD/5UtrkFX/tNbwjTSRAGC/+7CAo2pIcBaRgWmcBBHcsaCIw==", - "dev": true, - "license": "BlueOak-1.0.0" - }, "node_modules/package-manager-detector": { "version": "1.6.0", "resolved": "https://registry.npmjs.org/package-manager-detector/-/package-manager-detector-1.6.0.tgz", "integrity": "sha512-61A5ThoTiDG/C8s8UMZwSorAGwMJ0ERVGj2OjoW5pAalsNOg15+iQiPzrLJ4jhZ1HJzmC2PIHT2oEiH3R5fzNA==", + "dev": true, "license": "MIT" }, "node_modules/parent-module": { @@ -8743,6 +9039,7 @@ "version": "1.3.3", "resolved": "https://registry.npmjs.org/parseurl/-/parseurl-1.3.3.tgz", "integrity": "sha512-CiyeOxFT/JZyN5m0z9PfXw4SCBJ6Sygz1Dpl0wqjlhDEGGBP1GnsUVEL0p63hoG1fcj3fHynXi9NYO4nWOL+qQ==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.8" @@ -8752,6 +9049,7 @@ "version": "0.1.0", "resolved": "https://registry.npmjs.org/path-data-parser/-/path-data-parser-0.1.0.tgz", "integrity": "sha512-NOnmBpt5Y2RWbuv0LMzsayp3lVylAHLPUTut412ZA3l+C4uw4ZVkQbjShYCQ8TCpUMdPapr4YjUqLYD6v68j+w==", + "dev": true, "license": "MIT" }, "node_modules/path-exists": { @@ -8768,32 +9066,17 @@ "version": "3.1.1", "resolved": "https://registry.npmjs.org/path-key/-/path-key-3.1.1.tgz", "integrity": "sha512-ojmeN0qd+y0jszEtoY48r0Peq5dwMEkIlCOu6Q5f41lfkswXuKtYrhgoTpLnyIcHm24Uhqx+5Tqm2InSwLhE6Q==", + "dev": true, "license": "MIT", "engines": { "node": ">=8" } }, - "node_modules/path-scurry": { - "version": "1.11.1", - "resolved": "https://registry.npmjs.org/path-scurry/-/path-scurry-1.11.1.tgz", - "integrity": "sha512-Xa4Nw17FS9ApQFJ9umLiJS4orGjm7ZzwUrwamcGQuHSzDyth9boKDaycYdDcZDuqYATXw4HFXgaqWTctW/v1HA==", - "dev": true, - "license": "BlueOak-1.0.0", - "dependencies": { - "lru-cache": "^10.2.0", - "minipass": "^5.0.0 || ^6.0.2 || ^7.0.0" - }, - "engines": { - "node": ">=16 || 14 >=14.18" - }, - "funding": { - "url": "https://github.com/sponsors/isaacs" - } - }, "node_modules/path-to-regexp": { "version": "8.4.2", "resolved": "https://registry.npmjs.org/path-to-regexp/-/path-to-regexp-8.4.2.tgz", "integrity": "sha512-qRcuIdP69NPm4qbACK+aDogI5CBDMi1jKe0ry5rSQJz8JVLsC7jV8XpiJjGRLLol3N+R5ihGYcrPLTno6pAdBA==", + "dev": true, "license": "MIT", "funding": { "type": "opencollective", @@ -8821,6 +9104,7 @@ "version": "5.4.54", "resolved": "https://registry.npmjs.org/pdfjs-dist/-/pdfjs-dist-5.4.54.tgz", "integrity": "sha512-TBAiTfQw89gU/Z4LW98Vahzd2/LoCFprVGvGbTgFt+QCB1F+woyOPmNNVgLa6djX9Z9GGTnj7qE1UzpOVJiINw==", + "dev": true, "license": "Apache-2.0", "engines": { "node": ">=20.16.0 || >=22.3.0" @@ -8853,6 +9137,7 @@ "version": "5.0.1", "resolved": "https://registry.npmjs.org/pkce-challenge/-/pkce-challenge-5.0.1.tgz", "integrity": "sha512-wQ0b/W4Fr01qtpHlqSqspcj3EhBvimsdh0KlHhH8HRZnMsEa0ea2fTULOXOS9ccQr3om+GcGRk4e+isrZWV8qQ==", + "dev": true, "license": "MIT", "engines": { "node": ">=16.20.0" @@ -8890,16 +9175,28 @@ "node": ">=18" } }, + "node_modules/pngjs": { + "version": "7.0.0", + "resolved": "https://registry.npmjs.org/pngjs/-/pngjs-7.0.0.tgz", + "integrity": "sha512-LKWqWJRhstyYo9pGvgor/ivk2w94eSjE3RGVuzLGlr3NmD8bf7RcYGze1mNdEHRP6TRP6rMuDHk5t44hnTRyow==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.19.0" + } + }, "node_modules/points-on-curve": { "version": "0.2.0", "resolved": "https://registry.npmjs.org/points-on-curve/-/points-on-curve-0.2.0.tgz", "integrity": "sha512-0mYKnYYe9ZcqMCWhUjItv/oHjvgEsfKvnUTg8sAtnHr3GVy7rGkXCb6d5cSyqrWqL4k81b9CPg3urd+T7aop3A==", + "dev": true, "license": "MIT" }, "node_modules/points-on-path": { "version": "0.2.1", "resolved": "https://registry.npmjs.org/points-on-path/-/points-on-path-0.2.1.tgz", "integrity": "sha512-25ClnWWuw7JbWZcgqY/gJ4FQWadKxGWk+3kR/7kD0tCaDtPPMj7oHu2ToLaVhfpnHrZzYby2w6tUA0eOIuUg8g==", + "dev": true, "license": "MIT", "dependencies": { "path-data-parser": "0.1.0", @@ -9187,6 +9484,7 @@ "integrity": "sha512-Qb1gy5OrP5+zDf2Bvnzdl3jsTf1qXVMazbvCoKhtKqVs4/YK4ozX4gKQJJVyNe+cajNPn0KoC0MC3FUmaHWEmQ==", "dev": true, "license": "MIT", + "peer": true, "dependencies": { "ansi-regex": "^5.0.1", "ansi-styles": "^5.0.0", @@ -9202,6 +9500,7 @@ "integrity": "sha512-Cxwpt2SfTzTtXcfOlzGEee8O+c+MmUgGrNiBcXnuWxuFJHe6a5Hz7qwhwe5OgaSYI0IJvkLqWX1ASG+cJOkEiA==", "dev": true, "license": "MIT", + "peer": true, "engines": { "node": ">=10" }, @@ -9230,6 +9529,7 @@ "version": "7.1.0", "resolved": "https://registry.npmjs.org/property-information/-/property-information-7.1.0.tgz", "integrity": "sha512-TwEZ+X+yCJmYfL7TPUOcvBZ4QfoT5YenQiJuX//0th53DE6w0xxLEtfK3iyryQFddXuvkIk51EEgrJQ0WJkOmQ==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -9240,6 +9540,7 @@ "version": "2.0.7", "resolved": "https://registry.npmjs.org/proxy-addr/-/proxy-addr-2.0.7.tgz", "integrity": "sha512-llQsMLSUDUPT44jdrU/O37qlnifitDP+ZwrmmZcoSKyLKvtZxpyV0n2/bD/N4tBAAZ/gJEdZU7KMraoK1+XYAg==", + "dev": true, "license": "MIT", "dependencies": { "forwarded": "0.2.0", @@ -9260,9 +9561,10 @@ } }, "node_modules/qs": { - "version": "6.15.0", - "resolved": "https://registry.npmjs.org/qs/-/qs-6.15.0.tgz", - "integrity": "sha512-mAZTtNCeetKMH+pSjrb76NAM8V9a05I9aBZOHztWy/UqcJdQYNsf59vrRKWnojAT9Y+GbIvoTBC++CPHqpDBhQ==", + "version": "6.15.2", + "resolved": "https://registry.npmjs.org/qs/-/qs-6.15.2.tgz", + "integrity": "sha512-Rzq0KEyX/w/tEybncDgdkZrJgVUsUMk3xjh3t5bv3S1HTAtg+uOYt72+ZfwiQwKdysThkTBdL/rTi6HDmX9Ddw==", + "dev": true, "license": "BSD-3-Clause", "dependencies": { "side-channel": "^1.1.0" @@ -9278,6 +9580,7 @@ "version": "1.2.1", "resolved": "https://registry.npmjs.org/range-parser/-/range-parser-1.2.1.tgz", "integrity": "sha512-Hrgsx+orqoygnmhFbKaHE6c296J+HTAQXoxEF6gNupROmmGJRoyzfG3ccAveqCBrwr/2yxQ5BVd/GTl5agOwSg==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.6" @@ -9287,6 +9590,7 @@ "version": "3.0.2", "resolved": "https://registry.npmjs.org/raw-body/-/raw-body-3.0.2.tgz", "integrity": "sha512-K5zQjDllxWkf7Z5xJdV0/B0WTNqx6vxG70zJE4N0kBs4LovmEYWJzQGxC9bS9RAKu3bgM40lrd5zoLJ12MQ5BA==", + "dev": true, "license": "MIT", "dependencies": { "bytes": "~3.1.2", @@ -9302,6 +9606,7 @@ "version": "0.7.2", "resolved": "https://registry.npmjs.org/iconv-lite/-/iconv-lite-0.7.2.tgz", "integrity": "sha512-im9DjEDQ55s9fL4EYzOAv0yMqmMBSZp6G0VvFyTMPKWxiSBHUj9NW/qqLmXUwXrrM7AvqSlTCfvqRb0cM8yYqw==", + "dev": true, "license": "MIT", "dependencies": { "safer-buffer": ">= 2.1.2 < 3.0.0" @@ -9342,7 +9647,8 @@ "resolved": "https://registry.npmjs.org/react-is/-/react-is-17.0.2.tgz", "integrity": "sha512-w2GsyukL62IJnlaff/nRegPQR94C/XXamvMWmSHRJ4y7Ts/4ocGRmTHvOs8PSE6pB3dWOrD/nueuU5sduBsQ4w==", "dev": true, - "license": "MIT" + "license": "MIT", + "peer": true }, "node_modules/readdirp": { "version": "4.1.2", @@ -9393,6 +9699,7 @@ "version": "7.0.2", "resolved": "https://registry.npmjs.org/rehype-highlight/-/rehype-highlight-7.0.2.tgz", "integrity": "sha512-k158pK7wdC2qL3M5NcZROZ2tR/l7zOzjxXd5VGdcfIyoijjQqpHd3JKtYSBDpDZ38UI2WJWuFAtkMDxmx5kstA==", + "dev": true, "license": "MIT", "dependencies": { "@types/hast": "^3.0.0", @@ -9430,6 +9737,7 @@ "version": "10.0.1", "resolved": "https://registry.npmjs.org/rehype-stringify/-/rehype-stringify-10.0.1.tgz", "integrity": "sha512-k9ecfXHmIPuFVI61B9DeLPN0qFHfawM6RsuX48hoqlaKSF61RskNjSm1lI8PhBEM0MRdLxVVm4WmTqJQccH9mA==", + "dev": true, "license": "MIT", "dependencies": { "@types/hast": "^3.0.0", @@ -9445,6 +9753,7 @@ "version": "15.0.1", "resolved": "https://registry.npmjs.org/remark/-/remark-15.0.1.tgz", "integrity": "sha512-Eht5w30ruCXgFmxVUSlNWQ9iiimq07URKeFS3hNc8cUWy1llX4KDWfyEDZRycMc+znsN9Ux5/tJ/BFdgdOwA3A==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -9461,6 +9770,7 @@ "version": "4.0.0", "resolved": "https://registry.npmjs.org/remark-breaks/-/remark-breaks-4.0.0.tgz", "integrity": "sha512-IjEjJOkH4FuJvHZVIW0QCDWxcG96kCq7An/KVH2NfJe6rKZU2AsHeB3OEjPNRxi4QC34Xdx7I2KGYn6IpT7gxQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -9476,6 +9786,7 @@ "version": "4.0.1", "resolved": "https://registry.npmjs.org/remark-gfm/-/remark-gfm-4.0.1.tgz", "integrity": "sha512-1quofZ2RQ9EWdeN34S79+KExV1764+wCUGop5CPL1WGdD0ocPpu91lzPGbwWMECpEpd42kJGQwzRfyov9j4yNg==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -9494,6 +9805,7 @@ "version": "16.0.1", "resolved": "https://registry.npmjs.org/remark-html/-/remark-html-16.0.1.tgz", "integrity": "sha512-B9JqA5i0qZe0Nsf49q3OXyGvyXuZFDzAP2iOFLEumymuYJITVpiH1IgsTEwTpdptDmZlMDMWeDmSawdaJIGCXQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -9528,6 +9840,7 @@ "version": "11.0.0", "resolved": "https://registry.npmjs.org/remark-parse/-/remark-parse-11.0.0.tgz", "integrity": "sha512-FCxlKLNGknS5ba/1lmpYijMUzX2esxW5xQqjWxw2eHFfS2MSdaHVINFmhjo+qN1WhZhNimq0dZATN9pH0IDrpA==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -9544,6 +9857,7 @@ "version": "11.1.2", "resolved": "https://registry.npmjs.org/remark-rehype/-/remark-rehype-11.1.2.tgz", "integrity": "sha512-Dh7l57ianaEoIpzbp0PC9UKAdCSVklD8E5Rpw7ETfbTl3FqcOOgq5q2LVDhgGCkaBv7p24JXikPdvhhmHvKMsw==", + "dev": true, "license": "MIT", "dependencies": { "@types/hast": "^3.0.0", @@ -9561,6 +9875,7 @@ "version": "11.0.0", "resolved": "https://registry.npmjs.org/remark-stringify/-/remark-stringify-11.0.0.tgz", "integrity": "sha512-1OSmLd3awB/t8qdoEOMazZkNsfVTeY4fTsgzcQFdXNq8ToTN4ZGwrMnlda4K6smTFKD+GRV6O48i6Z4iKgPPpw==", + "dev": true, "license": "MIT", "dependencies": { "@types/mdast": "^4.0.0", @@ -9576,6 +9891,7 @@ "version": "2.0.2", "resolved": "https://registry.npmjs.org/require-from-string/-/require-from-string-2.0.2.tgz", "integrity": "sha512-Xf0nWe6RseziFMu+Ap9biiUbmplq6S9/p+7w7YXP/JBHhrUDDUhwa+vANyubuqfZWTveU//DYVGsDG7RKL/vEw==", + "dev": true, "license": "MIT", "engines": { "node": ">=0.10.0" @@ -9602,6 +9918,7 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/robust-predicates/-/robust-predicates-3.0.3.tgz", "integrity": "sha512-NS3levdsRIUOmiJ8FZWCP7LG3QpJyrs/TE0Zpf1yvZu8cAJJ6QMW92H1c7kWpdIHo8RvmLxN/o2JXTKHp74lUA==", + "dev": true, "license": "Unlicense" }, "node_modules/rollup": { @@ -9653,6 +9970,7 @@ "version": "4.6.6", "resolved": "https://registry.npmjs.org/roughjs/-/roughjs-4.6.6.tgz", "integrity": "sha512-ZUz/69+SYpFN/g/lUlo2FXcIjRkSu3nDarreVdGGndHEBJ6cXPdKguS8JGxwj5HA5xIbVKSmLgr5b3AWxtRfvQ==", + "dev": true, "license": "MIT", "dependencies": { "hachure-fill": "^0.5.2", @@ -9665,6 +9983,7 @@ "version": "2.2.0", "resolved": "https://registry.npmjs.org/router/-/router-2.2.0.tgz", "integrity": "sha512-nLTrUKm2UyiL7rlhapu/Zl45FwNgkZGaCpZbIHajDYgwlJCOzLSk+cIPAnsEqV955GjILJnKbdQC1nVPz+gAYQ==", + "dev": true, "license": "MIT", "dependencies": { "debug": "^4.4.0", @@ -9694,6 +10013,7 @@ "version": "0.25.0", "resolved": "https://registry.npmjs.org/runed/-/runed-0.25.0.tgz", "integrity": "sha512-7+ma4AG9FT2sWQEA0Egf6mb7PBT2vHyuHail1ie8ropfSjvZGtEAx8YTmUjv/APCsdRRxEVvArNjALk9zFSOrg==", + "dev": true, "funding": [ "https://github.com/sponsors/huntabyte", "https://github.com/sponsors/tglide" @@ -9709,6 +10029,7 @@ "version": "1.3.3", "resolved": "https://registry.npmjs.org/rw/-/rw-1.3.3.tgz", "integrity": "sha512-PdhdWy89SiZogBLaw42zdeqtRJ//zFd2PgQavcICDUgJT5oW10QCRKbJ6bg4r0/UY2M6BWd5tkxuGFRvCkgfHQ==", + "dev": true, "license": "BSD-3-Clause" }, "node_modules/sade": { @@ -9735,6 +10056,7 @@ "version": "2.1.2", "resolved": "https://registry.npmjs.org/safer-buffer/-/safer-buffer-2.1.2.tgz", "integrity": "sha512-YZo3K82SD7Riyi0E1EQPojLz7kpepnSQI9IyPbHHg1XXXevb5dJI7tpyN2ADxGcQbHG7vcyRHk0cbwqcQriUtg==", + "dev": true, "license": "MIT" }, "node_modules/sass": { @@ -9796,6 +10118,7 @@ "version": "1.2.1", "resolved": "https://registry.npmjs.org/send/-/send-1.2.1.tgz", "integrity": "sha512-1gnZf7DFcoIcajTjTwjwuDjzuz4PPcY2StKPlsGAQ1+YH20IRVrBaXSWmdjowTJ6u8Rc01PoYOGHXfP1mYcZNQ==", + "dev": true, "license": "MIT", "dependencies": { "debug": "^4.4.3", @@ -9822,6 +10145,7 @@ "version": "2.2.1", "resolved": "https://registry.npmjs.org/serve-static/-/serve-static-2.2.1.tgz", "integrity": "sha512-xRXBn0pPqQTVQiC8wyQrKs2MOlX24zQ0POGaj0kultvoOCstBQM5yvOhAVSUwOMjQtTvsPWoNCHfPGwaaQJhTw==", + "dev": true, "license": "MIT", "dependencies": { "encodeurl": "^2.0.0", @@ -9848,12 +10172,14 @@ "version": "1.2.0", "resolved": "https://registry.npmjs.org/setprototypeof/-/setprototypeof-1.2.0.tgz", "integrity": "sha512-E5LDX7Wrp85Kil5bhZv46j8jOeboKq5JMmYM3gVGdGH8xFpPWXUMsNrlODCrkoxMEeNi/XZIwuRvY4XNwYMJpw==", + "dev": true, "license": "ISC" }, "node_modules/shebang-command": { "version": "2.0.0", "resolved": "https://registry.npmjs.org/shebang-command/-/shebang-command-2.0.0.tgz", "integrity": "sha512-kHxr2zZpYtdmrN1qDjrrX/Z1rR1kG8Dx+gkpK1G4eXmvXswmcE1hTWBWYUzlraYw1/yZp6YuDY77YtvbN0dmDA==", + "dev": true, "license": "MIT", "dependencies": { "shebang-regex": "^3.0.0" @@ -9866,6 +10192,7 @@ "version": "3.0.0", "resolved": "https://registry.npmjs.org/shebang-regex/-/shebang-regex-3.0.0.tgz", "integrity": "sha512-7++dFhtcx3353uBaq8DDR4NuxBetBzC7ZQOhmTQInHEd6bSrXdiEyzCvG07Z44UYdLShWUyXt5M/yhz8ekcb1A==", + "dev": true, "license": "MIT", "engines": { "node": ">=8" @@ -9875,6 +10202,7 @@ "version": "1.1.0", "resolved": "https://registry.npmjs.org/side-channel/-/side-channel-1.1.0.tgz", "integrity": "sha512-ZX99e6tRweoUXqR+VBrslhda51Nh5MTQwou5tnUDgbtyM0dBgmhEDtWGP/xbKn6hqfPRHujUNwz5fy/wbbhnpw==", + "dev": true, "license": "MIT", "dependencies": { "es-errors": "^1.3.0", @@ -9894,6 +10222,7 @@ "version": "1.0.0", "resolved": "https://registry.npmjs.org/side-channel-list/-/side-channel-list-1.0.0.tgz", "integrity": "sha512-FCLHtRD/gnpCiCHEiJLOwdmFP+wzCmDEkc9y7NsYxeF4u7Btsn1ZuwgwJGxImImHicJArLP4R0yX4c2KCrMrTA==", + "dev": true, "license": "MIT", "dependencies": { "es-errors": "^1.3.0", @@ -9910,6 +10239,7 @@ "version": "1.0.1", "resolved": "https://registry.npmjs.org/side-channel-map/-/side-channel-map-1.0.1.tgz", "integrity": "sha512-VCjCNfgMsby3tTdo02nbjtM/ewra6jPHmpThenkTYh8pG9ucZ/1P8So4u4FGBek/BjpOVsDCMoLA/iuBKIFXRA==", + "dev": true, "license": "MIT", "dependencies": { "call-bound": "^1.0.2", @@ -9928,6 +10258,7 @@ "version": "1.0.2", "resolved": "https://registry.npmjs.org/side-channel-weakmap/-/side-channel-weakmap-1.0.2.tgz", "integrity": "sha512-WPS/HvHQTYnHisLo9McqBHOJk2FkHO/tlpvldyrnem4aeQp4hai3gythswg6p01oSoTl58rcpiFAjF2br2Ak2A==", + "dev": true, "license": "MIT", "dependencies": { "call-bound": "^1.0.2", @@ -9950,23 +10281,10 @@ "dev": true, "license": "ISC" }, - "node_modules/signal-exit": { - "version": "4.1.0", - "resolved": "https://registry.npmjs.org/signal-exit/-/signal-exit-4.1.0.tgz", - "integrity": "sha512-bzyZ1e88w9O1iNJbKnOlvYTrWPDl46O1bG0D3XInv+9tkPrxrN8jUUTiFlDkkmKWgn1M6CfIA13SuGqOa9Korw==", - "dev": true, - "license": "ISC", - "engines": { - "node": ">=14" - }, - "funding": { - "url": "https://github.com/sponsors/isaacs" - } - }, "node_modules/sirv": { - "version": "3.0.1", - "resolved": "https://registry.npmjs.org/sirv/-/sirv-3.0.1.tgz", - "integrity": "sha512-FoqMu0NCGBLCcAkS1qA+XJIQTR6/JHfQXl+uGteNCQ76T91DMUjPa9xfmeqMY3z80nLSg9yQmNjK0Px6RWsH/A==", + "version": "3.0.2", + "resolved": "https://registry.npmjs.org/sirv/-/sirv-3.0.2.tgz", + "integrity": "sha512-2wcC/oGxHis/BoHkkPwldgiPSYcpZK3JU28WoMVv55yHJgcZ8rlXvuG9iZggz+sU1d4bRgIGASwyWqjxu3FM0g==", "dev": true, "license": "MIT", "dependencies": { @@ -10002,6 +10320,7 @@ "version": "2.0.2", "resolved": "https://registry.npmjs.org/space-separated-tokens/-/space-separated-tokens-2.0.2.tgz", "integrity": "sha512-PEGlAwrG8yXGXRjW32fGbg66JAlOAwbObuqVoJpv/mRgoWDQfgH1wDPvtzWyUSNAXBGSk8h755YDbbcEy3SH2Q==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -10019,15 +10338,16 @@ "version": "2.0.2", "resolved": "https://registry.npmjs.org/statuses/-/statuses-2.0.2.tgz", "integrity": "sha512-DvEy55V3DB7uknRo+4iOGT5fP1slR8wQohVdknigZPMpMstaKJQWhwiYBACJE3Ul2pTnATihhBYnRhZQHGBiRw==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.8" } }, "node_modules/std-env": { - "version": "3.9.0", - "resolved": "https://registry.npmjs.org/std-env/-/std-env-3.9.0.tgz", - "integrity": "sha512-UGvjygr6F6tpH7o2qyqR6QYpwraIjKSdtzyBdyytFOHmPZY917kwdwLG0RbOjWOnKmnm3PeHjaoLLMie7kPLQw==", + "version": "4.1.0", + "resolved": "https://registry.npmjs.org/std-env/-/std-env-4.1.0.tgz", + "integrity": "sha512-Rq7ybcX2RuC55r9oaPVEW7/xu3tj8u4GeBYHBWCychFtzMIr86A7e3PPEBPT37sHStKX3+TiX/Fr/ACmJLVlLQ==", "dev": true, "license": "MIT" }, @@ -10067,64 +10387,11 @@ } } }, - "node_modules/string-width": { - "version": "5.1.2", - "resolved": "https://registry.npmjs.org/string-width/-/string-width-5.1.2.tgz", - "integrity": "sha512-HnLOCR3vjcY8beoNLtcjZ5/nxn2afmME6lhrDrebokqMap+XbeW8n9TXpPDOqdGK5qcI3oT0GKTW6wC7EMiVqA==", - "dev": true, - "license": "MIT", - "dependencies": { - "eastasianwidth": "^0.2.0", - "emoji-regex": "^9.2.2", - "strip-ansi": "^7.0.1" - }, - "engines": { - "node": ">=12" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/string-width-cjs": { - "name": "string-width", - "version": "4.2.3", - "resolved": "https://registry.npmjs.org/string-width/-/string-width-4.2.3.tgz", - "integrity": "sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==", - "dev": true, - "license": "MIT", - "dependencies": { - "emoji-regex": "^8.0.0", - "is-fullwidth-code-point": "^3.0.0", - "strip-ansi": "^6.0.1" - }, - "engines": { - "node": ">=8" - } - }, - "node_modules/string-width-cjs/node_modules/emoji-regex": { - "version": "8.0.0", - "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-8.0.0.tgz", - "integrity": "sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A==", - "dev": true, - "license": "MIT" - }, - "node_modules/string-width-cjs/node_modules/strip-ansi": { - "version": "6.0.1", - "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-6.0.1.tgz", - "integrity": "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==", - "dev": true, - "license": "MIT", - "dependencies": { - "ansi-regex": "^5.0.1" - }, - "engines": { - "node": ">=8" - } - }, "node_modules/stringify-entities": { "version": "4.0.4", "resolved": "https://registry.npmjs.org/stringify-entities/-/stringify-entities-4.0.4.tgz", "integrity": "sha512-IwfBptatlO+QCJUo19AqvrPNqlVMpW9YEL2LIVY+Rpv2qsjCGxaDLNRgeGsQWJhfItebuJhsGSLjaBbNSQ+ieg==", + "dev": true, "license": "MIT", "dependencies": { "character-entities-html4": "^2.0.0", @@ -10151,20 +10418,6 @@ "url": "https://github.com/chalk/strip-ansi?sponsor=1" } }, - "node_modules/strip-ansi-cjs": { - "name": "strip-ansi", - "version": "6.0.1", - "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-6.0.1.tgz", - "integrity": "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==", - "dev": true, - "license": "MIT", - "dependencies": { - "ansi-regex": "^5.0.1" - }, - "engines": { - "node": ">=8" - } - }, "node_modules/strip-ansi/node_modules/ansi-regex": { "version": "6.1.0", "resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-6.1.0.tgz", @@ -10204,30 +10457,11 @@ "url": "https://github.com/sponsors/sindresorhus" } }, - "node_modules/strip-literal": { - "version": "3.0.0", - "resolved": "https://registry.npmjs.org/strip-literal/-/strip-literal-3.0.0.tgz", - "integrity": "sha512-TcccoMhJOM3OebGhSBEmp3UZ2SfDMZUEBdRA/9ynfLi8yYajyWX3JiXArcJt4Umh4vISpspkQIY8ZZoCqjbviA==", - "dev": true, - "license": "MIT", - "dependencies": { - "js-tokens": "^9.0.1" - }, - "funding": { - "url": "https://github.com/sponsors/antfu" - } - }, - "node_modules/strip-literal/node_modules/js-tokens": { - "version": "9.0.1", - "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-9.0.1.tgz", - "integrity": "sha512-mxa9E9ITFOt0ban3j6L5MpjwegGz6lBQmM1IJkWeBZGcMxto50+eWdjC/52xDbS2vy0k7vIMK0Fe2wfL9OQSpQ==", - "dev": true, - "license": "MIT" - }, "node_modules/style-to-object": { "version": "1.0.9", "resolved": "https://registry.npmjs.org/style-to-object/-/style-to-object-1.0.9.tgz", "integrity": "sha512-G4qppLgKu/k6FwRpHiGiKPaPTFcG3g4wNVX/Qsfu+RqQM30E7Tyu/TEgxcL9PNLF5pdRLwQdE3YKKf+KF2Dzlw==", + "dev": true, "license": "MIT", "dependencies": { "inline-style-parser": "0.2.4" @@ -10237,6 +10471,7 @@ "version": "4.4.0", "resolved": "https://registry.npmjs.org/stylis/-/stylis-4.4.0.tgz", "integrity": "sha512-5Z9ZpRzfuH6l/UAvCPAPUo3665Nk2wLaZU3x+TLHKVzIz33+sbJqbtrYoC3KD4/uVOr2Zp+L0LySezP9OHV9yA==", + "dev": true, "license": "MIT" }, "node_modules/supports-color": { @@ -10256,6 +10491,7 @@ "version": "5.55.7", "resolved": "https://registry.npmjs.org/svelte/-/svelte-5.55.7.tgz", "integrity": "sha512-ymI5ykLPwIHW839E053FQbI1G+jnRFJEw3Kv5Y4njixVWywQBx+NUFpkkKyk5LIb36Fg9DVXSYpqiGekLD0hyw==", + "dev": true, "license": "MIT", "dependencies": { "@jridgewell/remapping": "^2.3.4", @@ -10389,6 +10625,7 @@ "version": "1.0.5", "resolved": "https://registry.npmjs.org/svelte-sonner/-/svelte-sonner-1.0.5.tgz", "integrity": "sha512-9dpGPFqKb/QWudYqGnEz93vuY+NgCEvyNvxoCLMVGw6sDN/3oVeKV1xiEirW2E1N3vJEyj5imSBNOGltQHA7mg==", + "dev": true, "license": "MIT", "dependencies": { "runed": "^0.28.0" @@ -10401,6 +10638,7 @@ "version": "0.28.0", "resolved": "https://registry.npmjs.org/runed/-/runed-0.28.0.tgz", "integrity": "sha512-k2xx7RuO9hWcdd9f+8JoBeqWtYrm5CALfgpkg2YDB80ds/QE4w0qqu34A7fqiAwiBBSBQOid7TLxwxVC27ymWQ==", + "dev": true, "funding": [ "https://github.com/sponsors/huntabyte", "https://github.com/sponsors/tglide" @@ -10417,6 +10655,7 @@ "version": "0.7.1", "resolved": "https://registry.npmjs.org/svelte-toolbelt/-/svelte-toolbelt-0.7.1.tgz", "integrity": "sha512-HcBOcR17Vx9bjaOceUvxkY3nGmbBmCBBbuWLLEWO6jtmWH8f/QoWmbyUfQZrpDINH39en1b8mptfPQT9VKQ1xQ==", + "dev": true, "funding": [ "https://github.com/sponsors/huntabyte" ], @@ -10437,6 +10676,7 @@ "version": "0.23.4", "resolved": "https://registry.npmjs.org/runed/-/runed-0.23.4.tgz", "integrity": "sha512-9q8oUiBYeXIDLWNK5DfCWlkL0EW3oGbk845VdKlPeia28l751VpfesaB/+7pI6rnbx1I6rqoZ2fZxptOJLxILA==", + "dev": true, "funding": [ "https://github.com/sponsors/huntabyte", "https://github.com/sponsors/tglide" @@ -10452,6 +10692,7 @@ "version": "5.3.1", "resolved": "https://registry.npmjs.org/aria-query/-/aria-query-5.3.1.tgz", "integrity": "sha512-Z/ZeOgVl7bcSYZ/u/rh0fOpvEpq//LZmdbkXyc7syVzjPAhfOa9ebsdTSjEBDU4vs5nC98Kfduj1uFo0qyET3g==", + "dev": true, "license": "Apache-2.0", "engines": { "node": ">= 0.4" @@ -10461,6 +10702,7 @@ "version": "2.2.4", "resolved": "https://registry.npmjs.org/esrap/-/esrap-2.2.4.tgz", "integrity": "sha512-suICpxAmZ9A8bzJjEl/+rLJiDKC0X4gYWUxT6URAWBLvlXmtbZd5ySMu/N2ZGEtMCAmflUDPSehrP9BQcsGcSg==", + "dev": true, "license": "MIT", "dependencies": { "@jridgewell/sourcemap-codec": "^1.4.15", @@ -10471,6 +10713,7 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/is-reference/-/is-reference-3.0.3.tgz", "integrity": "sha512-ixkJoqQvAP88E6wLydLGGqCJsrFUnqoH6HnaczB8XmDH1oaWU+xxdptvikTgaEhtZ53Ky6YXiBuUI2WXLMCwjw==", + "dev": true, "license": "MIT", "dependencies": { "@types/estree": "^1.0.6" @@ -10563,47 +10806,6 @@ "node": ">=18" } }, - "node_modules/test-exclude": { - "version": "7.0.1", - "resolved": "https://registry.npmjs.org/test-exclude/-/test-exclude-7.0.1.tgz", - "integrity": "sha512-pFYqmTw68LXVjeWJMST4+borgQP2AyMNbg1BpZh9LbyhUeNkeaPF9gzfPGUAnSMV3qPYdWUwDIjjCLiSDOl7vg==", - "dev": true, - "license": "ISC", - "dependencies": { - "@istanbuljs/schema": "^0.1.2", - "glob": "^10.4.1", - "minimatch": "^9.0.4" - }, - "engines": { - "node": ">=18" - } - }, - "node_modules/test-exclude/node_modules/brace-expansion": { - "version": "2.0.3", - "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-2.0.3.tgz", - "integrity": "sha512-MCV/fYJEbqx68aE58kv2cA/kiky1G8vux3OR6/jbS+jIMe/6fJWa0DTzJU7dqijOWYwHi1t29FlfYI9uytqlpA==", - "dev": true, - "license": "MIT", - "dependencies": { - "balanced-match": "^1.0.0" - } - }, - "node_modules/test-exclude/node_modules/minimatch": { - "version": "9.0.9", - "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-9.0.9.tgz", - "integrity": "sha512-OBwBN9AL4dqmETlpS2zasx+vTeWclWzkblfZk7KTA5j3jeOONz/tRCnZomUyvNg83wL5Zv9Ss6HMJXAgL8R2Yg==", - "dev": true, - "license": "ISC", - "dependencies": { - "brace-expansion": "^2.0.2" - }, - "engines": { - "node": ">=16 || 14 >=14.17" - }, - "funding": { - "url": "https://github.com/sponsors/isaacs" - } - }, "node_modules/tiny-invariant": { "version": "1.3.3", "resolved": "https://registry.npmjs.org/tiny-invariant/-/tiny-invariant-1.3.3.tgz", @@ -10619,11 +10821,14 @@ "license": "MIT" }, "node_modules/tinyexec": { - "version": "0.3.2", - "resolved": "https://registry.npmjs.org/tinyexec/-/tinyexec-0.3.2.tgz", - "integrity": "sha512-KQQR9yN7R5+OSwaK0XQoj22pwHoTlgYqmUscPYoknOoWCWfj/5/ABTMRi69FrKU5ffPVh5QcFikpWJI/P1ocHA==", + "version": "1.2.4", + "resolved": "https://registry.npmjs.org/tinyexec/-/tinyexec-1.2.4.tgz", + "integrity": "sha512-SHf/r48b7vOrjve9PxJo3MN5v5yuyjHvdUcrQffT3WXMUfnGmHDVbC4k3sHJaJTgZCwpUplIaAo5ANtMyp3YHg==", "dev": true, - "license": "MIT" + "license": "MIT", + "engines": { + "node": ">=18" + } }, "node_modules/tinyglobby": { "version": "0.2.15", @@ -10642,16 +10847,6 @@ "url": "https://github.com/sponsors/SuperchupuDev" } }, - "node_modules/tinypool": { - "version": "1.1.1", - "resolved": "https://registry.npmjs.org/tinypool/-/tinypool-1.1.1.tgz", - "integrity": "sha512-Zba82s87IFq9A9XmjiX5uZA/ARWDrB03OHlq+Vw1fSdt0I+4/Kutwy8BP4Y/y/aORMo61FQ0vIb5j44vSo5Pkg==", - "dev": true, - "license": "MIT", - "engines": { - "node": "^18.0.0 || >=20.0.0" - } - }, "node_modules/tinyrainbow": { "version": "2.0.0", "resolved": "https://registry.npmjs.org/tinyrainbow/-/tinyrainbow-2.0.0.tgz", @@ -10690,6 +10885,7 @@ "version": "1.0.1", "resolved": "https://registry.npmjs.org/toidentifier/-/toidentifier-1.0.1.tgz", "integrity": "sha512-o5sSPKEkg/DIQNmH43V0/uerLrpzVedkUh8tGNvaeXpfpuwjKenlSox/2O/BTlZUtEe+JG7s5YhEz608PlAHRA==", + "dev": true, "license": "MIT", "engines": { "node": ">=0.6" @@ -10709,6 +10905,7 @@ "version": "3.0.1", "resolved": "https://registry.npmjs.org/trim-lines/-/trim-lines-3.0.1.tgz", "integrity": "sha512-kRj8B+YHZCc9kQYdWfJB2/oUl9rA99qbowYYBtr4ui4mZyAQ2JpvVBd/6U2YloATfqBhBTSMhTpgBHtU0Mf3Rg==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -10719,6 +10916,7 @@ "version": "2.2.0", "resolved": "https://registry.npmjs.org/trough/-/trough-2.2.0.tgz", "integrity": "sha512-tmMpK00BjZiUyVyvrBK7knerNgmgvcV/KLVyuma/SC+TQN167GrMRciANTz09+k3zW8L8t60jWO1GpfkZdjTaw==", + "dev": true, "license": "MIT", "funding": { "type": "github", @@ -10742,6 +10940,7 @@ "version": "2.2.0", "resolved": "https://registry.npmjs.org/ts-dedent/-/ts-dedent-2.2.0.tgz", "integrity": "sha512-q5W7tVM71e2xjHZTlgfTDoPF/SmqKG5hddq9SzR49CH2hayqRKJtQ4mtRlSxKaJlR/+9rEM+mnBHf7I2/BQcpQ==", + "dev": true, "license": "MIT", "engines": { "node": ">=6.10" @@ -10794,6 +10993,7 @@ "version": "2.0.1", "resolved": "https://registry.npmjs.org/type-is/-/type-is-2.0.1.tgz", "integrity": "sha512-OZs6gsjF4vMp32qrCbiVSkrFmXtG/AZhY3t0iAMrMBiAZyV9oALtXO8hsrHbMXF9x6L3grlFuwW2oAz7cav+Gw==", + "dev": true, "license": "MIT", "dependencies": { "content-type": "^1.0.5", @@ -10853,6 +11053,7 @@ "version": "11.0.5", "resolved": "https://registry.npmjs.org/unified/-/unified-11.0.5.tgz", "integrity": "sha512-xKvGhPWw3k84Qjh8bI3ZeJjqnyadK+GEFtazSfZv/rKeTkTjOJho6mFqh2SM96iIcZokxiOpg78GazTSg8+KHA==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "^3.0.0", @@ -10872,6 +11073,7 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/union": { @@ -10890,6 +11092,7 @@ "version": "5.0.0", "resolved": "https://registry.npmjs.org/unist-util-find-after/-/unist-util-find-after-5.0.0.tgz", "integrity": "sha512-amQa0Ep2m6hE2g72AugUItjbuM8X8cGQnFoHk0pGfrFeT9GZhzN5SW8nRsiGKK7Aif4CrACPENkA6P/Lw6fHGQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "^3.0.0", @@ -10904,12 +11107,14 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/unist-util-is": { "version": "6.0.0", "resolved": "https://registry.npmjs.org/unist-util-is/-/unist-util-is-6.0.0.tgz", "integrity": "sha512-2qCTHimwdxLfz+YzdGfkqNlH0tLi9xjTnHddPmJwtIG9MGsdbutfTc4P+haPD7l7Cjxf/WZj+we5qfVPvvxfYw==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "^3.0.0" @@ -10923,12 +11128,14 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/unist-util-position": { "version": "5.0.0", "resolved": "https://registry.npmjs.org/unist-util-position/-/unist-util-position-5.0.0.tgz", "integrity": "sha512-fucsC7HjXvkB5R3kTCO7kUjRdrS0BJt3M/FPxmHMBOm8JQi2BsHAHFsy27E0EolP8rp0NzXsJ+jNPyDWvOJZPA==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "^3.0.0" @@ -10942,6 +11149,7 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/unist-util-remove-position": { @@ -10984,6 +11192,7 @@ "version": "5.0.0", "resolved": "https://registry.npmjs.org/unist-util-visit/-/unist-util-visit-5.0.0.tgz", "integrity": "sha512-MR04uvD+07cwl/yhVuVWAtw+3GOR/knlL55Nd/wAdblk27GCVt3lqpTivy/tkJcZoNPzTwS1Y+KMojlLDhoTzg==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "^3.0.0", @@ -10999,6 +11208,7 @@ "version": "6.0.1", "resolved": "https://registry.npmjs.org/unist-util-visit-parents/-/unist-util-visit-parents-6.0.1.tgz", "integrity": "sha512-L/PqWzfTP9lzzEa6CKs0k2nARxTdZduw3zyh8d2NVBnsyvHjSX4TWse388YrrQKbvI8w20fGjGlhgT96WwKykw==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "^3.0.0", @@ -11013,12 +11223,14 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/unist-util-visit/node_modules/@types/unist": { "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/universalify": { @@ -11035,6 +11247,7 @@ "version": "1.0.0", "resolved": "https://registry.npmjs.org/unpipe/-/unpipe-1.0.0.tgz", "integrity": "sha512-pjy2bYhSsufwWlKwPc+l3cN7+wuJlK6uz0YdJEOlQDbl6jo/YlPi4mb8agUkVC8BF7V8NuzeyPNqRksA3hztKQ==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.8" @@ -11094,6 +11307,7 @@ "version": "13.0.2", "resolved": "https://registry.npmjs.org/uuid/-/uuid-13.0.2.tgz", "integrity": "sha512-vzi9uRZ926x4XV73S/4qQaTwPXM2JBj6/6lI/byHH1jOpCzb0zDbfytgA9LcN/hzb2l7WQSQnxITOVx5un/wGw==", + "dev": true, "funding": [ "https://github.com/sponsors/broofa", "https://github.com/sponsors/ctavan" @@ -11107,6 +11321,7 @@ "version": "1.1.2", "resolved": "https://registry.npmjs.org/vary/-/vary-1.1.2.tgz", "integrity": "sha512-BNGbWLfd0eUPabhkXUVm0j8uuvREyTh5ovRa/dyow/BqAbZJyC+5fU+IzQOzmAKzYqYRAISoRhdQr3eIZ/PXqg==", + "dev": true, "license": "MIT", "engines": { "node": ">= 0.8" @@ -11116,6 +11331,7 @@ "version": "6.0.3", "resolved": "https://registry.npmjs.org/vfile/-/vfile-6.0.3.tgz", "integrity": "sha512-KzIbH/9tXat2u30jf+smMwFCsno4wHVdNmzFyL+T/L3UGqqk6JKfVqOFOZEpZSHADH1k40ab6NUIXZq422ov3Q==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "^3.0.0", @@ -11167,12 +11383,14 @@ "version": "3.0.3", "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, "license": "MIT" }, "node_modules/vfile/node_modules/unist-util-stringify-position": { "version": "4.0.0", "resolved": "https://registry.npmjs.org/unist-util-stringify-position/-/unist-util-stringify-position-4.0.0.tgz", "integrity": "sha512-0ASV06AAoKCDkS2+xw5RXJywruurpbC4JZSm7nr7MOt1ojAzvyyaO+UxZf18j8FCF6kmzCZKcAgN/yu2gm2XgQ==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "^3.0.0" @@ -11186,6 +11404,7 @@ "version": "4.0.2", "resolved": "https://registry.npmjs.org/vfile-message/-/vfile-message-4.0.2.tgz", "integrity": "sha512-jRDZ1IMLttGj41KcZvlrYAaI3CfqpLpfpf+Mfig13viT6NKvRzWZ+lXz0Y5D60w6uJIBAOGq9mSHf0gktF0duw==", + "dev": true, "license": "MIT", "dependencies": { "@types/unist": "^3.0.0", @@ -11271,29 +11490,6 @@ } } }, - "node_modules/vite-node": { - "version": "3.2.4", - "resolved": "https://registry.npmjs.org/vite-node/-/vite-node-3.2.4.tgz", - "integrity": "sha512-EbKSKh+bh1E1IFxeO0pg1n4dvoOTt0UDiXMd/qn++r98+jPO1xtJilvXldeuQ8giIB5IkpjCgMleHMNEsGH6pg==", - "dev": true, - "license": "MIT", - "dependencies": { - "cac": "^6.7.14", - "debug": "^4.4.1", - "es-module-lexer": "^1.7.0", - "pathe": "^2.0.3", - "vite": "^5.0.0 || ^6.0.0 || ^7.0.0-0" - }, - "bin": { - "vite-node": "vite-node.mjs" - }, - "engines": { - "node": "^18.0.0 || ^20.0.0 || >=22.0.0" - }, - "funding": { - "url": "https://opencollective.com/vitest" - } - }, "node_modules/vite-plugin-devtools-json": { "version": "0.2.1", "resolved": "https://registry.npmjs.org/vite-plugin-devtools-json/-/vite-plugin-devtools-json-0.2.1.tgz", @@ -11357,65 +11553,79 @@ } }, "node_modules/vitest": { - "version": "3.2.4", - "resolved": "https://registry.npmjs.org/vitest/-/vitest-3.2.4.tgz", - "integrity": "sha512-LUCP5ev3GURDysTWiP47wRRUpLKMOfPh+yKTx3kVIEiu5KOMeqzpnYNsKyOoVrULivR8tLcks4+lga33Whn90A==", - "dev": true, - "license": "MIT", - "dependencies": { - "@types/chai": "^5.2.2", - "@vitest/expect": "3.2.4", - "@vitest/mocker": "3.2.4", - "@vitest/pretty-format": "^3.2.4", - "@vitest/runner": "3.2.4", - "@vitest/snapshot": "3.2.4", - "@vitest/spy": "3.2.4", - "@vitest/utils": "3.2.4", - "chai": "^5.2.0", - "debug": "^4.4.1", - "expect-type": "^1.2.1", - "magic-string": "^0.30.17", + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/vitest/-/vitest-4.1.8.tgz", + "integrity": "sha512-flY6ScbCIt9HThs+C5HS7jvGOB560DJtk/Z15IQROTA6zEy49Nh8T/dofWTQL+n3vswqn87sbJNiuqw1SDp5Ig==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/expect": "4.1.8", + "@vitest/mocker": "4.1.8", + "@vitest/pretty-format": "4.1.8", + "@vitest/runner": "4.1.8", + "@vitest/snapshot": "4.1.8", + "@vitest/spy": "4.1.8", + "@vitest/utils": "4.1.8", + "es-module-lexer": "^2.0.0", + "expect-type": "^1.3.0", + "magic-string": "^0.30.21", + "obug": "^2.1.1", "pathe": "^2.0.3", - "picomatch": "^4.0.2", - "std-env": "^3.9.0", + "picomatch": "^4.0.3", + "std-env": "^4.0.0-rc.1", "tinybench": "^2.9.0", - "tinyexec": "^0.3.2", - "tinyglobby": "^0.2.14", - "tinypool": "^1.1.1", - "tinyrainbow": "^2.0.0", - "vite": "^5.0.0 || ^6.0.0 || ^7.0.0-0", - "vite-node": "3.2.4", + "tinyexec": "^1.0.2", + "tinyglobby": "^0.2.15", + "tinyrainbow": "^3.1.0", + "vite": "^6.0.0 || ^7.0.0 || ^8.0.0", "why-is-node-running": "^2.3.0" }, "bin": { "vitest": "vitest.mjs" }, "engines": { - "node": "^18.0.0 || ^20.0.0 || >=22.0.0" + "node": "^20.0.0 || ^22.0.0 || >=24.0.0" }, "funding": { "url": "https://opencollective.com/vitest" }, "peerDependencies": { "@edge-runtime/vm": "*", - "@types/debug": "^4.1.12", - "@types/node": "^18.0.0 || ^20.0.0 || >=22.0.0", - "@vitest/browser": "3.2.4", - "@vitest/ui": "3.2.4", + "@opentelemetry/api": "^1.9.0", + "@types/node": "^20.0.0 || ^22.0.0 || >=24.0.0", + "@vitest/browser-playwright": "4.1.8", + "@vitest/browser-preview": "4.1.8", + "@vitest/browser-webdriverio": "4.1.8", + "@vitest/coverage-istanbul": "4.1.8", + "@vitest/coverage-v8": "4.1.8", + "@vitest/ui": "4.1.8", "happy-dom": "*", - "jsdom": "*" + "jsdom": "*", + "vite": "^6.0.0 || ^7.0.0 || ^8.0.0" }, "peerDependenciesMeta": { "@edge-runtime/vm": { "optional": true }, - "@types/debug": { + "@opentelemetry/api": { "optional": true }, "@types/node": { "optional": true }, - "@vitest/browser": { + "@vitest/browser-playwright": { + "optional": true + }, + "@vitest/browser-preview": { + "optional": true + }, + "@vitest/browser-webdriverio": { + "optional": true + }, + "@vitest/coverage-istanbul": { + "optional": true + }, + "@vitest/coverage-v8": { "optional": true }, "@vitest/ui": { @@ -11426,25 +11636,103 @@ }, "jsdom": { "optional": true + }, + "vite": { + "optional": false } } }, "node_modules/vitest-browser-svelte": { - "version": "0.1.0", - "resolved": "https://registry.npmjs.org/vitest-browser-svelte/-/vitest-browser-svelte-0.1.0.tgz", - "integrity": "sha512-YB6ZUZZQNqU1T9NzvTEDpwpPv35Ng1NZMPBh81zDrLEdOgROGE6nJb79NWb1Eu/a8VkHifqArpOZfJfALge6xQ==", + "version": "2.1.1", + "resolved": "https://registry.npmjs.org/vitest-browser-svelte/-/vitest-browser-svelte-2.1.1.tgz", + "integrity": "sha512-qbunYRSm+N92r9bfTkdDTpBZESLmp4QFz2SluV3n/x8U7ysosfeXYJZ4vXbJ0Y0LzoqqDnV5LHprmFgn4Eo+Ug==", "dev": true, "license": "MIT", - "engines": { - "node": "^18.0.0 || >=20.0.0" + "dependencies": { + "@testing-library/svelte-core": "^1.0.0" }, "funding": { "url": "https://opencollective.com/vitest" }, "peerDependencies": { - "@vitest/browser": "^2.1.0 || ^3.0.0-0", - "svelte": ">3.0.0", - "vitest": "^2.1.0 || ^3.0.0-0" + "svelte": "^3 || ^4 || ^5 || ^5.0.0-next.0", + "vitest": "^4.0.0" + } + }, + "node_modules/vitest/node_modules/@vitest/expect": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/expect/-/expect-4.1.8.tgz", + "integrity": "sha512-h3nDO677RDLEGlBxyQ5CW8RlMThSKSRLUePLOx09gNIWRL40edgA1GCZSZgf1W55MFAG6/Sw14KeaAnqv0NKdQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "@standard-schema/spec": "^1.1.0", + "@types/chai": "^5.2.2", + "@vitest/spy": "4.1.8", + "@vitest/utils": "4.1.8", + "chai": "^6.2.2", + "tinyrainbow": "^3.1.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/vitest/node_modules/@vitest/pretty-format": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/pretty-format/-/pretty-format-4.1.8.tgz", + "integrity": "sha512-9GasEBxpZ1VYIpqHf/0+YGg121uSNwCKOJqIrTwWP/TB7DmFCiaBpNl3aPZzoLWfWkuqhbH8vJIVobZkvdo2cA==", + "dev": true, + "license": "MIT", + "dependencies": { + "tinyrainbow": "^3.1.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/vitest/node_modules/@vitest/spy": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/spy/-/spy-4.1.8.tgz", + "integrity": "sha512-6EevtBp6OZOPF7bmz36HrGMeP3txgVSrgebWxHOafDXGkhIzfXK14f8KF6MuFfgXXUeHxmpD3BQxkV00/3s5mA==", + "dev": true, + "license": "MIT", + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/vitest/node_modules/@vitest/utils": { + "version": "4.1.8", + "resolved": "https://registry.npmjs.org/@vitest/utils/-/utils-4.1.8.tgz", + "integrity": "sha512-uOJamYALNhfJ6iolExyQM40yIQwDqYnkKtQ5VCiSe17E33H0aQ/u+1GlRuz4LZBk6Mm3sg90G9hEbmEt37C1Zg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/pretty-format": "4.1.8", + "convert-source-map": "^2.0.0", + "tinyrainbow": "^3.1.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/vitest/node_modules/chai": { + "version": "6.2.2", + "resolved": "https://registry.npmjs.org/chai/-/chai-6.2.2.tgz", + "integrity": "sha512-NUPRluOfOiTKBKvWPtSD4PhFvWCqOi0BGStNWs57X9js7XGTprSmFoz5F0tWhR4WPjNeR9jXqdC7/UpSJTnlRg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=18" + } + }, + "node_modules/vitest/node_modules/tinyrainbow": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/tinyrainbow/-/tinyrainbow-3.1.0.tgz", + "integrity": "sha512-Bf+ILmBgretUrdJxzXM0SgXLZ3XfiaUuOj/IKQHuTXip+05Xn+uyEYdVg0kYDipTBcLrCVyUzAPz7QmArb0mmw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.0.0" } }, "node_modules/web-namespaces": { @@ -11482,6 +11770,7 @@ "version": "2.0.2", "resolved": "https://registry.npmjs.org/which/-/which-2.0.2.tgz", "integrity": "sha512-BLI3Tl1TW3Pvl70l3yq3Y64i+awpwXqsGBYWkkqMtnbXgrMD+yj7rhW0kuEDxzJaYXGjEW5ogapKNMEKNMjibA==", + "dev": true, "license": "ISC", "dependencies": { "isexe": "^2.0.0" @@ -11520,95 +11809,11 @@ "node": ">=0.10.0" } }, - "node_modules/wrap-ansi": { - "version": "8.1.0", - "resolved": "https://registry.npmjs.org/wrap-ansi/-/wrap-ansi-8.1.0.tgz", - "integrity": "sha512-si7QWI6zUMq56bESFvagtmzMdGOtoxfR+Sez11Mobfc7tm+VkUckk9bW2UeffTGVUbOksxmSw0AA2gs8g71NCQ==", - "dev": true, - "license": "MIT", - "dependencies": { - "ansi-styles": "^6.1.0", - "string-width": "^5.0.1", - "strip-ansi": "^7.0.1" - }, - "engines": { - "node": ">=12" - }, - "funding": { - "url": "https://github.com/chalk/wrap-ansi?sponsor=1" - } - }, - "node_modules/wrap-ansi-cjs": { - "name": "wrap-ansi", - "version": "7.0.0", - "resolved": "https://registry.npmjs.org/wrap-ansi/-/wrap-ansi-7.0.0.tgz", - "integrity": "sha512-YVGIj2kamLSTxw6NsZjoBxfSwsn0ycdesmc4p+Q21c5zPuZ1pl+NfxVdxPtdHvmNVOQ6XSYG4AUtyt/Fi7D16Q==", - "dev": true, - "license": "MIT", - "dependencies": { - "ansi-styles": "^4.0.0", - "string-width": "^4.1.0", - "strip-ansi": "^6.0.0" - }, - "engines": { - "node": ">=10" - }, - "funding": { - "url": "https://github.com/chalk/wrap-ansi?sponsor=1" - } - }, - "node_modules/wrap-ansi-cjs/node_modules/emoji-regex": { - "version": "8.0.0", - "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-8.0.0.tgz", - "integrity": "sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A==", - "dev": true, - "license": "MIT" - }, - "node_modules/wrap-ansi-cjs/node_modules/string-width": { - "version": "4.2.3", - "resolved": "https://registry.npmjs.org/string-width/-/string-width-4.2.3.tgz", - "integrity": "sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==", - "dev": true, - "license": "MIT", - "dependencies": { - "emoji-regex": "^8.0.0", - "is-fullwidth-code-point": "^3.0.0", - "strip-ansi": "^6.0.1" - }, - "engines": { - "node": ">=8" - } - }, - "node_modules/wrap-ansi-cjs/node_modules/strip-ansi": { - "version": "6.0.1", - "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-6.0.1.tgz", - "integrity": "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==", - "dev": true, - "license": "MIT", - "dependencies": { - "ansi-regex": "^5.0.1" - }, - "engines": { - "node": ">=8" - } - }, - "node_modules/wrap-ansi/node_modules/ansi-styles": { - "version": "6.2.3", - "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-6.2.3.tgz", - "integrity": "sha512-4Dj6M28JB+oAH8kFkTLUo+a2jwOFkuqb3yucU0CANcRRUbxS0cP0nZYCGjcc3BNXwRIsUVmDGgzawme7zvJHvg==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=12" - }, - "funding": { - "url": "https://github.com/chalk/ansi-styles?sponsor=1" - } - }, "node_modules/wrappy": { "version": "1.0.2", "resolved": "https://registry.npmjs.org/wrappy/-/wrappy-1.0.2.tgz", "integrity": "sha512-l4Sp/DRseor9wL6EvV2+TuQn63dMkPjZ/sp9XkghTEbV9KlPS1xUsZ3u7/IQO4wxtcFB4bgpQPRcR3QCvezPcQ==", + "dev": true, "license": "ISC" }, "node_modules/ws": { @@ -11676,12 +11881,14 @@ "version": "1.1.2", "resolved": "https://registry.npmjs.org/zimmerframe/-/zimmerframe-1.1.2.tgz", "integrity": "sha512-rAbqEGa8ovJy4pyBxZM70hg4pE6gDgaQ0Sl9M3enG3I0d6H4XSAM3GeNGLKnsBpuijUow064sf7ww1nutC5/3w==", + "dev": true, "license": "MIT" }, "node_modules/zod": { "version": "4.2.1", "resolved": "https://registry.npmjs.org/zod/-/zod-4.2.1.tgz", "integrity": "sha512-0wZ1IRqGGhMP76gLqz8EyfBXKk0J2qo2+H3fi4mcUP/KtTocoX08nmIAHl1Z2kJIZbZee8KOpBCSNPRgauucjw==", + "dev": true, "license": "MIT", "funding": { "url": "https://github.com/sponsors/colinhacks" @@ -11691,6 +11898,7 @@ "version": "3.25.1", "resolved": "https://registry.npmjs.org/zod-to-json-schema/-/zod-to-json-schema-3.25.1.tgz", "integrity": "sha512-pM/SU9d3YAggzi6MtR4h7ruuQlqKtad8e9S0fmxcMi+ueAK5Korys/aWcV9LIIHTVbj01NdzxcnXSN+O74ZIVA==", + "dev": true, "license": "ISC", "peerDependencies": { "zod": "^3.25 || ^4" @@ -11700,6 +11908,7 @@ "version": "2.0.4", "resolved": "https://registry.npmjs.org/zwitch/-/zwitch-2.0.4.tgz", "integrity": "sha512-bXE4cR/kVZhKZX/RjPEflHaKVhUVl85noU3v6b8apfQEc1x4A+zBxjZ4lN8LqGd6WZ3dl98pY4o717VFmoPp+A==", + "dev": true, "license": "MIT", "funding": { "type": "github", diff --git a/tools/ui/package.json b/tools/ui/package.json index 7c514fa8a6a0..4f5ef4d64fa0 100644 --- a/tools/ui/package.json +++ b/tools/ui/package.json @@ -23,75 +23,77 @@ "cleanup": "rm -rf .svelte-kit build node_modules test-results" }, "devDependencies": { - "@chromatic-com/storybook": "^5.0.0", - "@eslint/compat": "^1.2.5", - "@eslint/js": "^9.18.0", - "@internationalized/date": "^3.10.1", - "@lucide/svelte": "^0.515.0", - "@playwright/test": "^1.49.1", - "@storybook/addon-a11y": "^10.2.4", - "@storybook/addon-docs": "^10.2.4", - "@storybook/addon-svelte-csf": "^5.0.10", - "@storybook/addon-vitest": "^10.2.4", - "@storybook/sveltekit": "^10.2.4", - "@sveltejs/adapter-static": "^3.0.10", - "@sveltejs/kit": "^2.48.4", - "@sveltejs/vite-plugin-svelte": "^6.2.1", - "@tailwindcss/forms": "^0.5.9", - "@tailwindcss/typography": "^0.5.15", - "@tailwindcss/vite": "^4.0.0", + "@chromatic-com/storybook": "5.0.0", + "@eslint/compat": "1.4.1", + "@eslint/js": "9.39.2", + "@internationalized/date": "3.10.1", + "@lucide/svelte": "0.515.0", + "@modelcontextprotocol/sdk": "1.26.0", + "@playwright/test": "1.56.1", + "@storybook/addon-a11y": "10.2.4", + "@storybook/addon-docs": "10.2.4", + "@storybook/addon-svelte-csf": "5.0.10", + "@storybook/addon-vitest": "10.2.4", + "@storybook/sveltekit": "10.2.4", + "@sveltejs/adapter-static": "3.0.10", + "@sveltejs/kit": "2.60.1", + "@sveltejs/vite-plugin-svelte": "6.2.1", + "@tailwindcss/forms": "0.5.10", + "@tailwindcss/typography": "0.5.16", + "@tailwindcss/vite": "4.1.11", "@types/node": "^24", - "@vitest/browser": "^3.2.3", - "@vitest/coverage-v8": "^3.2.3", - "bits-ui": "^2.14.4", - "clsx": "^2.1.1", - "dexie": "^4.0.11", - "eslint": "^9.18.0", - "eslint-config-prettier": "^10.0.1", - "eslint-plugin-storybook": "^10.2.4", - "eslint-plugin-svelte": "^3.0.0", - "globals": "^16.0.0", - "http-server": "^14.1.1", - "mdast": "^3.0.0", - "mdsvex": "^0.12.3", - "playwright": "^1.56.1", - "prettier": "^3.4.2", - "prettier-plugin-svelte": "^3.3.3", - "prettier-plugin-tailwindcss": "^0.6.11", - "rehype-katex": "^7.0.1", - "remark-math": "^6.0.0", - "sass": "^1.93.3", - "storybook": "^10.2.4", - "svelte": "^5.38.2", - "svelte-check": "^4.0.0", - "tailwind-merge": "^3.3.1", - "tailwind-variants": "^3.2.2", - "tailwindcss": "^4.0.0", - "tw-animate-css": "^1.3.5", - "typescript": "^5.0.0", - "typescript-eslint": "^8.20.0", - "unified": "^11.0.5", - "uuid": "^13.0.0", - "vite": "^7.2.2", - "vite-plugin-devtools-json": "^0.2.0", - "vitest": "^3.2.3", - "vitest-browser-svelte": "^0.1.0" + "@vitest/browser": "4.1.8", + "@vitest/browser-playwright": "4.1.8", + "@vitest/coverage-v8": "4.1.8", + "bits-ui": "2.18.1", + "clsx": "2.1.1", + "dexie": "4.0.11", + "eslint": "9.39.2", + "eslint-config-prettier": "10.1.8", + "eslint-plugin-storybook": "10.2.4", + "eslint-plugin-svelte": "3.15.0", + "globals": "16.3.0", + "highlight.js": "11.11.1", + "http-server": "14.1.1", + "mdast": "3.0.0", + "mdsvex": "0.12.6", + "mermaid": "11.15.0", + "mode-watcher": "1.1.0", + "pdfjs-dist": "5.4.54", + "playwright": "1.56.1", + "prettier": "3.6.2", + "prettier-plugin-svelte": "3.4.0", + "prettier-plugin-tailwindcss": "0.6.14", + "rehype-highlight": "7.0.2", + "rehype-katex": "7.0.1", + "rehype-stringify": "10.0.1", + "remark": "15.0.1", + "remark-breaks": "4.0.0", + "remark-gfm": "4.0.1", + "remark-html": "16.0.1", + "remark-math": "6.0.0", + "remark-rehype": "11.1.2", + "sass": "1.93.3", + "storybook": "10.3.3", + "svelte": "5.55.7", + "svelte-check": "4.3.0", + "svelte-sonner": "1.0.5", + "tailwind-merge": "3.3.1", + "tailwind-variants": "3.2.2", + "tailwindcss": "4.1.11", + "tw-animate-css": "1.3.5", + "typescript": "5.8.3", + "typescript-eslint": "8.56.0", + "unified": "11.0.5", + "unist-util-visit": "5.0.0", + "uuid": "13.0.2", + "vite": "7.3.2", + "vite-plugin-devtools-json": "0.2.1", + "vitest": "4.1.8", + "vitest-browser-svelte": "2.1.1", + "zod": "4.2.1" }, - "dependencies": { - "@modelcontextprotocol/sdk": "^1.25.1", - "highlight.js": "^11.11.1", - "mermaid": "^11.15.0", - "mode-watcher": "^1.1.0", - "pdfjs-dist": "^5.4.54", - "rehype-highlight": "^7.0.2", - "rehype-stringify": "^10.0.1", - "remark": "^15.0.1", - "remark-breaks": "^4.0.0", - "remark-gfm": "^4.0.1", - "remark-html": "^16.0.1", - "remark-rehype": "^11.1.2", - "svelte-sonner": "^1.0.5", - "unist-util-visit": "^5.0.0", - "zod": "^4.2.1" + "overrides": { + "cookie": "1.1.1" } } diff --git a/tools/ui/tests/stories/SidebarNavigation.stories.svelte b/tools/ui/tests/stories/SidebarNavigation.stories.svelte index f64ee4f9b551..aae42f2a053c 100644 --- a/tools/ui/tests/stories/SidebarNavigation.stories.svelte +++ b/tools/ui/tests/stories/SidebarNavigation.stories.svelte @@ -58,10 +58,12 @@ name="Default" play={async () => { const { conversationsStore } = await import('$lib/stores/conversations.svelte'); - - waitFor(() => setTimeout(() => { - conversationsStore.conversations = mockConversations; - }, 0)); + + waitFor(() => + setTimeout(() => { + conversationsStore.conversations = mockConversations; + }, 0) + ); }} > @@ -76,11 +78,13 @@ name="SearchActive" play={async ({ userEvent }) => { const { conversationsStore } = await import('$lib/stores/conversations.svelte'); - - waitFor(() => setTimeout(() => { - conversationsStore.conversations = mockConversations; - }, 0)); - + + waitFor(() => + setTimeout(() => { + conversationsStore.conversations = mockConversations; + }, 0) + ); + const searchTrigger = screen.getByText('Search'); userEvent.click(searchTrigger); }} diff --git a/tools/ui/vite.config.ts b/tools/ui/vite.config.ts index 5b57eae3ad54..13e889dbc101 100644 --- a/tools/ui/vite.config.ts +++ b/tools/ui/vite.config.ts @@ -7,11 +7,23 @@ import { defineConfig, searchForWorkspaceRoot } from 'vite'; import devtoolsJson from 'vite-plugin-devtools-json'; import { storybookTest } from '@storybook/addon-vitest/vitest-plugin'; import { llamaCppBuildPlugin } from './scripts/vite-plugin-llama-cpp-build'; +import { playwright } from '@vitest/browser-playwright'; const __dirname = dirname(fileURLToPath(import.meta.url)); const SERVER_ORIGIN = import.meta.env?.VITE_PUBLIC_SERVER_ORIGIN || 'http://localhost:8080'; +// eslint-disable-next-line @typescript-eslint/no-explicit-any +const browserBaseConfig: any = { + enabled: true, + provider: playwright({ + launchOptions: { + args: ['--no-sandbox'] + } + }), + instances: [{ browser: 'chromium' }] +}; + export default defineConfig({ resolve: { alias: { @@ -33,12 +45,7 @@ export default defineConfig({ extends: './vite.config.ts', test: { name: 'client', - environment: 'browser', - browser: { - enabled: true, - provider: 'playwright', - instances: [{ browser: 'chromium' }] - }, + browser: browserBaseConfig, include: ['tests/client/**/*.svelte.{test,spec}.{js,ts}'], setupFiles: ['./vitest-setup-client.ts'] } @@ -57,13 +64,7 @@ export default defineConfig({ extends: './vite.config.ts', test: { name: 'ui', - environment: 'browser', - browser: { - enabled: true, - provider: 'playwright', - instances: [{ browser: 'chromium', headless: true }] - }, - include: ['tests/stories/**/*.stories.{js,ts,svelte}'], + browser: { ...browserBaseConfig, instances: [{ browser: 'chromium', headless: true }] }, setupFiles: ['./.storybook/vitest.setup.ts'] }, plugins: [ From da6dc9dc0c2aa256c300622f26c8c7cff5dfe61c Mon Sep 17 00:00:00 2001 From: Aman Gupta Date: Mon, 1 Jun 2026 07:54:39 +0200 Subject: [PATCH 03/13] convert: add dsv4 conversion --- conversion/__init__.py | 1 + conversion/base.py | 11 ++ conversion/deepseek.py | 337 +++++++++++++++++++++++++++++++++++++- gguf-py/gguf/constants.py | 89 ++++++++++ 4 files changed, 437 insertions(+), 1 deletion(-) diff --git a/conversion/__init__.py b/conversion/__init__.py index 2c79580f8a36..eecc0e3be6eb 100644 --- a/conversion/__init__.py +++ b/conversion/__init__.py @@ -48,6 +48,7 @@ "DeepseekV2ForCausalLM": "deepseek", "DeepseekV3ForCausalLM": "deepseek", "DeepseekV32ForCausalLM": "deepseek", + "DeepseekV4ForCausalLM": "deepseek", "DistilBertForMaskedLM": "bert", "DistilBertForSequenceClassification": "bert", "DistilBertModel": "bert", diff --git a/conversion/base.py b/conversion/base.py index 408e209aa884..af953f198853 100644 --- a/conversion/base.py +++ b/conversion/base.py @@ -2587,6 +2587,17 @@ def __torch_function__(cls, func, types, args=(), kwargs=None): return cls._wrap_fn(func)(*args, **kwargs) +if hasattr(torch, "float8_e8m0fnu"): + _torch_float8_e8m0 = torch.float8_e8m0fnu + LazyTorchTensor._dtype_map[_torch_float8_e8m0] = np.uint8 + LazyTorchTensor._dtype_byteswap_map[_torch_float8_e8m0] = np.uint8 + LazyTorchTensor._dtype_str_map["F8_E8M0"] = _torch_float8_e8m0 +else: + # Older torch builds do not expose F8_E8M0. Keep the raw bytes so callers + # that know the format can decode them explicitly. + LazyTorchTensor._dtype_str_map["F8_E8M0"] = torch.uint8 + + def get_model_architecture(hparams: dict[str, Any], model_type: ModelType) -> str: # TODO @ngxson : this won't work correctly if the model has both audio & vision encoders # maybe we should fallback to text model's arch in that case, since not many models have both diff --git a/conversion/deepseek.py b/conversion/deepseek.py index 72520cc9f6a5..bfa7ca36bd89 100644 --- a/conversion/deepseek.py +++ b/conversion/deepseek.py @@ -1,15 +1,17 @@ from __future__ import annotations +import json import re from typing import Any, Callable, Iterable, TYPE_CHECKING +import numpy as np import torch if TYPE_CHECKING: from torch import Tensor -from .base import MmprojModel, ModelBase, TextModel, gguf, logger +from .base import LazyTorchTensor, MmprojModel, ModelBase, TextModel, gguf, logger from .qwen import QwenModel @@ -459,3 +461,336 @@ def set_gguf_parameters(self): self.gguf_writer.add_indexer_head_count(self.hparams["index_n_heads"]) self.gguf_writer.add_indexer_key_length(self.hparams["index_head_dim"]) self.gguf_writer.add_indexer_top_k(self.hparams["index_topk"]) + + +@ModelBase.register("DeepseekV4ForCausalLM") +class DeepseekV4FlashModel(TextModel): + model_arch = gguf.MODEL_ARCH.DEEPSEEK_V4_FLASH + _skipped_mtp_tensors = 0 + + def __init__(self, *args, **kwargs): + type(self)._skipped_mtp_tensors = 0 + super().__init__(*args, **kwargs) + + with open(self.dir_model / "config.json", "r", encoding="utf-8") as f: + raw_hparams = json.load(f) + for key, value in raw_hparams.items(): + self.hparams.setdefault(key, value) + + self.block_count = self.hparams["num_hidden_layers"] + self.tensor_map = gguf.get_tensor_name_map(self.model_arch, self.block_count) + + self._dsv4_fp8_dequantized: set[str] = set() + self._dsv4_bf16_tensors: set[str] = set() + self._dsv4_f32_tensors: set[str] = set() + self._dsv4_mxfp4_generated = False + self._collect_source_dtypes() + + if type(self)._skipped_mtp_tensors: + logger.info("Skipping %d DeepSeek-V4 MTP tensor(s) for conversion v0", type(self)._skipped_mtp_tensors) + + @classmethod + def filter_tensors(cls, item: tuple[str, Callable[[], Tensor]]) -> tuple[str, Callable[[], Tensor]] | None: + name, _ = item + if name.startswith("mtp."): + cls._skipped_mtp_tensors += 1 + return None + return super().filter_tensors(item) + + def set_vocab(self): + self._set_vocab_gpt2() + + @staticmethod + def _float8_dtypes() -> tuple[torch.dtype, ...]: + return tuple( + dtype for dtype in ( + getattr(torch, "float8_e4m3fn", None), + getattr(torch, "float8_e5m2", None), + ) if dtype is not None + ) + + @staticmethod + def _e8m0_to_float(scale: Tensor) -> Tensor: + torch_float8_e8m0 = getattr(torch, "float8_e8m0fnu", None) + if torch_float8_e8m0 is not None and scale.dtype == torch_float8_e8m0: + return scale.float() + + bits = scale.view(torch.uint8).float() + return torch.pow(torch.tensor(2.0, device=bits.device), bits - 127.0) + + def _collect_source_dtypes(self) -> None: + for name, gen in self.model_tensors.items(): + dtype = gen().dtype + if dtype == torch.bfloat16: + self._dsv4_bf16_tensors.add(name) + elif dtype == torch.float32: + self._dsv4_f32_tensors.add(name) + + def set_gguf_parameters(self): + hparams = self.hparams + arch = gguf.MODEL_ARCH_NAMES[self.model_arch] + + self.gguf_writer.add_block_count(self.block_count) + self.gguf_writer.add_context_length(hparams["max_position_embeddings"]) + self.gguf_writer.add_embedding_length(hparams["hidden_size"]) + self.gguf_writer.add_vocab_size(hparams["vocab_size"]) + self.gguf_writer.add_head_count(hparams["num_attention_heads"]) + self.gguf_writer.add_head_count_kv(hparams["num_key_value_heads"]) + self.gguf_writer.add_key_length(hparams["head_dim"]) + self.gguf_writer.add_value_length(hparams["head_dim"]) + self.gguf_writer.add_rope_dimension_count(hparams["qk_rope_head_dim"]) + self.gguf_writer.add_rope_freq_base(hparams["rope_theta"]) + self.gguf_writer.add_q_lora_rank(hparams["q_lora_rank"]) + self.gguf_writer.add_sliding_window(hparams["sliding_window"]) + self.gguf_writer.add_layer_norm_rms_eps(hparams["rms_norm_eps"]) + + rope_scaling = hparams.get("rope_scaling") or {} + rope_type = rope_scaling.get("type", rope_scaling.get("rope_type")) + if rope_type == "yarn": + self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.YARN) + self.gguf_writer.add_rope_scaling_factor(rope_scaling["factor"]) + self.gguf_writer.add_rope_scaling_orig_ctx_len(rope_scaling["original_max_position_embeddings"]) + if (yarn_beta_fast := rope_scaling.get("beta_fast")) is not None: + self.gguf_writer.add_rope_scaling_yarn_beta_fast(yarn_beta_fast) + if (yarn_beta_slow := rope_scaling.get("beta_slow")) is not None: + self.gguf_writer.add_rope_scaling_yarn_beta_slow(yarn_beta_slow) + else: + self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.NONE) + + self.gguf_writer.add_expert_feed_forward_length(hparams["moe_intermediate_size"]) + self.gguf_writer.add_expert_count(hparams["n_routed_experts"]) + self.gguf_writer.add_expert_used_count(hparams["num_experts_per_tok"]) + self.gguf_writer.add_expert_shared_count(hparams["n_shared_experts"]) + self.gguf_writer.add_expert_weights_scale(hparams["routed_scaling_factor"]) + self.gguf_writer.add_expert_weights_norm(hparams["norm_topk_prob"]) + self.gguf_writer.add_swiglu_clamp_exp([hparams["swiglu_limit"]] * self.block_count) + self.gguf_writer.add_swiglu_clamp_shexp([hparams["swiglu_limit"]] * self.block_count) + + self.gguf_writer.add_indexer_head_count(hparams["index_n_heads"]) + self.gguf_writer.add_indexer_key_length(hparams["index_head_dim"]) + self.gguf_writer.add_indexer_top_k(hparams["index_topk"]) + + self.gguf_writer.add_uint32(f"{arch}.attention.o_group_count", hparams["o_groups"]) + self.gguf_writer.add_uint32(f"{arch}.attention.o_lora_rank", hparams["o_lora_rank"]) + self.gguf_writer.add_array(f"{arch}.attention.compress_ratios", hparams["compress_ratios"]) + self.gguf_writer.add_float32(f"{arch}.attention.compress_rope.freq_base", hparams["compress_rope_theta"]) + self.gguf_writer.add_uint32(f"{arch}.hc.mult", hparams["hc_mult"]) + self.gguf_writer.add_uint32(f"{arch}.hc.sinkhorn_iters", hparams["hc_sinkhorn_iters"]) + self.gguf_writer.add_float32(f"{arch}.hc.eps", hparams["hc_eps"]) + self.gguf_writer.add_uint32(f"{arch}.moe.hash_layer_count", hparams["num_hash_layers"]) + self.gguf_writer.add_string(f"{arch}.moe.score_func", hparams["scoring_func"]) + self.gguf_writer.add_string(f"{arch}.moe.topk_method", hparams["topk_method"]) + + self.gguf_writer.add_file_type(self.ftype) + logger.info(f"gguf: file type = {self.ftype}") + + def dequant_model(self): + fp8_dtypes = self._float8_dtypes() + tensors_to_remove: list[str] = [] + + def dequant_fp8_weight(weight: Tensor, scale: Tensor) -> Tensor: + out_features, in_features = weight.shape + scale_f = self._e8m0_to_float(scale) + scale_f = scale_f.repeat_interleave(128, 0)[:out_features] + scale_f = scale_f.repeat_interleave(128, 1)[:, :in_features] + return weight.float() * scale_f + + for name in list(self.model_tensors.keys()): + if not name.endswith(".scale"): + continue + weight_name = name.removesuffix(".scale") + ".weight" + if weight_name not in self.model_tensors: + continue + + weight = self.model_tensors[weight_name] + scale = self.model_tensors[name] + if weight().dtype not in fp8_dtypes: + continue + + self.model_tensors[weight_name] = lambda w=weight, s=scale: dequant_fp8_weight(w(), s()) + self._dsv4_fp8_dequantized.add(weight_name) + tensors_to_remove.append(name) + + for name in tensors_to_remove: + del self.model_tensors[name] + + @staticmethod + def _pack_mxfp4_blocks(weight: Tensor, scale: Tensor) -> np.ndarray: + packed = weight.contiguous().view(torch.uint8) + scale_u8 = scale.contiguous().view(torch.uint8) + + out_features, packed_cols = packed.shape + logical_cols = packed_cols * 2 + if logical_cols % 32 != 0: + raise ValueError(f"MXFP4 source row has {logical_cols} values, expected a multiple of 32") + + n_blocks = logical_cols // 32 + if tuple(scale_u8.shape) != (out_features, n_blocks): + raise ValueError(f"MXFP4 scale shape {tuple(scale_u8.shape)} does not match {(out_features, n_blocks)}") + + src = packed.reshape(out_features, n_blocks, 16) + low = src & 0x0F + high = (src >> 4) & 0x0F + + # The safetensors bytes store adjacent values as low/high nibbles. + # ggml MXFP4 blocks store values 0..15 in low nibbles and 16..31 in high nibbles. + vals = torch.stack((low, high), dim=-1).reshape(out_features, n_blocks, 32) + qs = vals[:, :, :16] | (vals[:, :, 16:] << 4) + raw = torch.cat((scale_u8.unsqueeze(-1), qs.to(torch.uint8)), dim=-1) + return raw.reshape(out_features, n_blocks * 17).cpu().numpy() + + def _write_mxfp4_expert_tensor(self, bid: int, proj: str, tensor_key: gguf.MODEL_TENSOR) -> list[str]: + n_experts = self.hparams["n_routed_experts"] + data: np.ndarray | None = None + consumed: list[str] = [] + + for eid in range(n_experts): + weight_name = f"layers.{bid}.ffn.experts.{eid}.{proj}.weight" + scale_name = f"layers.{bid}.ffn.experts.{eid}.{proj}.scale" + if weight_name not in self.model_tensors or scale_name not in self.model_tensors: + raise KeyError(f"Missing routed expert tensors for {weight_name}") + + weight = LazyTorchTensor.to_eager(self.model_tensors[weight_name]()) + scale = LazyTorchTensor.to_eager(self.model_tensors[scale_name]()) + packed = self._pack_mxfp4_blocks(weight, scale) + if data is None: + data = np.empty((n_experts, *packed.shape), dtype=packed.dtype) + data[eid] = packed + consumed.extend((weight_name, scale_name)) + + assert data is not None + new_name = self.format_tensor_name(tensor_key, bid) + shape = gguf.quant_shape_from_byte_shape(data.shape, gguf.GGMLQuantizationType.MXFP4) + logger.info(f"{new_name}: repacked routed experts to MXFP4, shape = {{{', '.join(str(n) for n in reversed(shape))}}}") + self.gguf_writer.add_tensor(new_name, data, raw_dtype=gguf.GGMLQuantizationType.MXFP4) + + return consumed + + def _write_hash_routing_tensors(self) -> list[str]: + consumed: list[str] = [] + + for bid in range(self.hparams["num_hash_layers"]): + name = f"layers.{bid}.ffn.gate.tid2eid" + if name not in self.model_tensors: + raise KeyError(f"Missing hash routing tensor {name}") + + data_torch = LazyTorchTensor.to_eager(self.model_tensors[name]()) + data = data_torch.to(torch.int32).cpu().numpy() + new_name = self.format_tensor_name(gguf.MODEL_TENSOR.FFN_GATE_TID2EID, bid, "") + logger.info(f"{new_name}: converted hash routing table to I32, shape = {{{', '.join(str(n) for n in reversed(data.shape))}}}") + self.gguf_writer.add_tensor(new_name, data) + consumed.append(name) + + return consumed + + def generate_extra_tensors(self) -> Iterable[tuple[str, Tensor]]: + if self._dsv4_mxfp4_generated: + return () + + consumed: list[str] = self._write_hash_routing_tensors() + for bid in range(self.block_count): + consumed.extend(self._write_mxfp4_expert_tensor(bid, "w1", gguf.MODEL_TENSOR.FFN_GATE_EXP)) + consumed.extend(self._write_mxfp4_expert_tensor(bid, "w2", gguf.MODEL_TENSOR.FFN_DOWN_EXP)) + consumed.extend(self._write_mxfp4_expert_tensor(bid, "w3", gguf.MODEL_TENSOR.FFN_UP_EXP)) + + for name in consumed: + del self.model_tensors[name] + + self._dsv4_mxfp4_generated = True + return () + + def _format_dsv4_tensor_name(self, key: gguf.MODEL_TENSOR, bid: int | None, suffix: str = ".weight") -> str: + return self.format_tensor_name(key, bid, suffix) + + def _map_dsv4_tensor_name(self, name: str, bid: int | None) -> tuple[gguf.MODEL_TENSOR, str]: + root_map: dict[str, tuple[gguf.MODEL_TENSOR, str]] = { + "embed.weight": (gguf.MODEL_TENSOR.TOKEN_EMBD, ".weight"), + "norm.weight": (gguf.MODEL_TENSOR.OUTPUT_NORM, ".weight"), + "head.weight": (gguf.MODEL_TENSOR.OUTPUT, ".weight"), + "hc_head_fn": (gguf.MODEL_TENSOR.HC_HEAD_FN, ""), + "hc_head_base": (gguf.MODEL_TENSOR.HC_HEAD_BASE, ""), + "hc_head_scale": (gguf.MODEL_TENSOR.HC_HEAD_SCALE, ""), + } + if name in root_map: + return root_map[name] + + match = re.match(r"layers\.(\d+)\.(.+)$", name) + if match is None: + raise ValueError(f"Unsupported DeepSeek-V4 tensor {name!r}") + + layer = int(match.group(1)) + if bid != layer: + raise ValueError(f"Tensor {name!r} parsed bid {bid} but layer name has {layer}") + + layer_map: dict[str, tuple[gguf.MODEL_TENSOR, str]] = { + "hc_attn_fn": (gguf.MODEL_TENSOR.HC_ATTN_FN, ""), + "hc_attn_base": (gguf.MODEL_TENSOR.HC_ATTN_BASE, ""), + "hc_attn_scale": (gguf.MODEL_TENSOR.HC_ATTN_SCALE, ""), + "hc_ffn_fn": (gguf.MODEL_TENSOR.HC_FFN_FN, ""), + "hc_ffn_base": (gguf.MODEL_TENSOR.HC_FFN_BASE, ""), + "hc_ffn_scale": (gguf.MODEL_TENSOR.HC_FFN_SCALE, ""), + "attn.attn_sink": (gguf.MODEL_TENSOR.ATTN_SINKS, ""), + "attn.wq_a.weight": (gguf.MODEL_TENSOR.ATTN_Q_A, ".weight"), + "attn.wq_b.weight": (gguf.MODEL_TENSOR.ATTN_Q_B, ".weight"), + "attn.q_norm.weight": (gguf.MODEL_TENSOR.ATTN_Q_A_NORM, ".weight"), + "attn.wkv.weight": (gguf.MODEL_TENSOR.ATTN_KV, ".weight"), + "attn.kv_norm.weight": (gguf.MODEL_TENSOR.ATTN_KV_NORM, ".weight"), + "attn.wo_a.weight": (gguf.MODEL_TENSOR.ATTN_OUT_A, ".weight"), + "attn.wo_b.weight": (gguf.MODEL_TENSOR.ATTN_OUT_B, ".weight"), + "attn.compressor.ape": (gguf.MODEL_TENSOR.ATTN_COMPRESSOR_APE, ""), + "attn.compressor.wkv.weight": (gguf.MODEL_TENSOR.ATTN_COMPRESSOR_WKV, ".weight"), + "attn.compressor.wgate.weight": (gguf.MODEL_TENSOR.ATTN_COMPRESSOR_WGATE, ".weight"), + "attn.compressor.norm.weight": (gguf.MODEL_TENSOR.ATTN_COMPRESSOR_NORM, ".weight"), + "attn.indexer.wq_b.weight": (gguf.MODEL_TENSOR.INDEXER_ATTN_Q_B, ".weight"), + "attn.indexer.weights_proj.weight": (gguf.MODEL_TENSOR.INDEXER_PROJ, ".weight"), + "attn.indexer.compressor.ape": (gguf.MODEL_TENSOR.INDEXER_COMPRESSOR_APE, ""), + "attn.indexer.compressor.wkv.weight": (gguf.MODEL_TENSOR.INDEXER_COMPRESSOR_WKV, ".weight"), + "attn.indexer.compressor.wgate.weight": (gguf.MODEL_TENSOR.INDEXER_COMPRESSOR_WGATE, ".weight"), + "attn.indexer.compressor.norm.weight": (gguf.MODEL_TENSOR.INDEXER_COMPRESSOR_NORM, ".weight"), + "attn_norm.weight": (gguf.MODEL_TENSOR.ATTN_NORM, ".weight"), + "ffn_norm.weight": (gguf.MODEL_TENSOR.FFN_NORM, ".weight"), + "ffn.gate.weight": (gguf.MODEL_TENSOR.FFN_GATE_INP, ".weight"), + "ffn.gate.bias": (gguf.MODEL_TENSOR.FFN_EXP_PROBS_B, ".bias"), + "ffn.gate.tid2eid": (gguf.MODEL_TENSOR.FFN_GATE_TID2EID, ""), + "ffn.shared_experts.w1.weight": (gguf.MODEL_TENSOR.FFN_GATE_SHEXP, ".weight"), + "ffn.shared_experts.w2.weight": (gguf.MODEL_TENSOR.FFN_DOWN_SHEXP, ".weight"), + "ffn.shared_experts.w3.weight": (gguf.MODEL_TENSOR.FFN_UP_SHEXP, ".weight"), + } + + tensor_name = match.group(2) + if tensor_name in layer_map: + return layer_map[tensor_name] + + if re.match(r"ffn\.experts\.\d+\.w[123]\.(weight|scale)$", tensor_name): + return gguf.MODEL_TENSOR.FFN_GATE_EXP, "" + + raise ValueError(f"Unsupported DeepSeek-V4 tensor {name!r}") + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + if re.match(r"layers\.\d+\.ffn\.experts\.\d+\.w[123]\.(weight|scale)$", name): + return [] + + tensor_key, suffix = self._map_dsv4_tensor_name(name, bid) + if tensor_key == gguf.MODEL_TENSOR.FFN_GATE_TID2EID: + return [] + elif tensor_key == gguf.MODEL_TENSOR.ATTN_OUT_A: + data_torch = data_torch.reshape(self.hparams["o_groups"], self.hparams["o_lora_rank"], self.hparams["hidden_size"]) + + return [(self._format_dsv4_tensor_name(tensor_key, bid, suffix), data_torch)] + + def tensor_force_quant(self, name: str, new_name: str, bid: int | None, n_dims: int) -> gguf.GGMLQuantizationType | bool: + del new_name, bid # unused + + if name in self._dsv4_fp8_dequantized and n_dims >= 2: + return gguf.GGMLQuantizationType.Q8_0 + if name in self._dsv4_f32_tensors: + return gguf.GGMLQuantizationType.F32 + if name in self._dsv4_bf16_tensors and n_dims >= 2: + return gguf.GGMLQuantizationType.BF16 + + return False + + def prepare_tensors(self): + super().prepare_tensors() + self._is_mxfp4 = True + self.ftype = gguf.LlamaFileType.MOSTLY_MXFP4_MOE diff --git a/gguf-py/gguf/constants.py b/gguf-py/gguf/constants.py index ce556ec9b655..9590894f8318 100644 --- a/gguf-py/gguf/constants.py +++ b/gguf-py/gguf/constants.py @@ -456,6 +456,7 @@ class MODEL_ARCH(IntEnum): DEEPSEEK2 = auto() DEEPSEEK2OCR = auto() DEEPSEEK32 = auto() + DEEPSEEK_V4_FLASH = auto() CHATGLM = auto() GLM4 = auto() GLM4_MOE = auto() @@ -537,6 +538,9 @@ class MODEL_TENSOR(IntEnum): DENSE_2_OUT = auto() # embeddinggemma 2_Dense DENSE_3_OUT = auto() # embeddinggemma 3_Dense OUTPUT_NORM = auto() + HC_HEAD_FN = auto() + HC_HEAD_BASE = auto() + HC_HEAD_SCALE = auto() ROPE_FREQS = auto() ROPE_FACTORS_LONG = auto() ROPE_FACTORS_SHORT = auto() @@ -576,6 +580,7 @@ class MODEL_TENSOR(IntEnum): FFN_DOWN_CHEXP = auto() FFN_UP_CHEXP = auto() FFN_EXP_PROBS_B = auto() + FFN_GATE_TID2EID = auto() MOE_LATENT_DOWN = auto() # nemotron 3 super MOE_LATENT_UP = auto() # nemotron 3 super ATTN_Q_NORM = auto() @@ -663,6 +668,20 @@ class MODEL_TENSOR(IntEnum): ATTN_V_B = auto() ATTN_Q_A_NORM = auto() ATTN_KV_A_NORM = auto() + ATTN_KV = auto() + ATTN_KV_NORM = auto() + ATTN_OUT_A = auto() + ATTN_OUT_B = auto() + HC_ATTN_FN = auto() + HC_ATTN_BASE = auto() + HC_ATTN_SCALE = auto() + HC_FFN_FN = auto() + HC_FFN_BASE = auto() + HC_FFN_SCALE = auto() + ATTN_COMPRESSOR_WKV = auto() + ATTN_COMPRESSOR_WGATE = auto() + ATTN_COMPRESSOR_APE = auto() + ATTN_COMPRESSOR_NORM = auto() FFN_SUB_NORM = auto() ATTN_SUB_NORM = auto() DEC_ATTN_NORM = auto() @@ -724,6 +743,10 @@ class MODEL_TENSOR(IntEnum): INDEXER_PROJ = auto() INDEXER_ATTN_K = auto() INDEXER_ATTN_Q_B = auto() + INDEXER_COMPRESSOR_WKV = auto() + INDEXER_COMPRESSOR_WGATE = auto() + INDEXER_COMPRESSOR_APE = auto() + INDEXER_COMPRESSOR_NORM = auto() # vision V_MMPROJ = auto() V_MMPROJ_FC = auto() @@ -977,6 +1000,7 @@ class MODEL_TENSOR(IntEnum): MODEL_ARCH.DEEPSEEK2: "deepseek2", MODEL_ARCH.DEEPSEEK2OCR: "deepseek2-ocr", MODEL_ARCH.DEEPSEEK32: "deepseek32", + MODEL_ARCH.DEEPSEEK_V4_FLASH: "deepseek-v4-flash", MODEL_ARCH.CHATGLM: "chatglm", MODEL_ARCH.GLM4: "glm4", MODEL_ARCH.GLM4_MOE: "glm4moe", @@ -1057,6 +1081,9 @@ class MODEL_TENSOR(IntEnum): MODEL_TENSOR.OUTPUT: "output", MODEL_TENSOR.DENSE_2_OUT: "dense_2", # embeddinggemma 2_Dense MODEL_TENSOR.DENSE_3_OUT: "dense_3", # embeddinggemma 2_Dense + MODEL_TENSOR.HC_HEAD_FN: "output_hc.fn", + MODEL_TENSOR.HC_HEAD_BASE: "output_hc.base", + MODEL_TENSOR.HC_HEAD_SCALE: "output_hc.scale", MODEL_TENSOR.ROPE_FREQS: "rope_freqs", MODEL_TENSOR.ROPE_FACTORS_LONG: "rope_factors_long", MODEL_TENSOR.ROPE_FACTORS_SHORT: "rope_factors_short", @@ -1098,6 +1125,7 @@ class MODEL_TENSOR(IntEnum): MODEL_TENSOR.FFN_UP_EXP: "blk.{bid}.ffn_up_exps", MODEL_TENSOR.FFN_GATE_UP_EXP: "blk.{bid}.ffn_gate_up_exps", MODEL_TENSOR.FFN_EXP_PROBS_B: "blk.{bid}.exp_probs_b", + MODEL_TENSOR.FFN_GATE_TID2EID: "blk.{bid}.ffn_gate_tid2eid", MODEL_TENSOR.MOE_LATENT_DOWN: "blk.{bid}.ffn_latent_down", # nemotron 3 super MODEL_TENSOR.MOE_LATENT_UP: "blk.{bid}.ffn_latent_up", # nemotron 3 super MODEL_TENSOR.LAYER_OUT_NORM: "blk.{bid}.layer_output_norm", @@ -1183,6 +1211,20 @@ class MODEL_TENSOR(IntEnum): MODEL_TENSOR.ATTN_V_B: "blk.{bid}.attn_v_b", MODEL_TENSOR.ATTN_Q_A_NORM: "blk.{bid}.attn_q_a_norm", MODEL_TENSOR.ATTN_KV_A_NORM: "blk.{bid}.attn_kv_a_norm", + MODEL_TENSOR.ATTN_KV: "blk.{bid}.attn_kv", + MODEL_TENSOR.ATTN_KV_NORM: "blk.{bid}.attn_kv_norm", + MODEL_TENSOR.ATTN_OUT_A: "blk.{bid}.attn_wo_a", + MODEL_TENSOR.ATTN_OUT_B: "blk.{bid}.attn_wo_b", + MODEL_TENSOR.HC_ATTN_FN: "blk.{bid}.hc_attn.fn", + MODEL_TENSOR.HC_ATTN_BASE: "blk.{bid}.hc_attn.base", + MODEL_TENSOR.HC_ATTN_SCALE: "blk.{bid}.hc_attn.scale", + MODEL_TENSOR.HC_FFN_FN: "blk.{bid}.hc_ffn.fn", + MODEL_TENSOR.HC_FFN_BASE: "blk.{bid}.hc_ffn.base", + MODEL_TENSOR.HC_FFN_SCALE: "blk.{bid}.hc_ffn.scale", + MODEL_TENSOR.ATTN_COMPRESSOR_WKV: "blk.{bid}.attn_comp.wkv", + MODEL_TENSOR.ATTN_COMPRESSOR_WGATE: "blk.{bid}.attn_comp.wgate", + MODEL_TENSOR.ATTN_COMPRESSOR_APE: "blk.{bid}.attn_comp.ape", + MODEL_TENSOR.ATTN_COMPRESSOR_NORM: "blk.{bid}.attn_comp.norm", MODEL_TENSOR.ATTN_SUB_NORM: "blk.{bid}.attn_sub_norm", MODEL_TENSOR.FFN_SUB_NORM: "blk.{bid}.ffn_sub_norm", MODEL_TENSOR.DEC_ATTN_NORM: "dec.blk.{bid}.attn_norm", @@ -1244,6 +1286,10 @@ class MODEL_TENSOR(IntEnum): MODEL_TENSOR.INDEXER_PROJ: "blk.{bid}.indexer.proj", MODEL_TENSOR.INDEXER_ATTN_K: "blk.{bid}.indexer.attn_k", MODEL_TENSOR.INDEXER_ATTN_Q_B: "blk.{bid}.indexer.attn_q_b", + MODEL_TENSOR.INDEXER_COMPRESSOR_WKV: "blk.{bid}.indexer_comp.wkv", + MODEL_TENSOR.INDEXER_COMPRESSOR_WGATE: "blk.{bid}.indexer_comp.wgate", + MODEL_TENSOR.INDEXER_COMPRESSOR_APE: "blk.{bid}.indexer_comp.ape", + MODEL_TENSOR.INDEXER_COMPRESSOR_NORM: "blk.{bid}.indexer_comp.norm", # vision MODEL_TENSOR.V_MMPROJ: "mm.{bid}", MODEL_TENSOR.V_MMPROJ_FC: "mm.model.fc", @@ -2987,6 +3033,49 @@ class MODEL_TENSOR(IntEnum): MODEL_TENSOR.NEXTN_SHARED_HEAD_HEAD, MODEL_TENSOR.NEXTN_SHARED_HEAD_NORM, ], + MODEL_ARCH.DEEPSEEK_V4_FLASH: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.HC_HEAD_FN, + MODEL_TENSOR.HC_HEAD_BASE, + MODEL_TENSOR.HC_HEAD_SCALE, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.ATTN_SINKS, + MODEL_TENSOR.ATTN_Q_A, + MODEL_TENSOR.ATTN_Q_B, + MODEL_TENSOR.ATTN_Q_A_NORM, + MODEL_TENSOR.ATTN_KV, + MODEL_TENSOR.ATTN_KV_NORM, + MODEL_TENSOR.ATTN_OUT_A, + MODEL_TENSOR.ATTN_OUT_B, + MODEL_TENSOR.HC_ATTN_FN, + MODEL_TENSOR.HC_ATTN_BASE, + MODEL_TENSOR.HC_ATTN_SCALE, + MODEL_TENSOR.HC_FFN_FN, + MODEL_TENSOR.HC_FFN_BASE, + MODEL_TENSOR.HC_FFN_SCALE, + MODEL_TENSOR.ATTN_COMPRESSOR_WKV, + MODEL_TENSOR.ATTN_COMPRESSOR_WGATE, + MODEL_TENSOR.ATTN_COMPRESSOR_APE, + MODEL_TENSOR.ATTN_COMPRESSOR_NORM, + MODEL_TENSOR.INDEXER_PROJ, + MODEL_TENSOR.INDEXER_ATTN_Q_B, + MODEL_TENSOR.INDEXER_COMPRESSOR_WKV, + MODEL_TENSOR.INDEXER_COMPRESSOR_WGATE, + MODEL_TENSOR.INDEXER_COMPRESSOR_APE, + MODEL_TENSOR.INDEXER_COMPRESSOR_NORM, + MODEL_TENSOR.FFN_GATE_INP, + MODEL_TENSOR.FFN_GATE_TID2EID, + MODEL_TENSOR.FFN_EXP_PROBS_B, + MODEL_TENSOR.FFN_NORM, + MODEL_TENSOR.FFN_GATE_EXP, + MODEL_TENSOR.FFN_DOWN_EXP, + MODEL_TENSOR.FFN_UP_EXP, + MODEL_TENSOR.FFN_GATE_SHEXP, + MODEL_TENSOR.FFN_DOWN_SHEXP, + MODEL_TENSOR.FFN_UP_SHEXP, + ], MODEL_ARCH.ERNIE4_5_MOE: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.OUTPUT_NORM, From df5506b940b8611104aee4a9ef9d5c4c00483e6d Mon Sep 17 00:00:00 2001 From: Aman Gupta Date: Mon, 1 Jun 2026 16:03:12 +0200 Subject: [PATCH 04/13] add basic setup --- src/llama-arch.cpp | 46 ++++ src/llama-arch.h | 23 ++ src/llama-context.cpp | 6 +- src/llama-graph.cpp | 63 +++-- src/llama-graph.h | 6 +- src/llama-hparams.h | 11 + src/llama-model-loader.cpp | 6 + src/llama-model.cpp | 4 + src/llama-model.h | 25 ++ src/models/deepseek-v4.cpp | 544 +++++++++++++++++++++++++++++++++++++ src/models/models.h | 50 ++++ 11 files changed, 760 insertions(+), 24 deletions(-) create mode 100644 src/models/deepseek-v4.cpp diff --git a/src/llama-arch.cpp b/src/llama-arch.cpp index fea898deaf2c..f8e0d2359c02 100644 --- a/src/llama-arch.cpp +++ b/src/llama-arch.cpp @@ -76,6 +76,7 @@ static const std::map LLM_ARCH_NAMES = { { LLM_ARCH_DEEPSEEK2, "deepseek2" }, { LLM_ARCH_DEEPSEEK2OCR, "deepseek2-ocr" }, { LLM_ARCH_DEEPSEEK32, "deepseek32" }, + { LLM_ARCH_DEEPSEEK_V4_FLASH, "deepseek-v4-flash" }, { LLM_ARCH_CHATGLM, "chatglm" }, { LLM_ARCH_GLM4, "glm4" }, { LLM_ARCH_GLM4_MOE, "glm4moe" }, @@ -431,6 +432,23 @@ static const std::map LLM_TENSOR_NAMES = { { LLM_TENSOR_ATTN_Q_B, "blk.%d.attn_q_b" }, { LLM_TENSOR_ATTN_KV_A_MQA, "blk.%d.attn_kv_a_mqa" }, { LLM_TENSOR_ATTN_KV_B, "blk.%d.attn_kv_b" }, + { LLM_TENSOR_ATTN_KV, "blk.%d.attn_kv" }, + { LLM_TENSOR_ATTN_KV_NORM, "blk.%d.attn_kv_norm" }, + { LLM_TENSOR_ATTN_OUT_A, "blk.%d.attn_wo_a" }, + { LLM_TENSOR_ATTN_OUT_B, "blk.%d.attn_wo_b" }, + { LLM_TENSOR_HC_HEAD_FN, "output_hc.fn" }, + { LLM_TENSOR_HC_HEAD_BASE, "output_hc.base" }, + { LLM_TENSOR_HC_HEAD_SCALE, "output_hc.scale" }, + { LLM_TENSOR_HC_ATTN_FN, "blk.%d.hc_attn.fn" }, + { LLM_TENSOR_HC_ATTN_BASE, "blk.%d.hc_attn.base" }, + { LLM_TENSOR_HC_ATTN_SCALE, "blk.%d.hc_attn.scale" }, + { LLM_TENSOR_HC_FFN_FN, "blk.%d.hc_ffn.fn" }, + { LLM_TENSOR_HC_FFN_BASE, "blk.%d.hc_ffn.base" }, + { LLM_TENSOR_HC_FFN_SCALE, "blk.%d.hc_ffn.scale" }, + { LLM_TENSOR_ATTN_COMPRESSOR_WKV, "blk.%d.attn_comp.wkv" }, + { LLM_TENSOR_ATTN_COMPRESSOR_WGATE, "blk.%d.attn_comp.wgate" }, + { LLM_TENSOR_ATTN_COMPRESSOR_APE, "blk.%d.attn_comp.ape" }, + { LLM_TENSOR_ATTN_COMPRESSOR_NORM, "blk.%d.attn_comp.norm" }, { LLM_TENSOR_PER_LAYER_TOKEN_EMBD, "per_layer_token_embd" }, { LLM_TENSOR_PER_LAYER_MODEL_PROJ, "per_layer_model_proj" }, { LLM_TENSOR_PER_LAYER_PROJ_NORM, "per_layer_proj_norm" }, @@ -555,6 +573,11 @@ static const std::map LLM_TENSOR_NAMES = { { LLM_TENSOR_INDEXER_PROJ, "blk.%d.indexer.proj" }, { LLM_TENSOR_INDEXER_ATTN_K, "blk.%d.indexer.attn_k" }, { LLM_TENSOR_INDEXER_ATTN_Q_B, "blk.%d.indexer.attn_q_b" }, + { LLM_TENSOR_INDEXER_COMPRESSOR_WKV, "blk.%d.indexer_comp.wkv" }, + { LLM_TENSOR_INDEXER_COMPRESSOR_WGATE, "blk.%d.indexer_comp.wgate" }, + { LLM_TENSOR_INDEXER_COMPRESSOR_APE, "blk.%d.indexer_comp.ape" }, + { LLM_TENSOR_INDEXER_COMPRESSOR_NORM, "blk.%d.indexer_comp.norm" }, + { LLM_TENSOR_FFN_GATE_TID2EID, "blk.%d.ffn_gate_tid2eid" }, }; // declare information about the model weight tensors: @@ -601,6 +624,23 @@ static const std::map LLM_TENSOR_INFOS = { {LLM_TENSOR_ATTN_Q_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, {LLM_TENSOR_ATTN_KV_A_MQA, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, {LLM_TENSOR_ATTN_KV_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_KV, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_KV_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ATTN_OUT_A, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_OUT_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_HC_HEAD_FN, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_HC_HEAD_BASE, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_ADD}}, + {LLM_TENSOR_HC_HEAD_SCALE, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL}}, + {LLM_TENSOR_HC_ATTN_FN, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_HC_ATTN_BASE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_HC_ATTN_SCALE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_HC_FFN_FN, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_HC_FFN_BASE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_HC_FFN_SCALE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ATTN_COMPRESSOR_WKV, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_COMPRESSOR_WGATE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_COMPRESSOR_APE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_ATTN_COMPRESSOR_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, {LLM_TENSOR_ATTN_K_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, {LLM_TENSOR_ATTN_V_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, {LLM_TENSOR_ATTN_SINKS, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_SCALE}}, @@ -764,6 +804,11 @@ static const std::map LLM_TENSOR_INFOS = { {LLM_TENSOR_INDEXER_PROJ, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, {LLM_TENSOR_INDEXER_ATTN_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, {LLM_TENSOR_INDEXER_ATTN_Q_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_INDEXER_COMPRESSOR_WKV, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_INDEXER_COMPRESSOR_WGATE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_INDEXER_COMPRESSOR_APE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_INDEXER_COMPRESSOR_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_FFN_GATE_TID2EID, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_GET_ROWS}}, // NextN/MTP tensors are stored per-block (blk.%d.nextn.*) even though only the // last nextn_predict_layers blocks carry them. Classify as LAYER_REPEATING so // the model loader doesn't fault on the block index. @@ -911,6 +956,7 @@ bool llm_arch_supports_sm_tensor(const llm_arch & arch) { case LLM_ARCH_OLMOE: case LLM_ARCH_DEEPSEEK2: case LLM_ARCH_DEEPSEEK32: + case LLM_ARCH_DEEPSEEK_V4_FLASH: case LLM_ARCH_GLM_DSA: case LLM_ARCH_BITNET: case LLM_ARCH_T5: diff --git a/src/llama-arch.h b/src/llama-arch.h index f364f6b0bae1..c6a012b2792c 100644 --- a/src/llama-arch.h +++ b/src/llama-arch.h @@ -80,6 +80,7 @@ enum llm_arch { LLM_ARCH_DEEPSEEK2, LLM_ARCH_DEEPSEEK2OCR, LLM_ARCH_DEEPSEEK32, + LLM_ARCH_DEEPSEEK_V4_FLASH, LLM_ARCH_CHATGLM, LLM_ARCH_GLM4, LLM_ARCH_GLM4_MOE, @@ -491,10 +492,27 @@ enum llm_tensor { LLM_TENSOR_ATTN_Q_B, LLM_TENSOR_ATTN_KV_A_MQA, LLM_TENSOR_ATTN_KV_B, + LLM_TENSOR_ATTN_KV, + LLM_TENSOR_ATTN_KV_NORM, + LLM_TENSOR_ATTN_OUT_A, + LLM_TENSOR_ATTN_OUT_B, LLM_TENSOR_ATTN_K_B, LLM_TENSOR_ATTN_V_B, LLM_TENSOR_ATTN_Q_A_NORM, LLM_TENSOR_ATTN_KV_A_NORM, + LLM_TENSOR_HC_HEAD_FN, + LLM_TENSOR_HC_HEAD_BASE, + LLM_TENSOR_HC_HEAD_SCALE, + LLM_TENSOR_HC_ATTN_FN, + LLM_TENSOR_HC_ATTN_BASE, + LLM_TENSOR_HC_ATTN_SCALE, + LLM_TENSOR_HC_FFN_FN, + LLM_TENSOR_HC_FFN_BASE, + LLM_TENSOR_HC_FFN_SCALE, + LLM_TENSOR_ATTN_COMPRESSOR_WKV, + LLM_TENSOR_ATTN_COMPRESSOR_WGATE, + LLM_TENSOR_ATTN_COMPRESSOR_APE, + LLM_TENSOR_ATTN_COMPRESSOR_NORM, LLM_TENSOR_ATTN_SUB_NORM, LLM_TENSOR_FFN_SUB_NORM, LLM_TENSOR_DEC_ATTN_NORM, @@ -556,6 +574,11 @@ enum llm_tensor { LLM_TENSOR_INDEXER_PROJ, LLM_TENSOR_INDEXER_ATTN_K, LLM_TENSOR_INDEXER_ATTN_Q_B, + LLM_TENSOR_INDEXER_COMPRESSOR_WKV, + LLM_TENSOR_INDEXER_COMPRESSOR_WGATE, + LLM_TENSOR_INDEXER_COMPRESSOR_APE, + LLM_TENSOR_INDEXER_COMPRESSOR_NORM, + LLM_TENSOR_FFN_GATE_TID2EID, LLM_TENSOR_NEXTN_EH_PROJ, LLM_TENSOR_NEXTN_EMBED_TOKENS, LLM_TENSOR_NEXTN_ENORM, diff --git a/src/llama-context.cpp b/src/llama-context.cpp index f59381a4d757..5704dccd32f2 100644 --- a/src/llama-context.cpp +++ b/src/llama-context.cpp @@ -2215,7 +2215,11 @@ void llama_context::output_reorder() { // uint32_t llama_context::graph_max_nodes(uint32_t n_tokens) const { - if (model.arch == LLM_ARCH_QWEN3NEXT || model.arch == LLM_ARCH_KIMI_LINEAR || model.arch == LLM_ARCH_QWEN35 || model.arch == LLM_ARCH_QWEN35MOE) { + if (model.arch == LLM_ARCH_QWEN3NEXT || + model.arch == LLM_ARCH_KIMI_LINEAR || + model.arch == LLM_ARCH_QWEN35 || + model.arch == LLM_ARCH_QWEN35MOE || + model.arch == LLM_ARCH_DEEPSEEK_V4_FLASH) { return std::max(n_tokens * 40, 32u * model.n_tensors()); } uint32_t res = std::max(1024u, 8u*model.n_tensors()); diff --git a/src/llama-graph.cpp b/src/llama-graph.cpp index f910528d21b3..b8bd0cd1729e 100644 --- a/src/llama-graph.cpp +++ b/src/llama-graph.cpp @@ -1286,20 +1286,24 @@ ggml_tensor * llm_graph_context::build_ffn( switch (type_op) { case LLM_FFN_SILU: if (gate && type_gate == LLM_FFN_PAR) { - // Step35: HF clamps gate (after SiLU) and up before multiplication - if (arch == LLM_ARCH_STEP35 && il >= 0) { + if (il >= 0) { const float limit = hparams.swiglu_clamp_shexp[il]; constexpr float eps = 1e-6f; if (limit > eps) { - ggml_tensor * gate_act = ggml_silu(ctx0, cur); - cb(gate_act, "ffn_silu", il); - gate_act = ggml_clamp(ctx0, gate_act, -INFINITY, limit); - cb(gate_act, "ffn_silu_clamped", il); - tmp = ggml_clamp(ctx0, tmp, -limit, limit); cb(tmp, "ffn_up_clamped", il); - cur = ggml_mul(ctx0, gate_act, tmp); + if (arch == LLM_ARCH_DEEPSEEK_V4_FLASH) { + cur = ggml_clamp(ctx0, cur, -INFINITY, limit); + cb(cur, "ffn_gate_clamped", il); + cur = ggml_swiglu_split(ctx0, cur, tmp); + } else { + ggml_tensor * gate_act = ggml_silu(ctx0, cur); + cb(gate_act, "ffn_silu", il); + gate_act = ggml_clamp(ctx0, gate_act, -INFINITY, limit); + cb(gate_act, "ffn_silu_clamped", il); + cur = ggml_mul(ctx0, gate_act, tmp); + } cb(cur, "ffn_swiglu_limited", il); type_gate = LLM_FFN_SEQ; break; @@ -1409,7 +1413,8 @@ ggml_tensor * llm_graph_context::build_moe_ffn( ggml_tensor * gate_up_exps, ggml_tensor * up_exps_s, ggml_tensor * gate_exps_s, - ggml_tensor * down_exps_s) const { + ggml_tensor * down_exps_s, + ggml_tensor * selected_experts_in) const { return build_moe_ffn( cur, gate_inp, /* gate_inp_b */ nullptr, @@ -1429,7 +1434,8 @@ ggml_tensor * llm_graph_context::build_moe_ffn( /* gate_up_exps_b */ nullptr, up_exps_s, gate_exps_s, - down_exps_s + down_exps_s, + selected_experts_in ); } @@ -1456,7 +1462,8 @@ ggml_tensor * llm_graph_context::build_moe_ffn( ggml_tensor * gate_up_exps_b, ggml_tensor * up_exps_s, ggml_tensor * gate_exps_s, - ggml_tensor * down_exps_s) const { + ggml_tensor * down_exps_s, + ggml_tensor * selected_experts_in) const { const int64_t n_embd = cur->ne[0]; const int64_t n_tokens = cur->ne[1]; const bool weight_before_ffn = arch == LLM_ARCH_LLAMA4; // for llama4, we apply the sigmoid-ed weights before the FFN @@ -1465,6 +1472,9 @@ ggml_tensor * llm_graph_context::build_moe_ffn( if (probs_in == nullptr) { logits = build_lora_mm(gate_inp, cur); // [n_expert, n_tokens] + if (gating_op == LLAMA_EXPERT_GATING_FUNC_TYPE_SQRT_SOFTPLUS) { + ggml_mul_mat_set_prec(logits, GGML_PREC_F32); + } cb(logits, "ffn_moe_logits", il); } else { logits = probs_in; @@ -1489,6 +1499,10 @@ ggml_tensor * llm_graph_context::build_moe_ffn( { probs = logits; // [n_expert, n_tokens] } break; + case LLAMA_EXPERT_GATING_FUNC_TYPE_SQRT_SOFTPLUS: + { + probs = ggml_sqrt(ctx0, ggml_softplus(ctx0, logits)); // [n_expert, n_tokens] + } break; default: GGML_ABORT("fatal error"); } @@ -1539,8 +1553,11 @@ ggml_tensor * llm_graph_context::build_moe_ffn( } // select experts - ggml_tensor * selected_experts = ggml_argsort_top_k(ctx0, selection_probs, n_expert_used); // [n_expert_used, n_tokens] - cb(selected_experts->src[0], "ffn_moe_argsort", il); + ggml_tensor * selected_experts = selected_experts_in; + if (selected_experts == nullptr) { + selected_experts = ggml_argsort_top_k(ctx0, selection_probs, n_expert_used); // [n_expert_used, n_tokens] + cb(selected_experts->src[0], "ffn_moe_argsort", il); + } cb(selected_experts, "ffn_moe_topk", il); if (arch == LLM_ARCH_GROVEMOE && n_expert != hparams.n_expert) { @@ -1668,20 +1685,24 @@ ggml_tensor * llm_graph_context::build_moe_ffn( switch (type_op) { case LLM_FFN_SILU: if (gate_exps) { - // Step35: per-layer clamp for routed experts - if (arch == LLM_ARCH_STEP35 && il >= 0) { + if (il >= 0) { const float limit = hparams.swiglu_clamp_exp[il]; constexpr float eps = 1e-6f; if (limit > eps) { - ggml_tensor * gate_act = ggml_silu(ctx0, cur); - cb(gate_act, "ffn_moe_silu", il); - gate_act = ggml_clamp(ctx0, gate_act, -INFINITY, limit); - cb(gate_act, "ffn_moe_silu_clamped", il); - up = ggml_clamp(ctx0, up, -limit, limit); cb(up, "ffn_moe_up_clamped", il); - cur = ggml_mul(ctx0, gate_act, up); + if (arch == LLM_ARCH_DEEPSEEK_V4_FLASH) { + cur = ggml_clamp(ctx0, cur, -INFINITY, limit); + cb(cur, "ffn_moe_gate_clamped", il); + cur = ggml_swiglu_split(ctx0, cur, up); + } else { + ggml_tensor * gate_act = ggml_silu(ctx0, cur); + cb(gate_act, "ffn_moe_silu", il); + gate_act = ggml_clamp(ctx0, gate_act, -INFINITY, limit); + cb(gate_act, "ffn_moe_silu_clamped", il); + cur = ggml_mul(ctx0, gate_act, up); + } cb(cur, "ffn_moe_swiglu_limited", il); break; } diff --git a/src/llama-graph.h b/src/llama-graph.h index bf5be09ac7fa..f14ccd3734dd 100644 --- a/src/llama-graph.h +++ b/src/llama-graph.h @@ -907,7 +907,8 @@ struct llm_graph_context { ggml_tensor * gate_up_exps = nullptr, ggml_tensor * up_exps_s = nullptr, ggml_tensor * gate_exps_s = nullptr, - ggml_tensor * down_exps_s = nullptr) const; + ggml_tensor * down_exps_s = nullptr, + ggml_tensor * selected_experts_in = nullptr) const; ggml_tensor * build_moe_ffn( ggml_tensor * cur, @@ -932,7 +933,8 @@ struct llm_graph_context { ggml_tensor * gate_up_exps_b = nullptr, ggml_tensor * up_exps_s = nullptr, ggml_tensor * gate_exps_s = nullptr, - ggml_tensor * down_exps_s = nullptr) const; + ggml_tensor * down_exps_s = nullptr, + ggml_tensor * selected_experts_in = nullptr) const; // // inputs diff --git a/src/llama-hparams.h b/src/llama-hparams.h index e8ed4dd74de3..1fed55968b11 100644 --- a/src/llama-hparams.h +++ b/src/llama-hparams.h @@ -14,6 +14,7 @@ enum llama_expert_gating_func_type { LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX = 1, LLAMA_EXPERT_GATING_FUNC_TYPE_SIGMOID = 2, LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX_WEIGHT = 3, // applied to the router weights instead of the logits + LLAMA_EXPERT_GATING_FUNC_TYPE_SQRT_SOFTPLUS = 4, }; enum llama_swa_type { @@ -218,6 +219,16 @@ struct llama_hparams { uint32_t indexer_head_size = 0; uint32_t indexer_top_k = 0; + // DeepSeek-V4 Flash + uint32_t dsv4_o_group_count = 0; + uint32_t dsv4_o_lora_rank = 0; + uint32_t dsv4_hc_mult = 0; + uint32_t dsv4_hc_sinkhorn_iters = 0; + uint32_t dsv4_hash_layer_count = 0; + float dsv4_compress_rope_base = 0.0f; + float dsv4_hc_eps = 0.0f; + std::array dsv4_compress_ratios; + // qwen3vl deepstack uint32_t n_deepstack_layers = 0; diff --git a/src/llama-model-loader.cpp b/src/llama-model-loader.cpp index 4d7b11067c97..6751a5b74ed7 100644 --- a/src/llama-model-loader.cpp +++ b/src/llama-model-loader.cpp @@ -294,6 +294,8 @@ namespace GGUFMeta { } template bool llama_model_loader::get_arr_n(enum llm_kv kid, uint32_t & result, bool required); + template std::enable_if::value, bool>::type + llama_model_loader::get_arr_n(const std::string & key, uint32_t & result, bool required); template bool llama_model_loader::get_arr(const std::string & key, std::vector & result, bool required) { @@ -393,6 +395,10 @@ namespace GGUFMeta { } template bool llama_model_loader::get_arr>(enum llm_kv kid, std::vector & result, bool required); + template bool llama_model_loader::get_arr( + const std::string & key, + std::array & result, + bool required); template bool llama_model_loader::get_key(const std::string & key, T & result, bool required) { diff --git a/src/llama-model.cpp b/src/llama-model.cpp index bc7a83b15f53..010af1e0728a 100644 --- a/src/llama-model.cpp +++ b/src/llama-model.cpp @@ -177,6 +177,8 @@ static llama_model * llama_model_mapping(llm_arch arch, const llama_model_params return new llama_model_deepseek2ocr(params); case LLM_ARCH_DEEPSEEK32: return new llama_model_deepseek32(params); + case LLM_ARCH_DEEPSEEK_V4_FLASH: + return new llama_model_deepseek_v4_flash(params); case LLM_ARCH_GLM_DSA: return new llama_model_glm_dsa(params); case LLM_ARCH_MISTRAL4: @@ -800,6 +802,7 @@ static const char * llama_expert_gating_func_name(llama_expert_gating_func_type switch (type) { case LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX: return "softmax"; case LLAMA_EXPERT_GATING_FUNC_TYPE_SIGMOID: return "sigmoid"; + case LLAMA_EXPERT_GATING_FUNC_TYPE_SQRT_SOFTPLUS: return "sqrtsoftplus"; default: return "unknown"; } } @@ -2334,6 +2337,7 @@ llama_rope_type llama_model_rope_type(const llama_model * model) { case LLM_ARCH_DEEPSEEK2: case LLM_ARCH_DEEPSEEK2OCR: case LLM_ARCH_DEEPSEEK32: + case LLM_ARCH_DEEPSEEK_V4_FLASH: case LLM_ARCH_PLM: case LLM_ARCH_CHATGLM: case LLM_ARCH_GRANITE: diff --git a/src/llama-model.h b/src/llama-model.h index a561374ed956..1127dc92d48b 100644 --- a/src/llama-model.h +++ b/src/llama-model.h @@ -254,9 +254,11 @@ struct llama_layer { struct ggml_tensor * wq_b = nullptr; struct ggml_tensor * wkv_a_mqa = nullptr; struct ggml_tensor * wkv_b = nullptr; + struct ggml_tensor * wkv = nullptr; struct ggml_tensor * wk_b = nullptr; struct ggml_tensor * wv_b = nullptr; struct ggml_tensor * wqkv_b = nullptr; + struct ggml_tensor * wo_a = nullptr; struct ggml_tensor * wo_b = nullptr; struct ggml_tensor * wq_cross = nullptr; struct ggml_tensor * wk_cross = nullptr; @@ -332,6 +334,7 @@ struct llama_layer { struct ggml_tensor * ffn_up_b = nullptr; // b3 struct ggml_tensor * ffn_act = nullptr; struct ggml_tensor * ffn_exp_probs_b = nullptr; + struct ggml_tensor * ffn_gate_tid2eid = nullptr; // mamba proj struct ggml_tensor * ssm_in = nullptr; @@ -462,6 +465,23 @@ struct llama_layer { // openai-moe struct ggml_tensor * attn_sinks = nullptr; + // DeepSeek-V4 Flash + struct ggml_tensor * attn_kv_norm = nullptr; + struct ggml_tensor * hc_attn_fn = nullptr; + struct ggml_tensor * hc_attn_base = nullptr; + struct ggml_tensor * hc_attn_scale = nullptr; + struct ggml_tensor * hc_ffn_fn = nullptr; + struct ggml_tensor * hc_ffn_base = nullptr; + struct ggml_tensor * hc_ffn_scale = nullptr; + struct ggml_tensor * attn_comp_wkv = nullptr; + struct ggml_tensor * attn_comp_wgate = nullptr; + struct ggml_tensor * attn_comp_ape = nullptr; + struct ggml_tensor * attn_comp_norm = nullptr; + struct ggml_tensor * indexer_comp_wkv = nullptr; + struct ggml_tensor * indexer_comp_wgate = nullptr; + struct ggml_tensor * indexer_comp_ape = nullptr; + struct ggml_tensor * indexer_comp_norm = nullptr; + // cogvlm struct ggml_tensor * visexp_attn_wqkv = nullptr; struct ggml_tensor * visexp_attn_wo = nullptr; @@ -548,6 +568,11 @@ struct llama_model { struct ggml_tensor * output_s = nullptr; struct ggml_tensor * output_in_s = nullptr; + // DeepSeek-V4 Flash + struct ggml_tensor * hc_head_fn = nullptr; + struct ggml_tensor * hc_head_base = nullptr; + struct ggml_tensor * hc_head_scale = nullptr; + // classifier struct ggml_tensor * cls = nullptr; struct ggml_tensor * cls_b = nullptr; diff --git a/src/models/deepseek-v4.cpp b/src/models/deepseek-v4.cpp new file mode 100644 index 000000000000..cd35bf60aaba --- /dev/null +++ b/src/models/deepseek-v4.cpp @@ -0,0 +1,544 @@ +#include "models.h" + +#include +#include +#include +#include + +static std::string dsv4_kv(const char * suffix) { + return std::string("deepseek-v4-flash.") + suffix; +} + +void llama_model_deepseek_v4_flash::load_arch_hparams(llama_model_loader & ml) { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + ml.get_key(LLM_KV_ATTENTION_Q_LORA_RANK, hparams.n_lora_q); + ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa); + + ml.get_key(LLM_KV_EXPERT_FEED_FORWARD_LENGTH, hparams.n_ff_exp); + ml.get_key(LLM_KV_EXPERT_SHARED_COUNT, hparams.n_expert_shared); + ml.get_key(LLM_KV_EXPERT_WEIGHTS_SCALE, hparams.expert_weights_scale); + ml.get_key(LLM_KV_EXPERT_WEIGHTS_NORM, hparams.expert_weights_norm); + ml.get_key_or_arr(LLM_KV_SWIGLU_CLAMP_EXP, hparams.swiglu_clamp_exp, hparams.n_layer); + ml.get_key_or_arr(LLM_KV_SWIGLU_CLAMP_SHEXP, hparams.swiglu_clamp_shexp, hparams.n_layer); + + ml.get_key(LLM_KV_ATTENTION_INDEXER_HEAD_COUNT, hparams.indexer_n_head); + ml.get_key(LLM_KV_ATTENTION_INDEXER_KEY_LENGTH, hparams.indexer_head_size); + ml.get_key(LLM_KV_ATTENTION_INDEXER_TOP_K, hparams.indexer_top_k); + + ml.get_key(dsv4_kv("attention.o_group_count"), hparams.dsv4_o_group_count); + ml.get_key(dsv4_kv("attention.o_lora_rank"), hparams.dsv4_o_lora_rank); + ml.get_key(dsv4_kv("attention.compress_rope.freq_base"), hparams.dsv4_compress_rope_base); + ml.get_key(dsv4_kv("hc.mult"), hparams.dsv4_hc_mult); + ml.get_key(dsv4_kv("hc.sinkhorn_iters"), hparams.dsv4_hc_sinkhorn_iters); + ml.get_key(dsv4_kv("hc.eps"), hparams.dsv4_hc_eps); + ml.get_key(dsv4_kv("moe.hash_layer_count"), hparams.dsv4_hash_layer_count); + + uint32_t n_compress_ratios = 0; + ml.get_arr_n(dsv4_kv("attention.compress_ratios"), n_compress_ratios); + if (n_compress_ratios < hparams.n_layer) { + throw std::runtime_error("DeepSeek-V4 Flash compress_ratios is shorter than block_count"); + } + ml.get_arr(dsv4_kv("attention.compress_ratios"), hparams.dsv4_compress_ratios); + + std::string score_func; + std::string topk_method; + ml.get_key(dsv4_kv("moe.score_func"), score_func); + ml.get_key(dsv4_kv("moe.topk_method"), topk_method); + if (score_func != "sqrtsoftplus") { + throw std::runtime_error("DeepSeek-V4 Flash loader currently expects sqrtsoftplus MoE scoring"); + } + if (topk_method != "noaux_tc") { + throw std::runtime_error("DeepSeek-V4 Flash loader currently expects noaux_tc MoE top-k"); + } + + hparams.swa_type = LLAMA_SWA_TYPE_STANDARD; + std::fill(hparams.swa_layers.begin(), hparams.swa_layers.begin() + hparams.n_layer, 1); + + switch (hparams.n_layer) { + case 43: type = LLM_TYPE_UNKNOWN; break; + default: type = LLM_TYPE_UNKNOWN; + } +} + +void llama_model_deepseek_v4_flash::load_arch_tensors(llama_model_loader &) { + LLAMA_LOAD_LOCALS; + + const int64_t q_lora_rank = hparams.n_lora_q; + const int64_t n_ff_exp = hparams.n_ff_exp; + const int64_t n_expert_shared = hparams.n_expert_shared; + + const int64_t n_embd_head = hparams.n_embd_head_k(); + const int64_t o_groups = hparams.dsv4_o_group_count; + const int64_t o_lora_rank = hparams.dsv4_o_lora_rank; + const int64_t hc_mult = hparams.dsv4_hc_mult; + const int64_t hc_dim = hc_mult * n_embd; + const int64_t hc_mix_dim = (2 + hc_mult) * hc_mult; + + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + hc_head_fn = create_tensor(tn(LLM_TENSOR_HC_HEAD_FN), {hc_dim, hc_mult}, 0); + hc_head_base = create_tensor(tn(LLM_TENSOR_HC_HEAD_BASE), {hc_mult}, 0); + hc_head_scale = create_tensor(tn(LLM_TENSOR_HC_HEAD_SCALE), {1}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_sinks = create_tensor(tn(LLM_TENSOR_ATTN_SINKS, nullptr, i), {n_head}, 0); + layer.wq_a = create_tensor(tn(LLM_TENSOR_ATTN_Q_A, "weight", i), {n_embd, q_lora_rank}, 0); + layer.attn_q_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_A_NORM, "weight", i), {q_lora_rank}, 0); + layer.wq_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_B, "weight", i), {q_lora_rank, n_head * n_embd_head}, 0); + layer.wkv = create_tensor(tn(LLM_TENSOR_ATTN_KV, "weight", i), {n_embd, n_embd_head}, 0); + layer.attn_kv_norm = create_tensor(tn(LLM_TENSOR_ATTN_KV_NORM, "weight", i), {n_embd_head}, 0); + layer.wo_a = create_tensor(tn(LLM_TENSOR_ATTN_OUT_A, "weight", i), {n_embd, o_lora_rank, o_groups}, 0); + layer.wo_b = create_tensor(tn(LLM_TENSOR_ATTN_OUT_B, "weight", i), {o_groups * o_lora_rank, n_embd}, 0); + + layer.hc_attn_fn = create_tensor(tn(LLM_TENSOR_HC_ATTN_FN, nullptr, i), {hc_dim, hc_mix_dim}, 0); + layer.hc_attn_base = create_tensor(tn(LLM_TENSOR_HC_ATTN_BASE, nullptr, i), {hc_mix_dim}, 0); + layer.hc_attn_scale = create_tensor(tn(LLM_TENSOR_HC_ATTN_SCALE, nullptr, i), {3}, 0); + layer.hc_ffn_fn = create_tensor(tn(LLM_TENSOR_HC_FFN_FN, nullptr, i), {hc_dim, hc_mix_dim}, 0); + layer.hc_ffn_base = create_tensor(tn(LLM_TENSOR_HC_FFN_BASE, nullptr, i), {hc_mix_dim}, 0); + layer.hc_ffn_scale = create_tensor(tn(LLM_TENSOR_HC_FFN_SCALE, nullptr, i), {3}, 0); + + const int64_t ratio = hparams.dsv4_compress_ratios[i]; + if (ratio != 0) { + const int64_t coff = ratio == 4 ? 2 : 1; + + layer.attn_comp_wkv = create_tensor(tn(LLM_TENSOR_ATTN_COMPRESSOR_WKV, "weight", i), {n_embd, coff * n_embd_head}, 0); + layer.attn_comp_wgate = create_tensor(tn(LLM_TENSOR_ATTN_COMPRESSOR_WGATE, "weight", i), {n_embd, coff * n_embd_head}, 0); + layer.attn_comp_ape = create_tensor(tn(LLM_TENSOR_ATTN_COMPRESSOR_APE, nullptr, i), {coff * n_embd_head, ratio}, 0); + layer.attn_comp_norm = create_tensor(tn(LLM_TENSOR_ATTN_COMPRESSOR_NORM, "weight", i), {n_embd_head}, 0); + + if (ratio == 4) { + const int64_t n_embd_indexer = hparams.indexer_head_size; + + layer.indexer_proj = create_tensor(tn(LLM_TENSOR_INDEXER_PROJ, "weight", i), {n_embd, hparams.indexer_n_head}, 0); + layer.indexer_attn_q_b = create_tensor(tn(LLM_TENSOR_INDEXER_ATTN_Q_B, "weight", i), {q_lora_rank, hparams.indexer_n_head * n_embd_indexer}, 0); + + layer.indexer_comp_wkv = create_tensor(tn(LLM_TENSOR_INDEXER_COMPRESSOR_WKV, "weight", i), {n_embd, 2 * n_embd_indexer}, 0); + layer.indexer_comp_wgate = create_tensor(tn(LLM_TENSOR_INDEXER_COMPRESSOR_WGATE, "weight", i), {n_embd, 2 * n_embd_indexer}, 0); + layer.indexer_comp_ape = create_tensor(tn(LLM_TENSOR_INDEXER_COMPRESSOR_APE, nullptr, i), {2 * n_embd_indexer, ratio}, 0); + layer.indexer_comp_norm = create_tensor(tn(LLM_TENSOR_INDEXER_COMPRESSOR_NORM, "weight", i), {n_embd_indexer}, 0); + } else if (ratio != 128) { + throw std::runtime_error("DeepSeek-V4 Flash loader only supports compression ratios 0, 4, and 128"); + } + } + + layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0); + if ((uint32_t) i < hparams.dsv4_hash_layer_count) { + layer.ffn_gate_tid2eid = create_tensor(tn(LLM_TENSOR_FFN_GATE_TID2EID, nullptr, i), {n_expert_used, n_vocab}, 0); + } else { + layer.ffn_exp_probs_b = create_tensor(tn(LLM_TENSOR_FFN_EXP_PROBS_B, "bias", i), {n_expert}, 0); + } + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff_exp, n_expert}, 0); + layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff_exp, n_embd, n_expert}, 0); + layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), {n_embd, n_ff_exp, n_expert}, 0); + + layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0); + layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), {n_ff_exp * n_expert_shared, n_embd }, 0); + layer.ffn_up_shexp = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP, "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0); + } +} + +std::unique_ptr llama_model_deepseek_v4_flash::build_arch_graph(const llm_graph_params & params) const { + return std::make_unique(*this, params); +} + +static size_t dsv4_elem_offset(const ggml_tensor * t, int64_t i) { + return ggml_row_size(t->type, i); +} + +static ggml_tensor * dsv4_view_1d(ggml_context * ctx, ggml_tensor * t, int64_t ne0, int64_t i0) { + return ggml_view_1d(ctx, t, ne0, dsv4_elem_offset(t, i0)); +} + +static ggml_tensor * dsv4_view_2d( + ggml_context * ctx, + ggml_tensor * t, + int64_t ne0, + int64_t ne1, + int64_t i0) { + return ggml_view_2d(ctx, t, ne0, ne1, t->nb[1], dsv4_elem_offset(t, i0)); +} + +static ggml_tensor * dsv4_hc_affine( + ggml_context * ctx, + ggml_tensor * x, + ggml_tensor * scale, + ggml_tensor * base) { + x = ggml_mul(ctx, x, scale); + x = ggml_add(ctx, x, base); + return x; +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hc_weighted_sum( + ggml_tensor * x, + ggml_tensor * weights) const { + const int64_t hc = hparams.dsv4_hc_mult; + const int64_t nt = x->ne[2]; + + ggml_tensor * acc = nullptr; + for (int64_t ih = 0; ih < hc; ++ih) { + ggml_tensor * xh = ggml_view_2d(ctx0, x, n_embd, nt, x->nb[2], ih*x->nb[1]); + ggml_tensor * wh = ggml_view_2d(ctx0, weights, 1, nt, weights->nb[1], ih*weights->nb[0]); + + ggml_tensor * cur = ggml_mul(ctx0, xh, wh); + acc = acc ? ggml_add(ctx0, acc, cur) : cur; + } + + return acc; +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hc_sinkhorn( + ggml_tensor * comb, + int il) const { + GGML_UNUSED(il); + + // comb is [dst_hc, src_hc, n_tokens]. Sinkhorn follows the reference: + // row softmax over dst, one column normalization, then repeated row/column normalization. + comb = ggml_soft_max(ctx0, comb); + + auto norm_cols = [&]() { + ggml_tensor * comb_src_dst = ggml_cont(ctx0, ggml_permute(ctx0, comb, 1, 0, 2, 3)); + ggml_tensor * col_sum = ggml_sum_rows(ctx0, comb_src_dst); + col_sum = ggml_clamp(ctx0, col_sum, hparams.dsv4_hc_eps, INFINITY); + col_sum = ggml_permute(ctx0, col_sum, 1, 0, 2, 3); + comb = ggml_div(ctx0, comb, col_sum); + }; + + auto norm_rows = [&]() { + ggml_tensor * row_sum = ggml_sum_rows(ctx0, comb); + row_sum = ggml_clamp(ctx0, row_sum, hparams.dsv4_hc_eps, INFINITY); + comb = ggml_div(ctx0, comb, row_sum); + }; + + norm_cols(); + for (uint32_t i = 1; i < hparams.dsv4_hc_sinkhorn_iters; ++i) { + norm_rows(); + norm_cols(); + } + + return comb; +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hc_pre( + ggml_tensor * x, + ggml_tensor * hc_fn, + ggml_tensor * hc_scale, + ggml_tensor * hc_base, + ggml_tensor ** post, + ggml_tensor ** comb, + int il) const { + const int64_t hc = hparams.dsv4_hc_mult; + const int64_t hc_dim = hc*n_embd; + const int64_t hc_mix_dim = (2 + hc)*hc; + const int64_t nt = x->ne[2]; + + GGML_ASSERT(hc == 4); + GGML_ASSERT(hc_fn->ne[1] == hc_mix_dim); + + ggml_tensor * flat = ggml_reshape_2d(ctx0, x, hc_dim, nt); + ggml_tensor * flat_norm = ggml_rms_norm(ctx0, flat, norm_rms_eps); + ggml_tensor * mixes = ggml_mul_mat(ctx0, hc_fn, flat_norm); + cb(mixes, "hc_mixes", il); + + ggml_tensor * scale_pre = dsv4_view_1d(ctx0, hc_scale, 1, 0); + ggml_tensor * scale_post = dsv4_view_1d(ctx0, hc_scale, 1, 1); + ggml_tensor * scale_comb = dsv4_view_1d(ctx0, hc_scale, 1, 2); + + ggml_tensor * base_pre = dsv4_view_1d(ctx0, hc_base, hc, 0); + ggml_tensor * base_post = dsv4_view_1d(ctx0, hc_base, hc, hc); + ggml_tensor * base_comb = dsv4_view_1d(ctx0, hc_base, hc*hc, 2*hc); + + ggml_tensor * pre = dsv4_view_2d(ctx0, mixes, hc, nt, 0); + pre = dsv4_hc_affine(ctx0, pre, scale_pre, base_pre); + pre = ggml_sigmoid(ctx0, pre); + pre = ggml_clamp(ctx0, pre, hparams.dsv4_hc_eps, INFINITY); + cb(pre, "hc_pre", il); + + *post = dsv4_view_2d(ctx0, mixes, hc, nt, hc); + *post = dsv4_hc_affine(ctx0, *post, scale_post, base_post); + *post = ggml_sigmoid(ctx0, *post); + *post = ggml_scale(ctx0, *post, 2.0f); + cb(*post, "hc_post", il); + + *comb = dsv4_view_2d(ctx0, mixes, hc*hc, nt, 2*hc); + *comb = dsv4_hc_affine(ctx0, *comb, scale_comb, base_comb); + *comb = ggml_reshape_3d(ctx0, *comb, hc, hc, nt); + *comb = build_hc_sinkhorn(*comb, il); + cb(*comb, "hc_comb", il); + + return build_hc_weighted_sum(x, pre); +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hc_post( + ggml_tensor * x, + ggml_tensor * residual, + ggml_tensor * post, + ggml_tensor * comb, + int il) const { + GGML_UNUSED(il); + + const int64_t hc = hparams.dsv4_hc_mult; + const int64_t nt = x->ne[1]; + + ggml_tensor * out = nullptr; + for (int64_t dst = 0; dst < hc; ++dst) { + ggml_tensor * post_dst = ggml_view_2d(ctx0, post, 1, nt, post->nb[1], dst*post->nb[0]); + ggml_tensor * cur = ggml_mul(ctx0, x, post_dst); + + for (int64_t src = 0; src < hc; ++src) { + ggml_tensor * res_src = ggml_view_2d(ctx0, residual, n_embd, nt, residual->nb[2], src*residual->nb[1]); + ggml_tensor * comb_src_dst = ggml_view_2d(ctx0, comb, 1, nt, comb->nb[2], dst*comb->nb[0] + src*comb->nb[1]); + cur = ggml_add(ctx0, cur, ggml_mul(ctx0, res_src, comb_src_dst)); + } + + cur = ggml_reshape_3d(ctx0, cur, n_embd, 1, nt); + out = out ? ggml_concat(ctx0, out, cur, 1) : cur; + } + + return out; +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hc_head( + ggml_tensor * x, + ggml_tensor * hc_fn, + ggml_tensor * hc_scale, + ggml_tensor * hc_base) const { + const int64_t hc = hparams.dsv4_hc_mult; + const int64_t hc_dim = hc*n_embd; + const int64_t nt = x->ne[2]; + + ggml_tensor * flat = ggml_reshape_2d(ctx0, x, hc_dim, nt); + ggml_tensor * flat_norm = ggml_rms_norm(ctx0, flat, norm_rms_eps); + ggml_tensor * mixes = ggml_mul_mat(ctx0, hc_fn, flat_norm); + cb(mixes, "hc_head_mixes", -1); + + ggml_tensor * pre = dsv4_hc_affine(ctx0, mixes, hc_scale, hc_base); + pre = ggml_sigmoid(ctx0, pre); + pre = ggml_clamp(ctx0, pre, hparams.dsv4_hc_eps, INFINITY); + cb(pre, "hc_head_pre", -1); + + return build_hc_weighted_sum(x, pre); +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( + const llama_model & model, + llm_graph_input_attn_no_cache * inp_attn, + ggml_tensor * cur, + ggml_tensor * inp_pos, + int il) const { + const auto & layer = model.layers[il]; + + const int64_t n_embd_head = hparams.n_embd_head_k(); + const int64_t n_embd_head_rope = hparams.n_rot(); + const int64_t n_embd_head_nope = n_embd_head - n_embd_head_rope; + const int64_t n_groups = hparams.dsv4_o_group_count; + const int64_t n_heads_group = n_head / n_groups; + const int64_t o_lora_rank = hparams.dsv4_o_lora_rank; + const int64_t o_group_dim = n_heads_group*n_embd_head; + const int64_t nt = cur->ne[1]; + + GGML_ASSERT(n_embd_head == n_embd_head_v); + GGML_ASSERT(n_head % n_groups == 0); + + const bool use_compress_rope = hparams.dsv4_compress_ratios[il] != 0; + const float freq_base_l = use_compress_rope ? hparams.dsv4_compress_rope_base : freq_base; + const float freq_scale_l = use_compress_rope ? freq_scale : 1.0f; + const float ext_factor_l = use_compress_rope ? ext_factor : 0.0f; + const float beta_fast_l = use_compress_rope ? beta_fast : 0.0f; + const float beta_slow_l = use_compress_rope ? beta_slow : 0.0f; + const int32_t n_ctx_orig_l = use_compress_rope ? n_ctx_orig : 0; + + ggml_tensor * qr = build_lora_mm(layer.wq_a, cur); + cb(qr, "qr", il); + + qr = build_norm(qr, layer.attn_q_a_norm, nullptr, LLM_NORM_RMS, il); + cb(qr, "qr_norm", il); + + ggml_tensor * q = build_lora_mm(layer.wq_b, qr); + q = ggml_reshape_3d(ctx0, q, n_embd_head, n_head, nt); + q = ggml_rms_norm(ctx0, q, norm_rms_eps); + cb(q, "q_norm", il); + + ggml_tensor * q_nope = ggml_view_3d(ctx0, q, n_embd_head_nope, n_head, nt, + ggml_row_size(q->type, n_embd_head), + ggml_row_size(q->type, n_embd_head)*n_head, + 0); + ggml_tensor * q_pe = ggml_view_3d(ctx0, q, n_embd_head_rope, n_head, nt, + ggml_row_size(q->type, n_embd_head), + ggml_row_size(q->type, n_embd_head)*n_head, + ggml_row_size(q->type, n_embd_head_nope)); + q_pe = ggml_rope_ext(ctx0, q_pe, inp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig_l, + freq_base_l, freq_scale_l, ext_factor_l, 1.0f, beta_fast_l, beta_slow_l); + cb(q_pe, "q_pe", il); + q = ggml_concat(ctx0, q_nope, q_pe, 0); + cb(q, "q", il); + + ggml_tensor * kv = build_lora_mm(layer.wkv, cur); + kv = build_norm(kv, layer.attn_kv_norm, nullptr, LLM_NORM_RMS, il); + kv = ggml_reshape_3d(ctx0, kv, n_embd_head, 1, nt); + cb(kv, "kv_norm", il); + + ggml_tensor * kv_nope = ggml_view_3d(ctx0, kv, n_embd_head_nope, 1, nt, + ggml_row_size(kv->type, n_embd_head), + ggml_row_size(kv->type, n_embd_head), + 0); + ggml_tensor * kv_pe = ggml_view_3d(ctx0, kv, n_embd_head_rope, 1, nt, + ggml_row_size(kv->type, n_embd_head), + ggml_row_size(kv->type, n_embd_head), + ggml_row_size(kv->type, n_embd_head_nope)); + kv_pe = ggml_rope_ext(ctx0, kv_pe, inp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig_l, + freq_base_l, freq_scale_l, ext_factor_l, 1.0f, beta_fast_l, beta_slow_l); + cb(kv_pe, "kv_pe", il); + kv = ggml_concat(ctx0, kv_nope, kv_pe, 0); + cb(kv, "kv", il); + + ggml_tensor * out = build_attn(inp_attn, + nullptr, nullptr, nullptr, + q, kv, kv, nullptr, layer.attn_sinks, nullptr, + 1.0f/sqrtf(float(n_embd_head)), il); + cb(out, "attn_raw", il); + + out = ggml_reshape_3d(ctx0, out, n_embd_head, n_head, nt); + ggml_tensor * out_nope = ggml_view_3d(ctx0, out, n_embd_head_nope, n_head, nt, + ggml_row_size(out->type, n_embd_head), + ggml_row_size(out->type, n_embd_head)*n_head, + 0); + ggml_tensor * out_pe = ggml_view_3d(ctx0, out, n_embd_head_rope, n_head, nt, + ggml_row_size(out->type, n_embd_head), + ggml_row_size(out->type, n_embd_head)*n_head, + ggml_row_size(out->type, n_embd_head_nope)); + out_pe = ggml_rope_ext_back(ctx0, out_pe, inp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig_l, + freq_base_l, freq_scale_l, ext_factor_l, 1.0f, beta_fast_l, beta_slow_l); + out = ggml_concat(ctx0, out_nope, out_pe, 0); + cb(out, "attn_derope", il); + + out = ggml_reshape_3d(ctx0, out, o_group_dim, n_groups, nt); + out = ggml_permute(ctx0, out, 0, 2, 1, 3); + ggml_tensor * oa = ggml_mul_mat(ctx0, layer.wo_a, out); + cb(oa, "attn_wo_a", il); + oa = ggml_permute(ctx0, oa, 0, 2, 1, 3); + oa = ggml_cont_2d(ctx0, oa, o_lora_rank*n_groups, nt); + + out = build_lora_mm(layer.wo_b, oa); + cb(out, "attn_out", il); + + return out; +} + +llama_model_deepseek_v4_flash::graph::graph(const llama_model & model, const llm_graph_params & params) : + llm_graph_context(params) { + ggml_tensor * cur; + + ggml_tensor * inp = build_inp_embd(model.tok_embd); + ggml_tensor * inp_pos = build_inp_pos(); + ggml_tensor * inp_out_ids = build_inp_out_ids(); + llm_graph_input_attn_no_cache * inp_attn = build_attn_inp_no_cache(); + ggml_build_forward_expand(gf, inp_attn->self_kq_mask); + if (inp_attn->self_kq_mask_swa) { + ggml_build_forward_expand(gf, inp_attn->self_kq_mask_swa); + } + + const int64_t hc = hparams.dsv4_hc_mult; + ggml_tensor * inpL = ggml_reshape_3d(ctx0, inp, n_embd, 1, n_tokens); + inpL = ggml_repeat_4d(ctx0, inpL, n_embd, hc, n_tokens, 1); + cb(inpL, "hc_init", -1); + + for (int il = 0; il < n_layer; ++il) { + ggml_tensor * residual = inpL; + ggml_tensor * post = nullptr; + ggml_tensor * comb = nullptr; + + cur = build_hc_pre(inpL, + model.layers[il].hc_attn_fn, + model.layers[il].hc_attn_scale, + model.layers[il].hc_attn_base, + &post, &comb, il); + cb(cur, "hc_attn_pre", il); + + cur = build_norm(cur, model.layers[il].attn_norm, nullptr, LLM_NORM_RMS, il); + cb(cur, "attn_norm", il); + + cur = build_attention(model, inp_attn, cur, inp_pos, il); + + inpL = build_hc_post(cur, residual, post, comb, il); + cb(inpL, "hc_attn_post", il); + + residual = inpL; + cur = build_hc_pre(inpL, + model.layers[il].hc_ffn_fn, + model.layers[il].hc_ffn_scale, + model.layers[il].hc_ffn_base, + &post, &comb, il); + cb(cur, "hc_ffn_pre", il); + + cur = build_norm(cur, model.layers[il].ffn_norm, nullptr, LLM_NORM_RMS, il); + cb(cur, "ffn_norm", il); + + const auto & layer = model.layers[il]; + ggml_tensor * selected_experts = nullptr; + ggml_tensor * exp_probs_b = layer.ffn_exp_probs_b; + if ((uint32_t) il < hparams.dsv4_hash_layer_count) { + selected_experts = ggml_get_rows(ctx0, layer.ffn_gate_tid2eid, res->t_inp_tokens); + exp_probs_b = nullptr; + } + + ggml_tensor * moe_out = build_moe_ffn(cur, + layer.ffn_gate_inp, + layer.ffn_up_exps, + layer.ffn_gate_exps, + layer.ffn_down_exps, + exp_probs_b, + n_expert, hparams.n_expert_used, + LLM_FFN_SILU, hparams.expert_weights_norm, + hparams.expert_weights_scale, + LLAMA_EXPERT_GATING_FUNC_TYPE_SQRT_SOFTPLUS, + il, + nullptr, + nullptr, + nullptr, + nullptr, + nullptr, + selected_experts); + cb(moe_out, "ffn_moe_out", il); + + ggml_tensor * ffn_shexp = build_ffn(cur, + layer.ffn_up_shexp, nullptr, nullptr, + layer.ffn_gate_shexp, nullptr, nullptr, + layer.ffn_down_shexp, nullptr, nullptr, + nullptr, LLM_FFN_SILU, LLM_FFN_PAR, il); + cb(ffn_shexp, "ffn_shexp", il); + + cur = ggml_add(ctx0, moe_out, ffn_shexp); + cb(cur, "ffn_out", il); + + inpL = build_hc_post(cur, residual, post, comb, il); + inpL = build_cvec(inpL, il); + cb(inpL, "l_out", il); + } + + if (inp_out_ids) { + ggml_tensor * flat = ggml_reshape_2d(ctx0, inpL, n_embd*hc, n_tokens); + flat = ggml_get_rows(ctx0, flat, inp_out_ids); + inpL = ggml_reshape_3d(ctx0, flat, n_embd, hc, n_outputs); + } + + cur = build_hc_head(inpL, model.hc_head_fn, model.hc_head_scale, model.hc_head_base); + cb(cur, "hc_head", -1); + + cur = build_norm(cur, model.output_norm, nullptr, LLM_NORM_RMS, -1); + cb(cur, "result_norm", -1); + res->t_embd = cur; + + cur = ggml_mul_mat(ctx0, model.output, cur); + cb(cur, "result_output", -1); + res->t_logits = cur; + + ggml_build_forward_expand(gf, cur); +} diff --git a/src/models/models.h b/src/models/models.h index 866e0d0be3ed..afe38c2b3a41 100644 --- a/src/models/models.h +++ b/src/models/models.h @@ -1055,6 +1055,56 @@ struct llama_model_deepseek32 : public llama_model_base { }; +struct llama_model_deepseek_v4_flash : public llama_model_base { + llama_model_deepseek_v4_flash(const struct llama_model_params & params) : llama_model_base(params) {} + void load_arch_hparams(llama_model_loader & ml) override; + void load_arch_tensors(llama_model_loader & ml) override; + + struct graph : public llm_graph_context { + graph(const llama_model & model, const llm_graph_params & params); + + ggml_tensor * build_hc_pre( + ggml_tensor * x, + ggml_tensor * hc_fn, + ggml_tensor * hc_scale, + ggml_tensor * hc_base, + ggml_tensor ** post, + ggml_tensor ** comb, + int il) const; + + ggml_tensor * build_hc_post( + ggml_tensor * x, + ggml_tensor * residual, + ggml_tensor * post, + ggml_tensor * comb, + int il) const; + + ggml_tensor * build_hc_head( + ggml_tensor * x, + ggml_tensor * hc_fn, + ggml_tensor * hc_scale, + ggml_tensor * hc_base) const; + + ggml_tensor * build_attention( + const llama_model & model, + llm_graph_input_attn_no_cache * inp_attn, + ggml_tensor * cur, + ggml_tensor * inp_pos, + int il) const; + + ggml_tensor * build_hc_weighted_sum( + ggml_tensor * x, + ggml_tensor * weights) const; + + ggml_tensor * build_hc_sinkhorn( + ggml_tensor * comb, + int il) const; + }; + + std::unique_ptr build_arch_graph(const llm_graph_params & params) const override; +}; + + struct llama_model_deepseek2ocr : public llama_model_base { llama_model_deepseek2ocr(const struct llama_model_params & params) : llama_model_base(params) {} void load_arch_hparams(llama_model_loader & ml) override; From 441b736659c53e8d421bd20927d5c1bf88d93412 Mon Sep 17 00:00:00 2001 From: Aman Gupta Date: Tue, 2 Jun 2026 14:23:05 +0200 Subject: [PATCH 05/13] add llm_graph_input_dsv4 --- src/CMakeLists.txt | 1 + src/llama-graph.cpp | 259 +++++++++++ src/llama-graph.h | 54 +++ src/llama-kv-cache-dsv4.cpp | 833 ++++++++++++++++++++++++++++++++++++ src/llama-kv-cache-dsv4.h | 259 +++++++++++ src/llama-kv-cache.cpp | 5 +- src/llama-model.cpp | 20 +- src/models/deepseek-v4.cpp | 827 ++++++++++++++++++++++++++++++++++- src/models/models.h | 77 +++- 9 files changed, 2323 insertions(+), 12 deletions(-) create mode 100644 src/llama-kv-cache-dsv4.cpp create mode 100644 src/llama-kv-cache-dsv4.h diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt index d15ccfd99f14..320784c3a8cc 100644 --- a/src/CMakeLists.txt +++ b/src/CMakeLists.txt @@ -25,6 +25,7 @@ add_library(llama llama-kv-cache.cpp llama-kv-cache-iswa.cpp llama-kv-cache-dsa.cpp + llama-kv-cache-dsv4.cpp llama-memory.cpp llama-memory-hybrid.cpp llama-memory-hybrid-iswa.cpp diff --git a/src/llama-graph.cpp b/src/llama-graph.cpp index b8bd0cd1729e..5b3dc2722b0f 100644 --- a/src/llama-graph.cpp +++ b/src/llama-graph.cpp @@ -8,6 +8,7 @@ #include "llama-kv-cache.h" #include "llama-kv-cache-iswa.h" #include "llama-kv-cache-dsa.h" +#include "llama-kv-cache-dsv4.h" #include "llama-memory-hybrid.h" #include "llama-memory-hybrid-iswa.h" #include "llama-memory-recurrent.h" @@ -17,6 +18,7 @@ #include #include #include +#include #include // dedup helpers @@ -620,6 +622,223 @@ bool llm_graph_input_attn_kv_iswa::can_reuse(const llm_graph_params & params) { return res; } +static void dsv4_set_i64(ggml_tensor * dst, const std::vector & src) { + if (!dst || !dst->buffer) { + return; + } + + GGML_ASSERT(dst->ne[0] == (int64_t) src.size()); + ggml_backend_tensor_set(dst, src.data(), 0, src.size()*ggml_element_size(dst)); +} + +static void dsv4_set_i32(ggml_tensor * dst, const std::vector & src) { + if (!dst || !dst->buffer) { + return; + } + + GGML_ASSERT(dst->ne[0] == (int64_t) src.size()); + ggml_backend_tensor_set(dst, src.data(), 0, src.size()*ggml_element_size(dst)); +} + +static void dsv4_set_kq_mask( + ggml_tensor * dst, + const llama_kv_cache_dsv4_context::comp_plan & plan, + uint32_t n_tokens) { + if (!dst || !dst->buffer) { + return; + } + + GGML_ASSERT(dst->type == GGML_TYPE_F32); + GGML_ASSERT(dst->ne[0] == plan.n_kv); + GGML_ASSERT(dst->ne[1] == (int64_t) n_tokens); + GGML_ASSERT(dst->ne[2] == 1); + GGML_ASSERT(dst->ne[3] == 1); + GGML_ASSERT((int64_t) plan.n_visible.size() == dst->ne[1]); + GGML_ASSERT(ggml_backend_buffer_is_host(dst->buffer)); + + float * data = (float *) dst->data; + + for (int64_t i = 0; i < dst->ne[1]; ++i) { + const int32_t n_visible = plan.n_visible[i]; + + for (int64_t j = 0; j < dst->ne[0]; ++j) { + data[i*dst->ne[0] + j] = j < n_visible ? 0.0f : -INFINITY; + } + } +} + +static std::string dsv4_plan_positions(const std::vector & values) { + std::ostringstream ss; + ss << "["; + for (size_t i = 0; i < values.size(); ++i) { + if (i > 0) { + ss << ", "; + } + ss << values[i]; + } + ss << "]"; + return ss.str(); +} + +static bool dsv4_compress_debug() { + static const bool debug = []() { + const char * env = getenv("LLAMA_DSV4_COMPRESS_DEBUG"); + return env && atoi(env) > 0; + }(); + + return debug; +} + +static void dsv4_set_comp_inputs( + const llm_graph_input_dsv4::comp_input & inp, + const llama_kv_cache_dsv4_context::comp_plan & plan, + const char * name, + bool debug, + uint32_t n_tokens) { + dsv4_set_i64(inp.write_idxs, plan.write_idxs); + dsv4_set_i32(inp.write_pos, plan.write_pos); + dsv4_set_i32(inp.write_end, plan.write_end); + dsv4_set_i32(inp.pending_end, plan.pending_end); + dsv4_set_i32(inp.state_idxs, plan.state_idxs); + dsv4_set_i32(inp.state_pos, plan.state_pos); + dsv4_set_i32(inp.state_read_idxs, plan.state_read_idxs); + dsv4_set_i64(inp.state_write_idxs, plan.state_write_idxs); + dsv4_set_i32(inp.state_write_pos, plan.state_write_pos); + dsv4_set_i32(inp.state_write_end, plan.state_write_end); + dsv4_set_i32(inp.n_visible, plan.n_visible); + dsv4_set_kq_mask(inp.kq_mask, plan, n_tokens); + + if (debug || dsv4_compress_debug()) { + LLAMA_LOG_INFO("%s: %s ratio=%u, n_tokens=%u, write_end=%s, state_write_end=%s, pending_end=%s\n", + __func__, name, plan.ratio, n_tokens, + dsv4_plan_positions(plan.write_end).c_str(), + dsv4_plan_positions(plan.state_write_end).c_str(), + dsv4_plan_positions(plan.pending_end).c_str()); + } +} + +static bool dsv4_can_reuse_tensor_1d(ggml_tensor * t, int64_t ne0) { + return (t == nullptr && ne0 == 0) || (t != nullptr && t->ne[0] == ne0); +} + +static bool dsv4_can_reuse_kq_mask( + ggml_tensor * t, + const llama_kv_cache_dsv4_context::comp_plan & plan, + uint32_t n_tokens) { + if (plan.n_kv == 0) { + return t == nullptr; + } + + return t != nullptr && + t->ne[0] == plan.n_kv && + t->ne[1] == (int64_t) n_tokens && + t->ne[2] == 1 && + t->ne[3] == 1; +} + +static bool dsv4_can_reuse_comp_input( + const llm_graph_input_dsv4::comp_input & inp, + const llama_kv_cache_dsv4_context::comp_plan & plan, + uint32_t n_tokens) { + const int64_t n_write = plan.write_idxs.size(); + + bool res = true; + res &= dsv4_can_reuse_tensor_1d(inp.write_idxs, n_write); + res &= dsv4_can_reuse_tensor_1d(inp.write_pos, n_write); + res &= dsv4_can_reuse_tensor_1d(inp.write_end, n_write); + res &= dsv4_can_reuse_tensor_1d(inp.pending_end, plan.pending_end.size()); + res &= dsv4_can_reuse_tensor_1d(inp.state_idxs, plan.state_idxs.size()); + res &= dsv4_can_reuse_tensor_1d(inp.state_pos, plan.state_pos.size()); + res &= dsv4_can_reuse_tensor_1d(inp.state_read_idxs, plan.state_read_idxs.size()); + res &= dsv4_can_reuse_tensor_1d(inp.state_write_idxs, plan.state_write_idxs.size()); + res &= dsv4_can_reuse_tensor_1d(inp.state_write_pos, plan.state_write_pos.size()); + res &= dsv4_can_reuse_tensor_1d(inp.state_write_end, plan.state_write_end.size()); + res &= dsv4_can_reuse_tensor_1d(inp.n_visible, plan.n_visible.size()); + res &= dsv4_can_reuse_kq_mask(inp.kq_mask, plan, n_tokens); + + return res; +} + +static ggml_tensor * dsv4_build_input_1d( + ggml_context * ctx, + ggml_type type, + int64_t ne0, + const std::string & name) { + if (ne0 == 0) { + return nullptr; + } + + ggml_tensor * res = ggml_new_tensor_1d(ctx, type, ne0); + ggml_set_input(res); + ggml_set_name(res, name.c_str()); + + return res; +} + +static void dsv4_build_comp_inputs( + ggml_context * ctx, + llm_graph_input_dsv4::comp_input & inp, + const llama_kv_cache_dsv4_context::comp_plan & plan, + const char * name) { + const int64_t n_write = plan.write_idxs.size(); + + inp.write_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I64, n_write, std::string("dsv4_") + name + "_write_idxs"); + inp.write_pos = dsv4_build_input_1d(ctx, GGML_TYPE_I32, n_write, std::string("dsv4_") + name + "_write_pos"); + inp.write_end = dsv4_build_input_1d(ctx, GGML_TYPE_I32, n_write, std::string("dsv4_") + name + "_write_end"); + inp.pending_end = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.pending_end.size(), std::string("dsv4_") + name + "_pending_end"); + inp.state_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_idxs.size(), std::string("dsv4_") + name + "_state_idxs"); + inp.state_pos = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_pos.size(), std::string("dsv4_") + name + "_state_pos"); + inp.state_read_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_read_idxs.size(), std::string("dsv4_") + name + "_state_read_idxs"); + inp.state_write_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I64, plan.state_write_idxs.size(), std::string("dsv4_") + name + "_state_write_idxs"); + inp.state_write_pos = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_write_pos.size(), std::string("dsv4_") + name + "_state_write_pos"); + inp.state_write_end = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_write_end.size(), std::string("dsv4_") + name + "_state_write_end"); + inp.n_visible = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.n_visible.size(), std::string("dsv4_") + name + "_n_visible"); + + if (plan.n_kv > 0) { + inp.kq_mask = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, plan.n_kv, plan.n_visible.size(), 1, 1); + ggml_set_input(inp.kq_mask); + ggml_set_name(inp.kq_mask, (std::string("dsv4_") + name + "_kq_mask").c_str()); + } +} + +void llm_graph_input_dsv4::set_input(const llama_ubatch * ubatch) { + inp_raw->mctx = mctx->get_raw(); + inp_raw->set_input(ubatch); + + dsv4_set_comp_inputs(inp_csa, mctx->get_csa_plan(), "csa", debug > 0, ubatch->n_tokens); + dsv4_set_comp_inputs(inp_hca, mctx->get_hca_plan(), "hca", debug > 0, ubatch->n_tokens); + dsv4_set_comp_inputs(inp_lid, mctx->get_lid_plan(), "lid", debug > 0, ubatch->n_tokens); + + if (inp_lid.k_rot && inp_lid.k_rot->buffer) { + mctx->get_lid()->set_input_k_rot(inp_lid.k_rot); + } +} + +bool llm_graph_input_dsv4::can_reuse(const llm_graph_params & params) { + const auto * mctx = static_cast(params.mctx); + + this->mctx = mctx; + inp_raw->mctx = mctx->get_raw(); + + bool res = true; + + if (inp_raw->self_k_idxs && inp_raw->self_k_idxs->buffer) { + res &= inp_raw->self_k_idxs->ne[0] == params.ubatch.n_tokens; + res &= can_reuse_kq_mask(inp_raw->self_kq_mask, mctx->get_raw()->get_base(), params.ubatch, params.cparams); + } + + if (inp_raw->self_k_idxs_swa && inp_raw->self_k_idxs_swa->buffer) { + res &= inp_raw->self_k_idxs_swa->ne[0] == params.ubatch.n_tokens; + res &= can_reuse_kq_mask(inp_raw->self_kq_mask_swa, mctx->get_raw()->get_swa(), params.ubatch, params.cparams); + } + + res &= dsv4_can_reuse_comp_input(inp_csa, mctx->get_csa_plan(), params.ubatch.n_tokens); + res &= dsv4_can_reuse_comp_input(inp_hca, mctx->get_hca_plan(), params.ubatch.n_tokens); + res &= dsv4_can_reuse_comp_input(inp_lid, mctx->get_lid_plan(), params.ubatch.n_tokens); + + return res; +} + void llm_graph_input_attn_cross::set_input(const llama_ubatch * ubatch) { GGML_ASSERT(cross_kq_mask); @@ -2731,6 +2950,46 @@ llm_graph_input_attn_kv_iswa * llm_graph_context::build_attn_inp_kv_iswa() const return (llm_graph_input_attn_kv_iswa *) res->add_input(std::move(inp)); } +llm_graph_input_dsv4 * llm_graph_context::build_inp_dsv4() const { + const auto * mctx_cur = static_cast(mctx); + const auto * raw_ctx = mctx_cur->get_raw(); + + auto inp_raw = std::make_unique(hparams, cparams, raw_ctx); + + { + inp_raw->self_k_idxs = raw_ctx->get_base()->build_input_k_idxs(ctx0, ubatch); + inp_raw->self_v_idxs = raw_ctx->get_base()->build_input_v_idxs(ctx0, ubatch); + + inp_raw->self_kq_mask = build_attn_inp_kq_mask(ctx0, raw_ctx->get_base(), ubatch, cparams); + inp_raw->self_kq_mask_cnv = inp_raw->self_kq_mask; + } + + { + GGML_ASSERT(hparams.swa_type != LLAMA_SWA_TYPE_NONE && "DSV4 expects SWA raw cache"); + + inp_raw->self_k_idxs_swa = raw_ctx->get_swa()->build_input_k_idxs(ctx0, ubatch); + inp_raw->self_v_idxs_swa = raw_ctx->get_swa()->build_input_v_idxs(ctx0, ubatch); + + inp_raw->self_kq_mask_swa = build_attn_inp_kq_mask(ctx0, raw_ctx->get_swa(), ubatch, cparams); + inp_raw->self_kq_mask_swa_cnv = inp_raw->self_kq_mask_swa; + } + + inp_raw->self_k_rot = raw_ctx->get_base()->build_input_k_rot(ctx0); + inp_raw->self_v_rot = raw_ctx->get_base()->build_input_v_rot(ctx0); + + inp_raw->self_k_rot_swa = raw_ctx->get_swa()->build_input_k_rot(ctx0); + inp_raw->self_v_rot_swa = raw_ctx->get_swa()->build_input_v_rot(ctx0); + + auto inp = std::make_unique(cparams, std::move(inp_raw), mctx_cur); + + dsv4_build_comp_inputs(ctx0, inp->inp_csa, mctx_cur->get_csa_plan(), "csa"); + dsv4_build_comp_inputs(ctx0, inp->inp_hca, mctx_cur->get_hca_plan(), "hca"); + dsv4_build_comp_inputs(ctx0, inp->inp_lid, mctx_cur->get_lid_plan(), "lid"); + inp->inp_lid.k_rot = mctx_cur->get_lid()->build_input_k_rot(ctx0); + + return (llm_graph_input_dsv4 *) res->add_input(std::move(inp)); +} + ggml_tensor * llm_graph_context::build_rs( ggml_tensor * s, ggml_tensor * state_copy_main, diff --git a/src/llama-graph.h b/src/llama-graph.h index f14ccd3734dd..58906534ad7f 100644 --- a/src/llama-graph.h +++ b/src/llama-graph.h @@ -23,6 +23,7 @@ struct llama_memory_context_i; class llama_kv_cache_context; class llama_kv_cache_dsa_context; +class llama_kv_cache_dsv4_context; class llama_kv_cache_iswa_context; class llama_memory_recurrent_context; class llama_memory_hybrid_context; @@ -459,6 +460,57 @@ class llm_graph_input_attn_kv_iswa : public llm_graph_input_i { const llama_kv_cache_iswa_context * mctx; }; +class llm_graph_input_dsv4 : public llm_graph_input_i { +public: + struct comp_input { + ggml_tensor * write_idxs = nullptr; // I64 [n_write] + ggml_tensor * write_pos = nullptr; // I32 [n_write] + ggml_tensor * write_end = nullptr; // I32 [n_write] + ggml_tensor * pending_end = nullptr; // I32 [n_pending] + + ggml_tensor * state_idxs = nullptr; // I32 [n_state] + ggml_tensor * state_pos = nullptr; // I32 [n_state] + ggml_tensor * state_read_idxs = nullptr; // I32 [ratio*n_state_write] + ggml_tensor * state_write_idxs = nullptr; // I64 [n_state_write] + ggml_tensor * state_write_pos = nullptr; // I32 [n_state_write] + ggml_tensor * state_write_end = nullptr; // I32 [n_state_write] + + ggml_tensor * n_visible = nullptr; // I32 [n_batch] + ggml_tensor * kq_mask = nullptr; // F32 [n_kv, n_batch] + + ggml_tensor * k_rot = nullptr; + }; + + llm_graph_input_dsv4( + const llama_cparams & cparams, + std::unique_ptr inp_raw, + const llama_kv_cache_dsv4_context * mctx) : + inp_raw(std::move(inp_raw)), + cparams(cparams), + mctx(mctx) { + } + ~llm_graph_input_dsv4() = default; + + void set_input(const llama_ubatch * ubatch) override; + + bool can_reuse(const llm_graph_params & params) override; + + llm_graph_input_attn_kv_iswa * get_raw() const { return inp_raw.get(); } + const comp_input & get_csa() const { return inp_csa; } + const comp_input & get_hca() const { return inp_hca; } + const comp_input & get_lid() const { return inp_lid; } + + std::unique_ptr inp_raw; + + comp_input inp_csa; + comp_input inp_hca; + comp_input inp_lid; + + const llama_cparams cparams; + + const llama_kv_cache_dsv4_context * mctx; +}; + class llm_graph_input_attn_cross : public llm_graph_input_i { public: llm_graph_input_attn_cross(const llama_cross * cross) : cross(cross) {} @@ -1034,6 +1086,8 @@ struct llm_graph_context { llm_graph_input_attn_kv_iswa * build_attn_inp_kv_iswa() const; + llm_graph_input_dsv4 * build_inp_dsv4() const; + // note: if k_cur or v_cur are not provided, they will not be stored in the memory ggml_tensor * build_attn( llm_graph_input_attn_kv_iswa * inp, diff --git a/src/llama-kv-cache-dsv4.cpp b/src/llama-kv-cache-dsv4.cpp new file mode 100644 index 000000000000..e0de2d91d728 --- /dev/null +++ b/src/llama-kv-cache-dsv4.cpp @@ -0,0 +1,833 @@ +#include "llama-kv-cache-dsv4.h" + +#include "ggml-backend.h" +#include "llama-impl.h" +#include "llama-batch.h" +#include "llama-model.h" + +#include +#include +#include +#include +#include +#include +#include +#include + +static constexpr uint32_t DSV4_CSA_RATIO = 4; +static constexpr uint32_t DSV4_HCA_RATIO = 128; + +static uint32_t dsv4_comp_size(uint32_t kv_size, uint32_t ratio) { + return std::max(1, (kv_size + ratio - 1)/ratio); +} + +static std::string dsv4_plan_positions(const std::vector & values) { + std::ostringstream ss; + ss << "["; + for (size_t i = 0; i < values.size(); ++i) { + if (i > 0) { + ss << ", "; + } + ss << values[i]; + } + ss << "]"; + return ss.str(); +} + +static llama_kv_cache_dsv4_context::comp_plan dsv4_build_comp_plan( + const llama_ubatch & ubatch, + uint32_t ratio, + bool overlap, + bool stateful, + uint32_t state_size, + uint32_t kv_size, + uint32_t n_stream) { + llama_kv_cache_dsv4_context::comp_plan plan; + plan.ratio = ratio; + plan.n_visible.resize(ubatch.n_tokens); + + const int64_t state_rows = (int64_t) state_size*n_stream; + + const auto current_token_idx = [&](llama_seq_id seq_id, llama_pos pos) -> int64_t { + for (uint32_t i = 0; i < ubatch.n_tokens; ++i) { + if (ubatch.pos[i] == pos && ubatch.seq_id[i][0] == seq_id) { + return i; + } + } + + return -1; + }; + + const auto state_source_idx = [&](llama_seq_id seq_id, llama_pos pos) -> int32_t { + if (pos < 0) { + // The overlap compressor needs a zero/-inf source for the first + // block's previous half. The graph appends that row after the + // current-ubatch scratch rows. + return (int32_t) (state_rows + ubatch.n_tokens); + } + + const int64_t tok_idx = current_token_idx(seq_id, pos); + if (tok_idx >= 0) { + return (int32_t) (state_rows + tok_idx); + } + + const int64_t stream_off = n_stream > 1 ? (int64_t) seq_id*state_size : 0; + return (int32_t) (stream_off + pos%state_size); + }; + + for (uint32_t i = 0; i < ubatch.n_tokens; ++i) { + const llama_pos pos = ubatch.pos[i]; + + if (pos < 0) { + continue; + } + + const llama_seq_id seq_id = ubatch.seq_id[i][0]; + + if (stateful) { + const int64_t stream_off = n_stream > 1 ? (int64_t) seq_id*state_size : 0; + + plan.state_idxs.push_back((int32_t) (stream_off + pos%state_size)); + plan.state_pos .push_back((int32_t) (pos%ratio)); + } + + const int64_t n_visible = (int64_t) (pos + 1)/ratio; + plan.n_visible[i] = (int32_t) n_visible; + plan.n_kv = std::max(plan.n_kv, n_visible); + + if ((pos + 1) % ratio != 0) { + continue; + } + + const llama_pos source_start = pos + 1 - ratio; + + if (stateful) { + const int64_t cache_off = n_stream > 1 ? (int64_t) seq_id*kv_size : 0; + + plan.state_write_idxs.push_back(cache_off + pos/ratio); + plan.state_write_pos .push_back((int32_t) source_start); + plan.state_write_end .push_back((int32_t) pos); + + if (overlap) { + const llama_pos prev_start = source_start - ratio; + + for (uint32_t j = 0; j < ratio; ++j) { + plan.state_read_idxs.push_back(state_source_idx(seq_id, prev_start + j)); + } + for (uint32_t j = 0; j < ratio; ++j) { + plan.state_read_idxs.push_back(state_source_idx(seq_id, source_start + j)); + } + } else { + for (uint32_t j = 0; j < ratio; ++j) { + plan.state_read_idxs.push_back(state_source_idx(seq_id, source_start + j)); + } + } + + continue; + } + + const int64_t stream_off = n_stream > 1 ? (int64_t) seq_id*kv_size : 0; + + plan.write_idxs.push_back(stream_off + pos/ratio); + plan.write_pos .push_back((int32_t) (pos + 1 - ratio)); + plan.write_end .push_back((int32_t) pos); + } + + static const bool debug = []() { + const char * env = getenv("LLAMA_DSV4_COMPRESS_DEBUG"); + return env && atoi(env) > 0; + }(); + + if (debug) { + LLAMA_LOG_INFO("%s: ratio=%u, n_tokens=%u, write_end=%s, state_write_end=%s, pending_end=%s\n", + __func__, ratio, ubatch.n_tokens, + dsv4_plan_positions(plan.write_end).c_str(), + dsv4_plan_positions(plan.state_write_end).c_str(), + dsv4_plan_positions(plan.pending_end).c_str()); + } + + return plan; +} + +static std::vector dsv4_build_comp_plans( + const std::vector & ubatches, + uint32_t ratio, + bool overlap, + bool stateful, + uint32_t state_size, + uint32_t kv_size, + uint32_t n_stream) { + std::vector plans; + plans.reserve(ubatches.size()); + + for (const llama_ubatch & ubatch : ubatches) { + plans.push_back(dsv4_build_comp_plan(ubatch, ratio, overlap, stateful, state_size, kv_size, n_stream)); + } + + return plans; +} + +static void dsv4_make_k_only(llama_hparams & hparams) { + // llama_kv_cache uses hparams.is_mla() to allocate K-only storage. + hparams.n_embd_head_k_mla_impl = hparams.n_embd_head_k(); + hparams.n_embd_head_v_mla_impl = hparams.n_embd_head_k(); +} + +// +// llama_dsv4_comp_state +// + +llama_dsv4_comp_state::llama_dsv4_comp_state( + const llama_model & model, + bool offload, + bool unified, + uint32_t n_seq_max, + uint32_t ratio, + uint32_t state_size, + uint32_t n_embd_state, + const char * name, + const llama_memory_i::layer_filter_cb & filter) : + ratio(ratio), + state_size(state_size), + n_embd_state(n_embd_state), + n_stream(unified ? 1 : n_seq_max) { + const llama_hparams & hparams = model.hparams; + + struct ggml_backend_buft_comparator { + bool operator()(const ggml_backend_buffer_type_t & lhs, const ggml_backend_buffer_type_t & rhs) const { + return strcmp(ggml_backend_buft_name(lhs), ggml_backend_buft_name(rhs)) < 0; + } + }; + + std::map ctx_map; + + auto ctx_for_buft = [&](ggml_backend_buffer_type_t buft) -> ggml_context * { + auto it = ctx_map.find(buft); + if (it == ctx_map.end()) { + ggml_init_params params = { + /*.mem_size =*/ size_t(2u*hparams.n_layer*ggml_tensor_overhead()), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + + ggml_context * ctx = ggml_init(params); + if (!ctx) { + return nullptr; + } + + ctx_map.emplace(buft, ctx); + + return ctx; + } + + return it->second.get(); + }; + + for (uint32_t il = 0; il < hparams.n_layer; ++il) { + if (filter && !filter(il)) { + continue; + } + + const char * dev_name = "CPU"; + + ggml_backend_buffer_type_t buft = ggml_backend_cpu_buffer_type(); + + if (offload) { + auto * dev = model.dev_layer(il); + buft = ggml_backend_dev_buffer_type(dev); + + dev_name = ggml_backend_dev_name(dev); + } + + LLAMA_LOG_DEBUG("%s: layer %3d: dev = %s\n", __func__, il, dev_name); + + ggml_context * ctx = ctx_for_buft(buft); + if (!ctx) { + throw std::runtime_error("failed to create ggml context for DSV4 compressor state"); + } + + ggml_tensor * kv = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, n_embd_state, state_size, n_stream); + ggml_tensor * score = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, n_embd_state, state_size, n_stream); + + ggml_format_name(kv, "dsv4_%s_state_kv_l%d", name, il); + ggml_format_name(score, "dsv4_%s_state_score_l%d", name, il); + + map_layer_ids[il] = layers.size(); + + layers.push_back({ il, kv, score }); + } + + for (auto & [buft, ctx] : ctx_map) { + ggml_backend_buffer_t buf = ggml_backend_alloc_ctx_tensors_from_buft(ctx.get(), buft); + if (!buf) { + throw std::runtime_error("failed to allocate buffer for DSV4 compressor state"); + } + + ggml_backend_buffer_clear(buf, 0); + + LLAMA_LOG_INFO("%s: %10s DSV4 %s state buffer size = %8.2f MiB\n", + __func__, ggml_backend_buffer_name(buf), name, ggml_backend_buffer_get_size(buf)/1024.0/1024.0); + + ctxs_bufs.emplace_back(std::move(ctx), buf); + } + + LLAMA_LOG_INFO("%s: %s ratio = %u, state = %u x %u, streams = %u, layers = %zu, size = %7.2f MiB\n", + __func__, name, ratio, state_size, n_embd_state, n_stream, layers.size(), total_size()/1024.0/1024.0); +} + +void llama_dsv4_comp_state::clear(bool data) { + if (!data) { + return; + } + + for (auto & [_, buf] : ctxs_bufs) { + ggml_backend_buffer_clear(buf.get(), 0); + } +} + +uint32_t llama_dsv4_comp_state::get_ratio() const { + return ratio; +} + +uint32_t llama_dsv4_comp_state::get_state_size() const { + return state_size; +} + +uint32_t llama_dsv4_comp_state::get_n_stream() const { + return n_stream; +} + +std::map llama_dsv4_comp_state::memory_breakdown() const { + std::map ret; + for (const auto & [_, buf] : ctxs_bufs) { + ggml_backend_buffer_type_t buft = ggml_backend_buffer_get_type(buf.get()); + ret[buft] += ggml_backend_buffer_get_size(buf.get()); + } + return ret; +} + +ggml_tensor * llama_dsv4_comp_state::get_kv(ggml_context * ctx, int32_t il) const { + const int32_t ids = map_layer_ids.at(il); + + ggml_tensor * state = layers[ids].kv; + + return ggml_reshape_2d(ctx, state, state->ne[0], state->ne[1]*state->ne[2]); +} + +ggml_tensor * llama_dsv4_comp_state::get_score(ggml_context * ctx, int32_t il) const { + const int32_t ids = map_layer_ids.at(il); + + ggml_tensor * state = layers[ids].score; + + return ggml_reshape_2d(ctx, state, state->ne[0], state->ne[1]*state->ne[2]); +} + +ggml_tensor * llama_dsv4_comp_state::cpy_kv(ggml_context * ctx, ggml_tensor * cur, ggml_tensor * idxs, int32_t il) const { + return ggml_set_rows(ctx, get_kv(ctx, il), cur, idxs); +} + +ggml_tensor * llama_dsv4_comp_state::cpy_score(ggml_context * ctx, ggml_tensor * cur, ggml_tensor * idxs, int32_t il) const { + return ggml_set_rows(ctx, get_score(ctx, il), cur, idxs); +} + +size_t llama_dsv4_comp_state::total_size() const { + size_t size = 0; + + for (const auto & [_, buf] : ctxs_bufs) { + size += ggml_backend_buffer_get_size(buf.get()); + } + + return size; +} + +// +// llama_kv_cache_dsv4 +// + +llama_kv_cache_dsv4::llama_kv_cache_dsv4( + const llama_model & model, + ggml_type type_k, + ggml_type type_v, + bool v_trans, + bool offload, + bool swa_full, + bool unified, + uint32_t kv_size, + uint32_t n_seq_max, + uint32_t n_ubatch, + uint32_t n_pad, + const layer_filter_cb & filter, + const layer_reuse_cb & reuse) : + hparams_csa(model.hparams), + hparams_hca(model.hparams), + hparams_lid(model.hparams) { + + const layer_filter_cb filter_raw = [&](int32_t il) { + if (filter && !filter(il)) { + return false; + } + + return true; + }; + + LLAMA_LOG_INFO("%s: creating DSV4 raw KV cache\n", __func__); + + kv_raw = std::make_unique( + model, type_k, type_v, + v_trans, offload, swa_full, unified, kv_size, n_seq_max, n_ubatch, n_pad, + filter_raw, reuse); + + dsv4_make_k_only(hparams_csa); + dsv4_make_k_only(hparams_hca); + + std::fill(hparams_lid.n_head_kv_arr.begin(), hparams_lid.n_head_kv_arr.end(), 1); + hparams_lid.n_embd_head_k_full = model.hparams.indexer_head_size; + hparams_lid.n_embd_head_v_full = model.hparams.indexer_head_size; + hparams_lid.n_embd_head_k_swa = model.hparams.indexer_head_size; + hparams_lid.n_embd_head_v_swa = model.hparams.indexer_head_size; + hparams_lid.rope_type = LLAMA_ROPE_TYPE_NEOX; + dsv4_make_k_only(hparams_lid); + + const layer_filter_cb filter_csa = [&](int32_t il) { + if (filter && !filter(il)) { + return false; + } + + return model.hparams.dsv4_compress_ratios[il] == DSV4_CSA_RATIO; + }; + + const layer_filter_cb filter_hca = [&](int32_t il) { + if (filter && !filter(il)) { + return false; + } + + return model.hparams.dsv4_compress_ratios[il] == DSV4_HCA_RATIO; + }; + + LLAMA_LOG_INFO("%s: creating DSV4 CSA compressed KV cache, size = %u cells\n", + __func__, dsv4_comp_size(kv_size, DSV4_CSA_RATIO)); + + kv_csa = std::make_unique( + model, hparams_csa, type_k, type_v, + v_trans, offload, unified, dsv4_comp_size(kv_size, DSV4_CSA_RATIO), n_seq_max, n_pad, + 0, LLAMA_SWA_TYPE_NONE, filter_csa, nullptr); + + LLAMA_LOG_INFO("%s: creating DSV4 HCA compressed KV cache, size = %u cells\n", + __func__, dsv4_comp_size(kv_size, DSV4_HCA_RATIO)); + + kv_hca = std::make_unique( + model, hparams_hca, type_k, type_v, + v_trans, offload, unified, dsv4_comp_size(kv_size, DSV4_HCA_RATIO), n_seq_max, n_pad, + 0, LLAMA_SWA_TYPE_NONE, filter_hca, nullptr); + + LLAMA_LOG_INFO("%s: creating DSV4 lightning-indexer KV cache, size = %u cells\n", + __func__, dsv4_comp_size(kv_size, DSV4_CSA_RATIO)); + + kv_lid = std::make_unique( + model, hparams_lid, type_k, type_v, + v_trans, offload, unified, dsv4_comp_size(kv_size, DSV4_CSA_RATIO), n_seq_max, n_pad, + 0, LLAMA_SWA_TYPE_NONE, filter_csa, nullptr); + + LLAMA_LOG_INFO("%s: creating DSV4 CSA compressor state\n", __func__); + + csa_state = std::make_unique( + model, offload, unified, n_seq_max, DSV4_CSA_RATIO, 2*DSV4_CSA_RATIO, + 2*model.hparams.n_embd_head_k(), "csa", filter_csa); + + LLAMA_LOG_INFO("%s: creating DSV4 HCA compressor state\n", __func__); + + hca_state = std::make_unique( + model, offload, unified, n_seq_max, DSV4_HCA_RATIO, DSV4_HCA_RATIO, + model.hparams.n_embd_head_k(), "hca", filter_hca); + + LLAMA_LOG_INFO("%s: creating DSV4 lightning-indexer compressor state\n", __func__); + + lid_state = std::make_unique( + model, offload, unified, n_seq_max, DSV4_CSA_RATIO, 2*DSV4_CSA_RATIO, + 2*model.hparams.indexer_head_size, "lid", filter_csa); +} + +llama_memory_context_ptr llama_kv_cache_dsv4::init_batch( + llama_batch_allocr & balloc, + uint32_t n_ubatch, + bool embd_all) { + GGML_UNUSED(embd_all); + + // Match llama_kv_cache_iswa splitting so the raw path remains identical. + do { + if (kv_raw->get_base()->get_n_stream() != 1) { + break; + } + + balloc.split_reset(); + + std::vector ubatches; + while (true) { + auto ubatch = balloc.split_simple(n_ubatch); + + if (ubatch.n_tokens == 0) { + break; + } + + ubatches.push_back(std::move(ubatch)); // NOLINT + } + + if (balloc.get_n_used() < balloc.get_n_tokens()) { + break; + } + + auto sinfos_raw_base = kv_raw->get_base()->prepare(ubatches); + if (sinfos_raw_base.empty()) { + break; + } + + auto sinfos_raw_swa = kv_raw->get_swa()->prepare(ubatches); + if (sinfos_raw_swa.empty()) { + break; + } + + return std::make_unique( + this, std::move(sinfos_raw_base), std::move(sinfos_raw_swa), std::move(ubatches)); + } while (false); + + do { + balloc.split_reset(); + + std::vector ubatches; + while (true) { + auto ubatch = balloc.split_equal(n_ubatch, kv_raw->get_base()->get_n_stream() != 1); + + if (ubatch.n_tokens == 0) { + break; + } + + ubatches.push_back(std::move(ubatch)); // NOLINT + } + + if (balloc.get_n_used() < balloc.get_n_tokens()) { + break; + } + + auto sinfos_raw_base = kv_raw->get_base()->prepare(ubatches); + if (sinfos_raw_base.empty()) { + break; + } + + auto sinfos_raw_swa = kv_raw->get_swa()->prepare(ubatches); + if (sinfos_raw_swa.empty()) { + break; + } + + return std::make_unique( + this, std::move(sinfos_raw_base), std::move(sinfos_raw_swa), std::move(ubatches)); + } while (false); + + return std::make_unique(LLAMA_MEMORY_STATUS_FAILED_PREPARE); +} + +llama_memory_context_ptr llama_kv_cache_dsv4::init_full() { + return std::make_unique(this); +} + +llama_memory_context_ptr llama_kv_cache_dsv4::init_update(llama_context * lctx, bool optimize) { + return std::make_unique(this, lctx, optimize); +} + +bool llama_kv_cache_dsv4::get_can_shift() const { + // Compressed row metadata uses block-derived positions. Keep shifting + // disabled until DSV4 compressed-cache shift semantics are wired. + return false; +} + +void llama_kv_cache_dsv4::clear(bool data) { + kv_raw->clear(data); + clear_compressed(data); +} + +bool llama_kv_cache_dsv4::seq_rm(llama_seq_id seq_id, llama_pos p0, llama_pos p1) { + if (p1 >= 0) { + return false; + } + + if (p0 > 0) { + // DSV4 compressed cache rows are derived from running compressor state, + // so arbitrary rollback is not reconstructible from the raw cache alone. + // Allow the common prompt-cache cleanup no-op: remove [end, infinity). + if (seq_id >= 0 && p0 > kv_raw->seq_pos_max(seq_id)) { + return true; + } + + return false; + } + + const bool res = kv_raw->seq_rm(seq_id, p0, p1); + + if (res) { + clear_compressed(false); + } + + return res; +} + +void llama_kv_cache_dsv4::seq_cp(llama_seq_id seq_id_src, llama_seq_id seq_id_dst, llama_pos p0, llama_pos p1) { + kv_raw->seq_cp(seq_id_src, seq_id_dst, p0, p1); + clear_compressed(false); +} + +void llama_kv_cache_dsv4::seq_keep(llama_seq_id seq_id) { + kv_raw->seq_keep(seq_id); + clear_compressed(false); +} + +void llama_kv_cache_dsv4::seq_add(llama_seq_id seq_id, llama_pos p0, llama_pos p1, llama_pos shift) { + kv_raw->seq_add(seq_id, p0, p1, shift); + clear_compressed(false); +} + +void llama_kv_cache_dsv4::seq_div(llama_seq_id seq_id, llama_pos p0, llama_pos p1, int d) { + kv_raw->seq_div(seq_id, p0, p1, d); + clear_compressed(false); +} + +llama_pos llama_kv_cache_dsv4::seq_pos_min(llama_seq_id seq_id) const { + return kv_raw->seq_pos_min(seq_id); +} + +llama_pos llama_kv_cache_dsv4::seq_pos_max(llama_seq_id seq_id) const { + return kv_raw->seq_pos_max(seq_id); +} + +std::map llama_kv_cache_dsv4::memory_breakdown() const { + std::map mb = kv_raw->memory_breakdown(); + for (const auto & buft_size : kv_csa->memory_breakdown()) { + mb[buft_size.first] += buft_size.second; + } + for (const auto & buft_size : kv_hca->memory_breakdown()) { + mb[buft_size.first] += buft_size.second; + } + for (const auto & buft_size : kv_lid->memory_breakdown()) { + mb[buft_size.first] += buft_size.second; + } + for (const auto & buft_size : csa_state->memory_breakdown()) { + mb[buft_size.first] += buft_size.second; + } + for (const auto & buft_size : hca_state->memory_breakdown()) { + mb[buft_size.first] += buft_size.second; + } + for (const auto & buft_size : lid_state->memory_breakdown()) { + mb[buft_size.first] += buft_size.second; + } + return mb; +} + +void llama_kv_cache_dsv4::state_write(llama_io_write_i & io, llama_seq_id seq_id, llama_state_seq_flags flags) const { + kv_raw->state_write(io, seq_id, flags); +} + +void llama_kv_cache_dsv4::state_read(llama_io_read_i & io, llama_seq_id seq_id, llama_state_seq_flags flags) { + kv_raw->state_read(io, seq_id, flags); + clear_compressed(false); +} + +llama_kv_cache_iswa * llama_kv_cache_dsv4::get_raw() const { + return kv_raw.get(); +} + +llama_kv_cache * llama_kv_cache_dsv4::get_csa() const { + return kv_csa.get(); +} + +llama_kv_cache * llama_kv_cache_dsv4::get_hca() const { + return kv_hca.get(); +} + +llama_kv_cache * llama_kv_cache_dsv4::get_lid() const { + return kv_lid.get(); +} + +llama_dsv4_comp_state * llama_kv_cache_dsv4::get_csa_state() const { + return csa_state.get(); +} + +llama_dsv4_comp_state * llama_kv_cache_dsv4::get_hca_state() const { + return hca_state.get(); +} + +llama_dsv4_comp_state * llama_kv_cache_dsv4::get_lid_state() const { + return lid_state.get(); +} + +void llama_kv_cache_dsv4::clear_compressed(bool data) { + kv_csa->clear(data); + kv_hca->clear(data); + kv_lid->clear(data); + csa_state->clear(data); + hca_state->clear(data); + lid_state->clear(data); +} + +// +// llama_kv_cache_dsv4_context +// + +llama_kv_cache_dsv4_context::llama_kv_cache_dsv4_context(llama_memory_status status) : status(status) {} + +llama_kv_cache_dsv4_context::llama_kv_cache_dsv4_context( + llama_kv_cache_dsv4 * kv) : + ctx_raw(kv->get_raw()->init_full()), + ctx_csa(kv->get_csa()->init_full()), + ctx_hca(kv->get_hca()->init_full()), + ctx_lid(kv->get_lid()->init_full()), + csa_state(kv->get_csa_state()), + hca_state(kv->get_hca_state()), + lid_state(kv->get_lid_state()), + status(llama_memory_status_combine( + llama_memory_status_combine(ctx_raw->get_status(), ctx_csa->get_status()), + llama_memory_status_combine(ctx_hca->get_status(), ctx_lid->get_status()))) { +} + +llama_kv_cache_dsv4_context::llama_kv_cache_dsv4_context( + llama_kv_cache_dsv4 * kv, + llama_context * lctx, + bool optimize) : + ctx_raw(kv->get_raw()->init_update(lctx, optimize)), + ctx_csa(kv->get_csa()->init_update(lctx, optimize)), + ctx_hca(kv->get_hca()->init_update(lctx, optimize)), + ctx_lid(kv->get_lid()->init_update(lctx, optimize)), + csa_state(kv->get_csa_state()), + hca_state(kv->get_hca_state()), + lid_state(kv->get_lid_state()), + status(llama_memory_status_combine( + llama_memory_status_combine(ctx_raw->get_status(), ctx_csa->get_status()), + llama_memory_status_combine(ctx_hca->get_status(), ctx_lid->get_status()))) { +} + +llama_kv_cache_dsv4_context::llama_kv_cache_dsv4_context( + llama_kv_cache_dsv4 * kv, + slot_info_vec_t sinfos_raw_base, + slot_info_vec_t sinfos_raw_swa, + std::vector ubatches) : + ubatches(std::move(ubatches)), + plans_csa(dsv4_build_comp_plans(this->ubatches, DSV4_CSA_RATIO, true, true, + kv->get_csa_state()->get_state_size(), kv->get_csa()->get_size(), kv->get_csa_state()->get_n_stream())), + plans_hca(dsv4_build_comp_plans(this->ubatches, DSV4_HCA_RATIO, false, true, + kv->get_hca_state()->get_state_size(), kv->get_hca()->get_size(), kv->get_hca_state()->get_n_stream())), + plans_lid(plans_csa), + ctx_raw(new llama_kv_cache_iswa_context(kv->get_raw(), std::move(sinfos_raw_base), std::move(sinfos_raw_swa), this->ubatches)), + ctx_csa(new llama_kv_cache_context(kv->get_csa())), + ctx_hca(new llama_kv_cache_context(kv->get_hca())), + ctx_lid(new llama_kv_cache_context(kv->get_lid())), + csa_state(kv->get_csa_state()), + hca_state(kv->get_hca_state()), + lid_state(kv->get_lid_state()), + status(ctx_raw->get_status()) { +} + +llama_kv_cache_dsv4_context::~llama_kv_cache_dsv4_context() = default; + +bool llama_kv_cache_dsv4_context::next() { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + ctx_raw->next(); + + if (++i_next >= ubatches.size()) { + return false; + } + + return true; +} + +bool llama_kv_cache_dsv4_context::apply() { + assert(!llama_memory_status_is_fail(status)); + + bool res = true; + + res = res & ctx_raw->apply(); + + return res; +} + +llama_memory_status llama_kv_cache_dsv4_context::get_status() const { + return status; +} + +const llama_ubatch & llama_kv_cache_dsv4_context::get_ubatch() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + return ubatches[i_next]; +} + +const llama_kv_cache_iswa_context * llama_kv_cache_dsv4_context::get_raw() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + return static_cast(ctx_raw.get()); +} + +const llama_kv_cache_context * llama_kv_cache_dsv4_context::get_csa() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + return static_cast(ctx_csa.get()); +} + +const llama_kv_cache_context * llama_kv_cache_dsv4_context::get_hca() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + return static_cast(ctx_hca.get()); +} + +const llama_kv_cache_context * llama_kv_cache_dsv4_context::get_lid() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + return static_cast(ctx_lid.get()); +} + +const llama_dsv4_comp_state * llama_kv_cache_dsv4_context::get_csa_state() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + return csa_state; +} + +const llama_dsv4_comp_state * llama_kv_cache_dsv4_context::get_hca_state() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + return hca_state; +} + +const llama_dsv4_comp_state * llama_kv_cache_dsv4_context::get_lid_state() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + return lid_state; +} + +const llama_kv_cache_dsv4_context::comp_plan & llama_kv_cache_dsv4_context::get_csa_plan() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + static const comp_plan empty; + if (plans_csa.empty()) { + return empty; + } + + return plans_csa[i_next]; +} + +const llama_kv_cache_dsv4_context::comp_plan & llama_kv_cache_dsv4_context::get_hca_plan() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + static const comp_plan empty; + if (plans_hca.empty()) { + return empty; + } + + return plans_hca[i_next]; +} + +const llama_kv_cache_dsv4_context::comp_plan & llama_kv_cache_dsv4_context::get_lid_plan() const { + assert(status == LLAMA_MEMORY_STATUS_SUCCESS); + + static const comp_plan empty; + if (plans_lid.empty()) { + return empty; + } + + return plans_lid[i_next]; +} diff --git a/src/llama-kv-cache-dsv4.h b/src/llama-kv-cache-dsv4.h new file mode 100644 index 000000000000..57bf4caf7019 --- /dev/null +++ b/src/llama-kv-cache-dsv4.h @@ -0,0 +1,259 @@ +#pragma once + +#include "llama-kv-cache.h" +#include "llama-kv-cache-iswa.h" + +#include +#include +#include +#include + +class llama_dsv4_comp_state { +public: + llama_dsv4_comp_state( + const llama_model & model, + bool offload, + bool unified, + uint32_t n_seq_max, + uint32_t ratio, + uint32_t state_size, + uint32_t n_embd_state, + const char * name, + const llama_memory_i::layer_filter_cb & filter); + + void clear(bool data); + + uint32_t get_ratio() const; + uint32_t get_state_size() const; + uint32_t get_n_stream() const; + + std::map memory_breakdown() const; + + ggml_tensor * get_kv (ggml_context * ctx, int32_t il) const; + ggml_tensor * get_score(ggml_context * ctx, int32_t il) const; + + ggml_tensor * cpy_kv (ggml_context * ctx, ggml_tensor * cur, ggml_tensor * idxs, int32_t il) const; + ggml_tensor * cpy_score(ggml_context * ctx, ggml_tensor * cur, ggml_tensor * idxs, int32_t il) const; + +private: + struct layer { + uint32_t il; + + ggml_tensor * kv; + ggml_tensor * score; + }; + + const uint32_t ratio; + const uint32_t state_size; + const uint32_t n_embd_state; + const uint32_t n_stream; + + std::vector> ctxs_bufs; + + std::vector layers; + + std::unordered_map map_layer_ids; + + size_t total_size() const; +}; + +// +// llama_kv_cache_dsv4 +// + +// DSV4 uses a normal raw/SWA token cache plus compressed K-only block caches. +// The compressed caches are storage only; DSV4-specific visibility and block +// planning are handled by llama_kv_cache_dsv4_context / llm_graph_input_dsv4. + +class llama_kv_cache_dsv4 : public llama_memory_i { +public: + llama_kv_cache_dsv4( + const llama_model & model, + ggml_type type_k, + ggml_type type_v, + bool v_trans, + bool offload, + bool swa_full, + bool unified, + uint32_t kv_size, + uint32_t n_seq_max, + uint32_t n_ubatch, + uint32_t n_pad, + const layer_filter_cb & filter, + const layer_reuse_cb & reuse); + + ~llama_kv_cache_dsv4() = default; + + // + // llama_memory_i + // + + llama_memory_context_ptr init_batch( + llama_batch_allocr & balloc, + uint32_t n_ubatch, + bool embd_all) override; + + llama_memory_context_ptr init_full() override; + + llama_memory_context_ptr init_update(llama_context * lctx, bool optimize) override; + + bool get_can_shift() const override; + + void clear(bool data) override; + + bool seq_rm (llama_seq_id seq_id, llama_pos p0, llama_pos p1) override; + void seq_cp (llama_seq_id seq_id_src, llama_seq_id seq_id_dst, llama_pos p0, llama_pos p1) override; + void seq_keep(llama_seq_id seq_id) override; + void seq_add (llama_seq_id seq_id, llama_pos p0, llama_pos p1, llama_pos shift) override; + void seq_div (llama_seq_id seq_id, llama_pos p0, llama_pos p1, int d) override; + + llama_pos seq_pos_min(llama_seq_id seq_id) const override; + llama_pos seq_pos_max(llama_seq_id seq_id) const override; + + std::map memory_breakdown() const override; + + void state_write(llama_io_write_i & io, llama_seq_id seq_id = -1, llama_state_seq_flags flags = 0) const override; + void state_read (llama_io_read_i & io, llama_seq_id seq_id = -1, llama_state_seq_flags flags = 0) override; + + // + // llama_kv_cache_dsv4 specific API + // + + llama_kv_cache_iswa * get_raw() const; + llama_kv_cache * get_csa() const; + llama_kv_cache * get_hca() const; + llama_kv_cache * get_lid() const; + llama_dsv4_comp_state * get_csa_state() const; + llama_dsv4_comp_state * get_hca_state() const; + llama_dsv4_comp_state * get_lid_state() const; + +private: + llama_hparams hparams_csa; + llama_hparams hparams_hca; + llama_hparams hparams_lid; + + std::unique_ptr kv_raw; + std::unique_ptr kv_csa; + std::unique_ptr kv_hca; + std::unique_ptr kv_lid; + std::unique_ptr csa_state; + std::unique_ptr hca_state; + std::unique_ptr lid_state; + + void clear_compressed(bool data); +}; + +class llama_kv_cache_dsv4_context : public llama_memory_context_i { +public: + using slot_info_vec_t = llama_kv_cache::slot_info_vec_t; + + struct comp_plan { + uint32_t ratio = 0; + + // Logical compressed row ids written by the current graph. + std::vector write_idxs; + + // Position used for compressor RoPE. For a completed block this is the + // first source-token position of that block. + std::vector write_pos; + + // Position at which the compressed row becomes visible to attention. + std::vector write_end; + + // Completed blocks that could not be planned. This should remain empty + // for the scratch-backed state path. + std::vector pending_end; + + // Compressor-state row ids updated by the current graph. + std::vector state_idxs; + + // APE row ids, i.e. pos % ratio, for the compressor-state updates. + std::vector state_pos; + + // Flattened source row ids used for state-backed commits. Source rows + // index the graph-local [persistent_state | current_ubatch_scratch] + // tensor. For overlapped compression the first half is previous rows + // and the second half is current rows; a final synthetic zero/-inf row + // may be addressed for the first block's previous half. + std::vector state_read_idxs; + + // Final compressed-cache row ids written by state-backed commits. + std::vector state_write_idxs; + + // RoPE positions for state-backed commits. + std::vector state_write_pos; + + // End positions for state-backed commits. + std::vector state_write_end; + + // Number of completed compressed rows visible for each query token. + std::vector n_visible; + + // Maximum compressed rows visible to this ubatch. + int64_t n_kv = 0; + }; + + llama_kv_cache_dsv4_context(llama_memory_status status); + + llama_kv_cache_dsv4_context( + llama_kv_cache_dsv4 * kv); + + llama_kv_cache_dsv4_context( + llama_kv_cache_dsv4 * kv, + llama_context * lctx, + bool optimize); + + llama_kv_cache_dsv4_context( + llama_kv_cache_dsv4 * kv, + slot_info_vec_t sinfos_raw_base, + slot_info_vec_t sinfos_raw_swa, + std::vector ubatches); + + virtual ~llama_kv_cache_dsv4_context(); + + // + // llama_memory_context_i + // + + bool next() override; + bool apply() override; + + llama_memory_status get_status() const override; + const llama_ubatch & get_ubatch() const override; + + // + // llama_kv_cache_dsv4_context specific API + // + + const llama_kv_cache_iswa_context * get_raw() const; + const llama_kv_cache_context * get_csa() const; + const llama_kv_cache_context * get_hca() const; + const llama_kv_cache_context * get_lid() const; + const llama_dsv4_comp_state * get_csa_state() const; + const llama_dsv4_comp_state * get_hca_state() const; + const llama_dsv4_comp_state * get_lid_state() const; + + const comp_plan & get_csa_plan() const; + const comp_plan & get_hca_plan() const; + const comp_plan & get_lid_plan() const; + +private: + size_t i_next = 0; + + std::vector ubatches; + + std::vector plans_csa; + std::vector plans_hca; + std::vector plans_lid; + + const llama_memory_context_ptr ctx_raw; + const llama_memory_context_ptr ctx_csa; + const llama_memory_context_ptr ctx_hca; + const llama_memory_context_ptr ctx_lid; + + const llama_dsv4_comp_state * csa_state = nullptr; + const llama_dsv4_comp_state * hca_state = nullptr; + const llama_dsv4_comp_state * lid_state = nullptr; + + const llama_memory_status status; +}; diff --git a/src/llama-kv-cache.cpp b/src/llama-kv-cache.cpp index 82da38e0b611..03bc81935f9d 100644 --- a/src/llama-kv-cache.cpp +++ b/src/llama-kv-cache.cpp @@ -294,8 +294,9 @@ llama_kv_cache::llama_kv_cache( ggml_is_quantized(type_k) && hparams.n_embd_head_k() % 64 == 0; - // always create Hadamard rotation tensors for DeepSeek V3.2 DSA lightning indexer - if (model.arch == LLM_ARCH_DEEPSEEK32 && hparams.n_embd_head_k_full == hparams.indexer_head_size) { + // always create Hadamard rotation tensors for DeepSeek lightning indexers + if ((model.arch == LLM_ARCH_DEEPSEEK32 || model.arch == LLM_ARCH_DEEPSEEK_V4_FLASH) && + hparams.n_embd_head_k_full == hparams.indexer_head_size) { attn_rot_k = true; } diff --git a/src/llama-model.cpp b/src/llama-model.cpp index 010af1e0728a..e6e4cd42f088 100644 --- a/src/llama-model.cpp +++ b/src/llama-model.cpp @@ -11,6 +11,7 @@ #include "llama-kv-cache.h" #include "llama-kv-cache-iswa.h" #include "llama-kv-cache-dsa.h" +#include "llama-kv-cache-dsv4.h" #include "llama-memory-hybrid.h" #include "llama-memory-hybrid-iswa.h" #include "llama-memory-recurrent.h" @@ -2124,7 +2125,24 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params, } } - if (hparams.swa_type != LLAMA_SWA_TYPE_NONE) { + if (arch == LLM_ARCH_DEEPSEEK_V4_FLASH) { + GGML_ASSERT(hparams.swa_type != LLAMA_SWA_TYPE_NONE); + + res = new llama_kv_cache_dsv4( + *this, + params.type_k, + params.type_v, + !cparams.flash_attn, + cparams.offload_kqv, + params.swa_full, + cparams.kv_unified, + cparams.n_ctx_seq, + cparams.n_seq_max, + cparams.n_ubatch, + 1, + filter, + reuse); + } else if (hparams.swa_type != LLAMA_SWA_TYPE_NONE) { GGML_ASSERT(hparams.is_swa_any()); res = new llama_kv_cache_iswa( diff --git a/src/models/deepseek-v4.cpp b/src/models/deepseek-v4.cpp index cd35bf60aaba..d67b19fab29b 100644 --- a/src/models/deepseek-v4.cpp +++ b/src/models/deepseek-v4.cpp @@ -1,5 +1,7 @@ #include "models.h" +#include "llama-kv-cache-dsv4.h" + #include #include #include @@ -166,6 +168,27 @@ static ggml_tensor * dsv4_view_2d( return ggml_view_2d(ctx, t, ne0, ne1, t->nb[1], dsv4_elem_offset(t, i0)); } +static ggml_tensor * dsv4_append_zero_row(ggml_context * ctx, ggml_tensor * t, bool neg_inf) { + ggml_tensor * row = ggml_view_1d(ctx, t, t->ne[0], 0); + row = neg_inf ? ggml_scale_bias(ctx, row, 0.0f, -INFINITY) : ggml_scale(ctx, row, 0.0f); + row = ggml_reshape_2d(ctx, row, t->ne[0], 1); + + return ggml_concat(ctx, t, row, 1); +} + +static ggml_tensor * dsv4_with_zero_dep(ggml_context * ctx, ggml_tensor * t, ggml_tensor * dep) { + if (dep == nullptr) { + return t; + } + + ggml_tensor * zero = ggml_scale(ctx, ggml_sum(ctx, dep), 0.0f); + return ggml_add(ctx, t, zero); +} + +static constexpr int64_t DSV4_CSA_RATIO = 4; +static constexpr int64_t DSV4_HCA_RATIO = 128; +static constexpr int64_t DSV4_SWA_WINDOW = 128; + static ggml_tensor * dsv4_hc_affine( ggml_context * ctx, ggml_tensor * x, @@ -327,13 +350,532 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hc_head( return build_hc_weighted_sum(x, pre); } +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_compressed_kv( + ggml_tensor * cur, + ggml_tensor * comp_pos, + ggml_tensor * wkv, + ggml_tensor * wgate, + ggml_tensor * ape, + ggml_tensor * norm, + int64_t ratio, + int64_t n_embd_head, + bool overlap, + const char * name, + int il) const { + const int64_t n_embd_head_rope = hparams.n_rot(); + const int64_t n_embd_head_nope = n_embd_head - n_embd_head_rope; + const int64_t nt = cur->ne[1]; + const int64_t n_blocks = comp_pos ? comp_pos->ne[0] : 0; + const int64_t n_complete = n_blocks*ratio; + const int64_t coff = overlap ? 2 : 1; + + GGML_ASSERT(n_blocks > 0); + GGML_ASSERT(nt >= n_complete); + GGML_ASSERT(n_embd_head >= n_embd_head_rope); + + ggml_tensor * kv = build_lora_mm(wkv, cur); + kv = ggml_cont(ctx0, ggml_cast(ctx0, kv, GGML_TYPE_F32)); + cb(kv, name, il); + + ggml_tensor * score = build_lora_mm(wgate, cur); + score = ggml_cont(ctx0, ggml_cast(ctx0, score, GGML_TYPE_F32)); + cb(score, name, il); + + if (ape->type != GGML_TYPE_F32) { + ape = ggml_cast(ctx0, ape, GGML_TYPE_F32); + } + + kv = ggml_view_2d(ctx0, kv, coff*n_embd_head, n_complete, kv->nb[1], 0); + kv = ggml_reshape_3d(ctx0, kv, coff*n_embd_head, ratio, n_blocks); + + score = ggml_view_2d(ctx0, score, coff*n_embd_head, n_complete, score->nb[1], 0); + score = ggml_reshape_3d(ctx0, score, coff*n_embd_head, ratio, n_blocks); + score = ggml_add(ctx0, score, ape); + + ggml_tensor * values = kv; + ggml_tensor * scores = score; + if (overlap) { + ggml_tensor * kv_prev_src = ggml_view_3d(ctx0, kv, n_embd_head, ratio, n_blocks, + kv->nb[1], kv->nb[2], 0); + ggml_tensor * kv_cur = ggml_view_3d(ctx0, kv, n_embd_head, ratio, n_blocks, + kv->nb[1], kv->nb[2], ggml_row_size(kv->type, n_embd_head)); + + ggml_tensor * score_prev_src = ggml_view_3d(ctx0, score, n_embd_head, ratio, n_blocks, + score->nb[1], score->nb[2], 0); + ggml_tensor * score_cur = ggml_view_3d(ctx0, score, n_embd_head, ratio, n_blocks, + score->nb[1], score->nb[2], ggml_row_size(score->type, n_embd_head)); + + ggml_tensor * kv_prev_head = ggml_cont(ctx0, ggml_view_3d(ctx0, kv_prev_src, n_embd_head, ratio, 1, + kv_prev_src->nb[1], kv_prev_src->nb[2], 0)); + ggml_tensor * score_prev_head = ggml_cont(ctx0, ggml_view_3d(ctx0, score_prev_src, n_embd_head, ratio, 1, + score_prev_src->nb[1], score_prev_src->nb[2], 0)); + + ggml_tensor * kv_prev_zero = ggml_scale(ctx0, kv_prev_head, 0.0f); + ggml_tensor * score_prev_zero = ggml_scale(ctx0, score_prev_head, 0.0f); + ggml_tensor * score_prev_inf = ggml_scale_bias(ctx0, score_prev_zero, 0.0f, -INFINITY); + + ggml_tensor * kv_prev = kv_prev_zero; + ggml_tensor * score_prev = score_prev_inf; + if (n_blocks > 1) { + ggml_tensor * kv_prev_tail = ggml_view_3d(ctx0, kv_prev_src, n_embd_head, ratio, n_blocks - 1, + kv_prev_src->nb[1], kv_prev_src->nb[2], 0); + ggml_tensor * score_prev_tail = ggml_view_3d(ctx0, score_prev_src, n_embd_head, ratio, n_blocks - 1, + score_prev_src->nb[1], score_prev_src->nb[2], 0); + + kv_prev = ggml_concat(ctx0, kv_prev_zero, kv_prev_tail, 2); + score_prev = ggml_concat(ctx0, score_prev_inf, score_prev_tail, 2); + } + + values = ggml_concat(ctx0, kv_prev, kv_cur, 1); + scores = ggml_concat(ctx0, score_prev, score_cur, 1); + } + + values = ggml_cont(ctx0, ggml_permute(ctx0, values, 1, 0, 2, 3)); + + scores = ggml_cont(ctx0, ggml_permute(ctx0, scores, 1, 0, 2, 3)); + + ggml_tensor * weights = ggml_soft_max(ctx0, scores); + ggml_tensor * comp = ggml_mul(ctx0, values, weights); + comp = ggml_sum_rows(ctx0, comp); + comp = ggml_cont(ctx0, ggml_permute(ctx0, comp, 1, 0, 2, 3)); + cb(comp, name, il); + + comp = build_norm(comp, norm, nullptr, LLM_NORM_RMS, il); + cb(comp, name, il); + + ggml_tensor * comp_nope = ggml_view_3d(ctx0, comp, n_embd_head_nope, 1, n_blocks, + ggml_row_size(comp->type, n_embd_head), + ggml_row_size(comp->type, n_embd_head), + 0); + ggml_tensor * comp_pe = ggml_view_3d(ctx0, comp, n_embd_head_rope, 1, n_blocks, + ggml_row_size(comp->type, n_embd_head), + ggml_row_size(comp->type, n_embd_head), + ggml_row_size(comp->type, n_embd_head_nope)); + + comp_pe = ggml_rope_ext(ctx0, comp_pe, comp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig, + hparams.dsv4_compress_rope_base, freq_scale, ext_factor, 1.0f, beta_fast, beta_slow); + cb(comp_pe, name, il); + + comp = ggml_concat(ctx0, comp_nope, comp_pe, 0); + cb(comp, name, il); + + return comp; +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hca_compressed_kv_from_state( + ggml_tensor * kv_state, + ggml_tensor * score_state, + ggml_tensor * state_read_idxs, + ggml_tensor * comp_pos, + ggml_tensor * norm, + int64_t n_embd_head, + const char * name, + int il) const { + const int64_t n_embd_head_rope = hparams.n_rot(); + const int64_t n_embd_head_nope = n_embd_head - n_embd_head_rope; + const int64_t n_blocks = comp_pos ? comp_pos->ne[0] : 0; + + GGML_ASSERT(n_blocks > 0); + GGML_ASSERT(state_read_idxs); + GGML_ASSERT(state_read_idxs->ne[0] == DSV4_HCA_RATIO*n_blocks); + GGML_ASSERT(n_embd_head >= n_embd_head_rope); + + ggml_tensor * kv = ggml_get_rows(ctx0, kv_state, state_read_idxs); + kv = ggml_reshape_3d(ctx0, kv, n_embd_head, DSV4_HCA_RATIO, n_blocks); + cb(kv, name, il); + + ggml_tensor * score = ggml_get_rows(ctx0, score_state, state_read_idxs); + score = ggml_reshape_3d(ctx0, score, n_embd_head, DSV4_HCA_RATIO, n_blocks); + cb(score, name, il); + + ggml_tensor * values = ggml_cont(ctx0, ggml_permute(ctx0, kv, 1, 0, 2, 3)); + ggml_tensor * scores = ggml_cont(ctx0, ggml_permute(ctx0, score, 1, 0, 2, 3)); + + ggml_tensor * weights = ggml_soft_max(ctx0, scores); + ggml_tensor * comp = ggml_mul(ctx0, values, weights); + comp = ggml_sum_rows(ctx0, comp); + comp = ggml_cont(ctx0, ggml_permute(ctx0, comp, 1, 0, 2, 3)); + cb(comp, name, il); + + comp = build_norm(comp, norm, nullptr, LLM_NORM_RMS, il); + cb(comp, name, il); + + ggml_tensor * comp_nope = ggml_view_3d(ctx0, comp, n_embd_head_nope, 1, n_blocks, + ggml_row_size(comp->type, n_embd_head), + ggml_row_size(comp->type, n_embd_head), + 0); + ggml_tensor * comp_pe = ggml_view_3d(ctx0, comp, n_embd_head_rope, 1, n_blocks, + ggml_row_size(comp->type, n_embd_head), + ggml_row_size(comp->type, n_embd_head), + ggml_row_size(comp->type, n_embd_head_nope)); + + comp_pe = ggml_rope_ext(ctx0, comp_pe, comp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig, + hparams.dsv4_compress_rope_base, freq_scale, ext_factor, 1.0f, beta_fast, beta_slow); + cb(comp_pe, name, il); + + comp = ggml_concat(ctx0, comp_nope, comp_pe, 0); + cb(comp, name, il); + + return comp; +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_overlap_compressed_kv_from_state( + ggml_tensor * kv_state, + ggml_tensor * score_state, + ggml_tensor * state_read_idxs, + ggml_tensor * comp_pos, + ggml_tensor * norm, + int64_t ratio, + int64_t n_embd_head, + const char * name, + int il) const { + const int64_t n_embd_head_rope = hparams.n_rot(); + const int64_t n_embd_head_nope = n_embd_head - n_embd_head_rope; + const int64_t n_blocks = comp_pos ? comp_pos->ne[0] : 0; + + GGML_ASSERT(n_blocks > 0); + GGML_ASSERT(state_read_idxs); + GGML_ASSERT(state_read_idxs->ne[0] == 2*ratio*n_blocks); + GGML_ASSERT(kv_state->ne[0] == 2*n_embd_head); + GGML_ASSERT(score_state->ne[0] == 2*n_embd_head); + GGML_ASSERT(n_embd_head >= n_embd_head_rope); + + kv_state = dsv4_append_zero_row(ctx0, kv_state, false); + score_state = dsv4_append_zero_row(ctx0, score_state, true); + + ggml_tensor * prev_idxs = dsv4_view_1d(ctx0, state_read_idxs, ratio*n_blocks, 0); + ggml_tensor * cur_idxs = dsv4_view_1d(ctx0, state_read_idxs, ratio*n_blocks, ratio*n_blocks); + + ggml_tensor * kv_prev = ggml_get_rows(ctx0, kv_state, prev_idxs); + kv_prev = ggml_cont(ctx0, ggml_view_2d(ctx0, kv_prev, n_embd_head, ratio*n_blocks, kv_prev->nb[1], 0)); + kv_prev = ggml_reshape_3d(ctx0, kv_prev, n_embd_head, ratio, n_blocks); + cb(kv_prev, name, il); + + ggml_tensor * score_prev = ggml_get_rows(ctx0, score_state, prev_idxs); + score_prev = ggml_cont(ctx0, ggml_view_2d(ctx0, score_prev, n_embd_head, ratio*n_blocks, score_prev->nb[1], 0)); + score_prev = ggml_reshape_3d(ctx0, score_prev, n_embd_head, ratio, n_blocks); + cb(score_prev, name, il); + + ggml_tensor * kv_cur = ggml_get_rows(ctx0, kv_state, cur_idxs); + kv_cur = ggml_cont(ctx0, ggml_view_2d(ctx0, kv_cur, n_embd_head, ratio*n_blocks, kv_cur->nb[1], + ggml_row_size(kv_cur->type, n_embd_head))); + kv_cur = ggml_reshape_3d(ctx0, kv_cur, n_embd_head, ratio, n_blocks); + + ggml_tensor * score_cur = ggml_get_rows(ctx0, score_state, cur_idxs); + score_cur = ggml_cont(ctx0, ggml_view_2d(ctx0, score_cur, n_embd_head, ratio*n_blocks, score_cur->nb[1], + ggml_row_size(score_cur->type, n_embd_head))); + score_cur = ggml_reshape_3d(ctx0, score_cur, n_embd_head, ratio, n_blocks); + + ggml_tensor * values = ggml_concat(ctx0, kv_prev, kv_cur, 1); + ggml_tensor * scores = ggml_concat(ctx0, score_prev, score_cur, 1); + + values = ggml_cont(ctx0, ggml_permute(ctx0, values, 1, 0, 2, 3)); + scores = ggml_cont(ctx0, ggml_permute(ctx0, scores, 1, 0, 2, 3)); + + ggml_tensor * weights = ggml_soft_max(ctx0, scores); + ggml_tensor * comp = ggml_mul(ctx0, values, weights); + comp = ggml_sum_rows(ctx0, comp); + comp = ggml_cont(ctx0, ggml_permute(ctx0, comp, 1, 0, 2, 3)); + cb(comp, name, il); + + comp = build_norm(comp, norm, nullptr, LLM_NORM_RMS, il); + cb(comp, name, il); + + ggml_tensor * comp_nope = ggml_view_3d(ctx0, comp, n_embd_head_nope, 1, n_blocks, + ggml_row_size(comp->type, n_embd_head), + ggml_row_size(comp->type, n_embd_head), + 0); + ggml_tensor * comp_pe = ggml_view_3d(ctx0, comp, n_embd_head_rope, 1, n_blocks, + ggml_row_size(comp->type, n_embd_head), + ggml_row_size(comp->type, n_embd_head), + ggml_row_size(comp->type, n_embd_head_nope)); + + comp_pe = ggml_rope_ext(ctx0, comp_pe, comp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig, + hparams.dsv4_compress_rope_base, freq_scale, ext_factor, 1.0f, beta_fast, beta_slow); + cb(comp_pe, name, il); + + comp = ggml_concat(ctx0, comp_nope, comp_pe, 0); + cb(comp, name, il); + + return comp; +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_csa_mask( + ggml_tensor * inp_pos, + int64_t n_kv, + int64_t n_tokens) const { + GGML_ASSERT(hparams.n_swa == DSV4_SWA_WINDOW); + GGML_ASSERT(n_tokens <= DSV4_SWA_WINDOW); + GGML_ASSERT(n_kv == n_tokens + n_tokens/DSV4_CSA_RATIO); + + const int64_t n_blocks = n_tokens/DSV4_CSA_RATIO; + + ggml_tensor * pos_f = ggml_cast(ctx0, inp_pos, GGML_TYPE_F32); + ggml_tensor * zero_1d = ggml_scale(ctx0, pos_f, 0.0f); + + ggml_tensor * raw_shape = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_tokens, n_tokens); + ggml_tensor * raw_mask = ggml_repeat(ctx0, zero_1d, raw_shape); + raw_mask = ggml_diag_mask_inf(ctx0, raw_mask, 0); + + ggml_tensor * block_end = ggml_arange(ctx0, float(DSV4_CSA_RATIO - 1), float(n_blocks*DSV4_CSA_RATIO), float(DSV4_CSA_RATIO)); + block_end = ggml_cast(ctx0, block_end, GGML_TYPE_I32); + + ggml_tensor * raw_mask_t = ggml_cont(ctx0, ggml_transpose(ctx0, raw_mask)); + ggml_tensor * comp_mask_t = ggml_get_rows(ctx0, raw_mask_t, block_end); + ggml_tensor * comp_mask = ggml_cont(ctx0, ggml_transpose(ctx0, comp_mask_t)); + + return ggml_concat(ctx0, raw_mask, comp_mask, 0); +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_lid_top_k( + const llama_model & model, + llm_graph_input_dsv4 * inp_dsv4, + ggml_tensor * qr, + ggml_tensor * cur, + ggml_tensor * inp_pos, + int il) const { + const auto & layer = model.layers[il]; + const auto & inp_lid = inp_dsv4->get_lid(); + const int64_t n_embd_indexer_head = hparams.indexer_head_size; + const int64_t n_embd_indexer_head_rope = hparams.n_rot(); + const int64_t n_embd_indexer_head_nope = n_embd_indexer_head - n_embd_indexer_head_rope; + const int64_t n_indexer_head = hparams.indexer_n_head; + const int64_t nt = cur->ne[1]; + + GGML_ASSERT(inp_lid.kq_mask); + GGML_ASSERT(inp_lid.k_rot); + GGML_ASSERT(n_embd_indexer_head >= n_embd_indexer_head_rope); + + ggml_tensor * indexer_q = build_lora_mm(layer.indexer_attn_q_b, qr); + indexer_q = ggml_reshape_3d(ctx0, indexer_q, n_embd_indexer_head, n_indexer_head, nt); + cb(indexer_q, "lid_q", il); + + ggml_tensor * indexer_q_nope = ggml_view_3d(ctx0, indexer_q, n_embd_indexer_head_nope, n_indexer_head, nt, + ggml_row_size(indexer_q->type, n_embd_indexer_head), + ggml_row_size(indexer_q->type, n_embd_indexer_head)*n_indexer_head, + 0); + ggml_tensor * indexer_q_pe = ggml_view_3d(ctx0, indexer_q, n_embd_indexer_head_rope, n_indexer_head, nt, + ggml_row_size(indexer_q->type, n_embd_indexer_head), + ggml_row_size(indexer_q->type, n_embd_indexer_head)*n_indexer_head, + ggml_row_size(indexer_q->type, n_embd_indexer_head_nope)); + + indexer_q_pe = ggml_rope_ext(ctx0, indexer_q_pe, inp_pos, nullptr, n_embd_indexer_head_rope, + rope_type, n_ctx_orig, hparams.dsv4_compress_rope_base, freq_scale, + ext_factor, 1.0f, beta_fast, beta_slow); + cb(indexer_q_pe, "lid_q_pe", il); + + indexer_q = ggml_concat(ctx0, indexer_q_nope, indexer_q_pe, 0); + indexer_q = ggml_mul_mat(ctx0, inp_lid.k_rot, indexer_q); + cb(indexer_q, "lid_q_rot", il); + + ggml_tensor * indexer_weights = build_lora_mm(layer.indexer_proj, cur); + indexer_weights = ggml_scale(ctx0, indexer_weights, 1.0f/sqrtf(float(n_embd_indexer_head*n_indexer_head))); + cb(indexer_weights, "lid_weights", il); + + ggml_tensor * indexer_k = inp_dsv4->mctx->get_lid()->get_k(ctx0, il); + const int64_t n_lid = inp_lid.kq_mask->ne[0]; + GGML_ASSERT(n_lid > 0); + GGML_ASSERT(n_lid <= indexer_k->ne[2]); + + indexer_k = ggml_view_4d(ctx0, indexer_k, + indexer_k->ne[0], indexer_k->ne[1], n_lid, indexer_k->ne[3], + indexer_k->nb[1], indexer_k->nb[2], indexer_k->nb[3], 0); + cb(indexer_k, "lid_k", il); + + const int64_t n_stream = indexer_k->ne[3]; + indexer_q = ggml_view_4d(ctx0, indexer_q, + indexer_q->ne[0], indexer_q->ne[1], indexer_q->ne[2]/n_stream, n_stream, + indexer_q->nb[1], indexer_q->nb[2], indexer_q->nb[3]/n_stream, 0); + indexer_weights = ggml_view_4d(ctx0, indexer_weights, + indexer_weights->ne[0], indexer_weights->ne[1]/n_stream, indexer_weights->ne[2], n_stream, + indexer_weights->nb[1], indexer_weights->nb[2]/n_stream, indexer_weights->nb[3]/n_stream, 0); + + indexer_q = ggml_permute(ctx0, indexer_q, 0, 2, 1, 3); + cb(indexer_q, "lid_q", il); + indexer_k = ggml_permute(ctx0, indexer_k, 0, 2, 1, 3); + cb(indexer_k, "lid_k", il); + + ggml_tensor * indexer_kq = ggml_mul_mat(ctx0, indexer_k, indexer_q); + cb(indexer_kq, "lid_kq", il); + + indexer_kq = ggml_cont(ctx0, ggml_permute(ctx0, indexer_kq, 2, 1, 0, 3)); + cb(indexer_kq, "lid_kq", il); + + ggml_tensor * indexer_score = ggml_relu(ctx0, indexer_kq); + indexer_score = ggml_mul(ctx0, indexer_score, indexer_weights); + indexer_score = ggml_sum_rows(ctx0, indexer_score); + indexer_score = ggml_cont(ctx0, ggml_permute(ctx0, indexer_score, 2, 1, 0, 3)); + cb(indexer_score, "lid_score", il); + + indexer_score = ggml_add(ctx0, indexer_score, inp_lid.kq_mask); + cb(indexer_score, "lid_score_masked", il); + + const uint32_t n_top_k = indexer_score->ne[0] < hparams.indexer_top_k ? indexer_score->ne[0] : hparams.indexer_top_k; + ggml_tensor * top_k = ggml_cont(ctx0, ggml_top_k(ctx0, indexer_score, n_top_k)); + cb(top_k, "lid_top_k", il); + + return top_k; +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_top_k_mask( + ggml_tensor * kq_mask, + ggml_tensor * top_k, + const char * name, + int il) const { + GGML_ASSERT(kq_mask); + GGML_ASSERT(top_k); + + ggml_tensor * kq_mask_all = ggml_fill(ctx0, kq_mask, -INFINITY); + kq_mask_all = ggml_view_4d(ctx0, kq_mask_all, 1, kq_mask_all->ne[0], kq_mask_all->ne[1], kq_mask_all->ne[3], + kq_mask_all->nb[0], kq_mask_all->nb[1], kq_mask_all->nb[2], 0); + + ggml_tensor * top_k_3d = ggml_view_4d(ctx0, top_k, top_k->ne[0], top_k->ne[1], top_k->ne[3], 1, + top_k->nb[1], top_k->nb[2], top_k->ne[3]*top_k->nb[3], 0); + + ggml_tensor * zeros = ggml_new_tensor_4d(ctx0, GGML_TYPE_F32, 1, top_k_3d->ne[0], top_k_3d->ne[1], top_k_3d->ne[2]); + zeros = ggml_fill(ctx0, zeros, 0.0f); + + ggml_tensor * kq_mask_top_k = ggml_set_rows(ctx0, kq_mask_all, zeros, top_k_3d); + kq_mask_top_k = ggml_view_4d(ctx0, kq_mask_top_k, + kq_mask_top_k->ne[1], kq_mask_top_k->ne[2], 1, kq_mask_top_k->ne[3], + kq_mask_top_k->nb[2], kq_mask_top_k->nb[3], kq_mask_top_k->nb[3], 0); + + kq_mask_top_k = ggml_add(ctx0, kq_mask_top_k, kq_mask); + cb(kq_mask_top_k, name, il); + + return kq_mask_top_k; +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_csa_lid_attention( + const llama_model & model, + llm_graph_input_dsv4 * inp_dsv4, + llm_graph_input_attn_kv_iswa * inp_attn, + ggml_tensor * q, + ggml_tensor * kv, + ggml_tensor * qr, + ggml_tensor * cur, + ggml_tensor * inp_pos, + ggml_tensor * sinks, + float kq_scale, + int il) const { + const auto & inp_csa = inp_dsv4->get_csa(); + GGML_ASSERT(inp_csa.kq_mask); + GGML_ASSERT(inp_attn->self_k_rot_swa == nullptr); + GGML_ASSERT(inp_attn->self_v_rot_swa == nullptr); + + ggml_tensor * top_k = build_lid_top_k(model, inp_dsv4, qr, cur, inp_pos, il); + + ggml_build_forward_expand(gf, q); + ggml_build_forward_expand(gf, kv); + + const llama_kv_cache_context * mctx_swa = inp_attn->mctx->get_swa(); + + ggml_build_forward_expand(gf, mctx_swa->cpy_k(ctx0, kv, inp_attn->get_k_idxs_swa(), il)); + ggml_build_forward_expand(gf, mctx_swa->cpy_v(ctx0, kv, inp_attn->get_v_idxs_swa(), il)); + + ggml_tensor * raw_k = mctx_swa->get_k(ctx0, il); + if (raw_k->type != GGML_TYPE_F32) { + raw_k = ggml_cast(ctx0, raw_k, GGML_TYPE_F32); + } + cb(raw_k, "csa_raw_k", il); + + ggml_tensor * csa_k = inp_dsv4->mctx->get_csa()->get_k(ctx0, il); + const int64_t n_csa = inp_csa.kq_mask->ne[0]; + GGML_ASSERT(n_csa > 0); + GGML_ASSERT(n_csa <= csa_k->ne[2]); + if (csa_k->type != GGML_TYPE_F32) { + csa_k = ggml_cast(ctx0, csa_k, GGML_TYPE_F32); + } + + csa_k = ggml_view_4d(ctx0, csa_k, + csa_k->ne[0], csa_k->ne[1], n_csa, csa_k->ne[3], + csa_k->nb[1], csa_k->nb[2], csa_k->nb[3], 0); + cb(csa_k, "csa_comp_k", il); + + ggml_tensor * k_all = ggml_concat(ctx0, raw_k, csa_k, 2); + cb(k_all, "csa_k_all", il); + + ggml_tensor * raw_mask = inp_attn->get_kq_mask_swa(); + ggml_tensor * csa_mask = build_top_k_mask(inp_csa.kq_mask, top_k, "csa_top_k_mask", il); + if (raw_mask->type != csa_mask->type) { + raw_mask = ggml_cast(ctx0, raw_mask, csa_mask->type); + } + + ggml_tensor * kq_mask = ggml_concat(ctx0, raw_mask, csa_mask, 0); + cb(kq_mask, "csa_lid_kq_mask", il); + + ggml_tensor * out = build_attn_mha(q, k_all, k_all, nullptr, kq_mask, sinks, nullptr, kq_scale, il); + cb(out, "attn_csa_lid", il); + + return out; +} + +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hca_attention( + llm_graph_input_dsv4 * inp_dsv4, + llm_graph_input_attn_kv_iswa * inp_attn, + ggml_tensor * q, + ggml_tensor * kv, + ggml_tensor * sinks, + float kq_scale, + int il) const { + const auto & inp_hca = inp_dsv4->get_hca(); + GGML_ASSERT(inp_hca.kq_mask); + GGML_ASSERT(inp_attn->self_k_rot_swa == nullptr); + GGML_ASSERT(inp_attn->self_v_rot_swa == nullptr); + + ggml_build_forward_expand(gf, q); + ggml_build_forward_expand(gf, kv); + + const llama_kv_cache_context * mctx_swa = inp_attn->mctx->get_swa(); + + ggml_build_forward_expand(gf, mctx_swa->cpy_k(ctx0, kv, inp_attn->get_k_idxs_swa(), il)); + ggml_build_forward_expand(gf, mctx_swa->cpy_v(ctx0, kv, inp_attn->get_v_idxs_swa(), il)); + + ggml_tensor * raw_k = mctx_swa->get_k(ctx0, il); + if (raw_k->type != GGML_TYPE_F32) { + raw_k = ggml_cast(ctx0, raw_k, GGML_TYPE_F32); + } + cb(raw_k, "hca_raw_k", il); + + ggml_tensor * hca_k = inp_dsv4->mctx->get_hca()->get_k(ctx0, il); + const int64_t n_hca = inp_hca.kq_mask->ne[0]; + GGML_ASSERT(n_hca > 0); + GGML_ASSERT(n_hca <= hca_k->ne[2]); + if (hca_k->type != GGML_TYPE_F32) { + hca_k = ggml_cast(ctx0, hca_k, GGML_TYPE_F32); + } + + hca_k = ggml_view_4d(ctx0, hca_k, + hca_k->ne[0], hca_k->ne[1], n_hca, hca_k->ne[3], + hca_k->nb[1], hca_k->nb[2], hca_k->nb[3], 0); + cb(hca_k, "hca_comp_k", il); + + ggml_tensor * k_all = ggml_concat(ctx0, raw_k, hca_k, 2); + cb(k_all, "hca_k_all", il); + + ggml_tensor * raw_mask = inp_attn->get_kq_mask_swa(); + ggml_tensor * hca_mask = inp_hca.kq_mask; + if (raw_mask->type != hca_mask->type) { + raw_mask = ggml_cast(ctx0, raw_mask, hca_mask->type); + } + + ggml_tensor * kq_mask = ggml_concat(ctx0, raw_mask, hca_mask, 0); + cb(kq_mask, "hca_kq_mask", il); + + ggml_tensor * out = build_attn_mha(q, k_all, k_all, nullptr, kq_mask, sinks, nullptr, kq_scale, il); + cb(out, "attn_hca", il); + + return out; +} + ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( const llama_model & model, - llm_graph_input_attn_no_cache * inp_attn, + llm_graph_input_dsv4 * inp_dsv4, ggml_tensor * cur, ggml_tensor * inp_pos, int il) const { const auto & layer = model.layers[il]; + llm_graph_input_attn_kv_iswa * inp_attn = inp_dsv4->get_raw(); const int64_t n_embd_head = hparams.n_embd_head_k(); const int64_t n_embd_head_rope = hparams.n_rot(); @@ -399,11 +941,279 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( kv = ggml_concat(ctx0, kv_nope, kv_pe, 0); cb(kv, "kv", il); - ggml_tensor * out = build_attn(inp_attn, - nullptr, nullptr, nullptr, - q, kv, kv, nullptr, layer.attn_sinks, nullptr, - 1.0f/sqrtf(float(n_embd_head)), il); - cb(out, "attn_raw", il); + const int64_t ratio = hparams.dsv4_compress_ratios[il]; + + ggml_tensor * hca_state_kv = nullptr; + ggml_tensor * hca_state_score = nullptr; + if (ratio == DSV4_HCA_RATIO && inp_dsv4->get_hca().state_idxs) { + hca_state_kv = build_lora_mm(layer.attn_comp_wkv, cur); + hca_state_kv = ggml_cont(ctx0, ggml_cast(ctx0, hca_state_kv, GGML_TYPE_F32)); + cb(hca_state_kv, "hca_state_kv", il); + + hca_state_score = build_lora_mm(layer.attn_comp_wgate, cur); + hca_state_score = ggml_cont(ctx0, ggml_cast(ctx0, hca_state_score, GGML_TYPE_F32)); + cb(hca_state_score, "hca_state_score", il); + + ggml_tensor * ape = layer.attn_comp_ape; + if (ape->type != GGML_TYPE_F32) { + ape = ggml_cast(ctx0, ape, GGML_TYPE_F32); + } + + ggml_tensor * ape_rows = ggml_get_rows(ctx0, ape, inp_dsv4->get_hca().state_pos); + hca_state_score = ggml_add(ctx0, hca_state_score, ape_rows); + cb(hca_state_score, "hca_state_score_ape", il); + + } + + ggml_tensor * kv_comp_csa = nullptr; + if (ratio == DSV4_CSA_RATIO && inp_dsv4->get_csa().write_idxs) { + kv_comp_csa = build_compressed_kv(cur, + inp_dsv4->get_csa().write_pos, + layer.attn_comp_wkv, + layer.attn_comp_wgate, + layer.attn_comp_ape, + layer.attn_comp_norm, + DSV4_CSA_RATIO, + n_embd_head, + true, + "csa_compress", + il); + + ggml_build_forward_expand(gf, inp_dsv4->mctx->get_csa()->cpy_k(ctx0, + kv_comp_csa, inp_dsv4->get_csa().write_idxs, il)); + + if (inp_dsv4->get_lid().write_idxs) { + ggml_tensor * kv_comp_lid = build_compressed_kv(cur, + inp_dsv4->get_lid().write_pos, + layer.indexer_comp_wkv, + layer.indexer_comp_wgate, + layer.indexer_comp_ape, + layer.indexer_comp_norm, + DSV4_CSA_RATIO, + hparams.indexer_head_size, + true, + "lid_compress", + il); + + if (inp_dsv4->get_lid().k_rot) { + kv_comp_lid = ggml_mul_mat(ctx0, inp_dsv4->get_lid().k_rot, kv_comp_lid); + cb(kv_comp_lid, "lid_compress_rot", il); + } + + ggml_build_forward_expand(gf, inp_dsv4->mctx->get_lid()->cpy_k(ctx0, + kv_comp_lid, inp_dsv4->get_lid().write_idxs, il)); + } + } else if (ratio == DSV4_HCA_RATIO && inp_dsv4->get_hca().write_idxs) { + ggml_tensor * kv_comp_hca = build_compressed_kv(cur, + inp_dsv4->get_hca().write_pos, + layer.attn_comp_wkv, + layer.attn_comp_wgate, + layer.attn_comp_ape, + layer.attn_comp_norm, + DSV4_HCA_RATIO, + n_embd_head, + false, + "hca_compress", + il); + + ggml_build_forward_expand(gf, inp_dsv4->mctx->get_hca()->cpy_k(ctx0, + kv_comp_hca, inp_dsv4->get_hca().write_idxs, il)); + } + + if (ratio == DSV4_CSA_RATIO && inp_dsv4->get_csa().state_idxs) { + ggml_tensor * csa_state_kv = build_lora_mm(layer.attn_comp_wkv, cur); + csa_state_kv = ggml_cont(ctx0, ggml_cast(ctx0, csa_state_kv, GGML_TYPE_F32)); + cb(csa_state_kv, "csa_state_kv", il); + + ggml_tensor * csa_state_score = build_lora_mm(layer.attn_comp_wgate, cur); + csa_state_score = ggml_cont(ctx0, ggml_cast(ctx0, csa_state_score, GGML_TYPE_F32)); + cb(csa_state_score, "csa_state_score", il); + + ggml_tensor * csa_ape = layer.attn_comp_ape; + if (csa_ape->type != GGML_TYPE_F32) { + csa_ape = ggml_cast(ctx0, csa_ape, GGML_TYPE_F32); + } + + ggml_tensor * csa_ape_rows = ggml_get_rows(ctx0, csa_ape, inp_dsv4->get_csa().state_pos); + csa_state_score = ggml_add(ctx0, csa_state_score, csa_ape_rows); + cb(csa_state_score, "csa_state_score_ape", il); + + ggml_tensor * csa_state_dep = nullptr; + if (inp_dsv4->get_csa().state_write_idxs) { + ggml_tensor * csa_source_kv = ggml_concat(ctx0, + inp_dsv4->mctx->get_csa_state()->get_kv(ctx0, il), csa_state_kv, 1); + ggml_tensor * csa_source_score = ggml_concat(ctx0, + inp_dsv4->mctx->get_csa_state()->get_score(ctx0, il), csa_state_score, 1); + + ggml_tensor * kv_comp_csa_state = build_overlap_compressed_kv_from_state( + csa_source_kv, + csa_source_score, + inp_dsv4->get_csa().state_read_idxs, + inp_dsv4->get_csa().state_write_pos, + layer.attn_comp_norm, + DSV4_CSA_RATIO, + n_embd_head, + "csa_state_compress", + il); + + ggml_build_forward_expand(gf, inp_dsv4->mctx->get_csa()->cpy_k(ctx0, + kv_comp_csa_state, inp_dsv4->get_csa().state_write_idxs, il)); + csa_state_dep = kv_comp_csa_state; + } + + csa_state_kv = dsv4_with_zero_dep(ctx0, csa_state_kv, csa_state_dep); + csa_state_score = dsv4_with_zero_dep(ctx0, csa_state_score, csa_state_dep); + + csa_state_kv = inp_dsv4->mctx->get_csa_state()->cpy_kv(ctx0, + csa_state_kv, inp_dsv4->get_csa().state_idxs, il); + csa_state_score = inp_dsv4->mctx->get_csa_state()->cpy_score(ctx0, + csa_state_score, inp_dsv4->get_csa().state_idxs, il); + + ggml_build_forward_expand(gf, csa_state_kv); + ggml_build_forward_expand(gf, csa_state_score); + + ggml_tensor * lid_state_kv = build_lora_mm(layer.indexer_comp_wkv, cur); + lid_state_kv = ggml_cont(ctx0, ggml_cast(ctx0, lid_state_kv, GGML_TYPE_F32)); + cb(lid_state_kv, "lid_state_kv", il); + + ggml_tensor * lid_state_score = build_lora_mm(layer.indexer_comp_wgate, cur); + lid_state_score = ggml_cont(ctx0, ggml_cast(ctx0, lid_state_score, GGML_TYPE_F32)); + cb(lid_state_score, "lid_state_score", il); + + ggml_tensor * lid_ape = layer.indexer_comp_ape; + if (lid_ape->type != GGML_TYPE_F32) { + lid_ape = ggml_cast(ctx0, lid_ape, GGML_TYPE_F32); + } + + ggml_tensor * lid_ape_rows = ggml_get_rows(ctx0, lid_ape, inp_dsv4->get_lid().state_pos); + lid_state_score = ggml_add(ctx0, lid_state_score, lid_ape_rows); + cb(lid_state_score, "lid_state_score_ape", il); + + ggml_tensor * lid_state_dep = nullptr; + if (inp_dsv4->get_lid().state_write_idxs) { + ggml_tensor * lid_source_kv = ggml_concat(ctx0, + inp_dsv4->mctx->get_lid_state()->get_kv(ctx0, il), lid_state_kv, 1); + ggml_tensor * lid_source_score = ggml_concat(ctx0, + inp_dsv4->mctx->get_lid_state()->get_score(ctx0, il), lid_state_score, 1); + + ggml_tensor * kv_comp_lid_state = build_overlap_compressed_kv_from_state( + lid_source_kv, + lid_source_score, + inp_dsv4->get_lid().state_read_idxs, + inp_dsv4->get_lid().state_write_pos, + layer.indexer_comp_norm, + DSV4_CSA_RATIO, + hparams.indexer_head_size, + "lid_state_compress", + il); + + if (inp_dsv4->get_lid().k_rot) { + kv_comp_lid_state = ggml_mul_mat(ctx0, inp_dsv4->get_lid().k_rot, kv_comp_lid_state); + cb(kv_comp_lid_state, "lid_state_compress_rot", il); + } + + ggml_build_forward_expand(gf, inp_dsv4->mctx->get_lid()->cpy_k(ctx0, + kv_comp_lid_state, inp_dsv4->get_lid().state_write_idxs, il)); + lid_state_dep = kv_comp_lid_state; + } + + lid_state_kv = dsv4_with_zero_dep(ctx0, lid_state_kv, lid_state_dep); + lid_state_score = dsv4_with_zero_dep(ctx0, lid_state_score, lid_state_dep); + + lid_state_kv = inp_dsv4->mctx->get_lid_state()->cpy_kv(ctx0, + lid_state_kv, inp_dsv4->get_lid().state_idxs, il); + lid_state_score = inp_dsv4->mctx->get_lid_state()->cpy_score(ctx0, + lid_state_score, inp_dsv4->get_lid().state_idxs, il); + + ggml_build_forward_expand(gf, lid_state_kv); + ggml_build_forward_expand(gf, lid_state_score); + } + + ggml_tensor * hca_state_dep = nullptr; + if (ratio == DSV4_HCA_RATIO && inp_dsv4->get_hca().state_write_idxs) { + GGML_ASSERT(hca_state_kv); + GGML_ASSERT(hca_state_score); + + ggml_tensor * hca_source_kv = ggml_concat(ctx0, + inp_dsv4->mctx->get_hca_state()->get_kv(ctx0, il), hca_state_kv, 1); + ggml_tensor * hca_source_score = ggml_concat(ctx0, + inp_dsv4->mctx->get_hca_state()->get_score(ctx0, il), hca_state_score, 1); + + ggml_tensor * kv_comp_hca = build_hca_compressed_kv_from_state( + hca_source_kv, + hca_source_score, + inp_dsv4->get_hca().state_read_idxs, + inp_dsv4->get_hca().state_write_pos, + layer.attn_comp_norm, + n_embd_head, + "hca_state_compress", + il); + + ggml_build_forward_expand(gf, inp_dsv4->mctx->get_hca()->cpy_k(ctx0, + kv_comp_hca, inp_dsv4->get_hca().state_write_idxs, il)); + hca_state_dep = kv_comp_hca; + } + + if (ratio == DSV4_HCA_RATIO && inp_dsv4->get_hca().state_idxs) { + GGML_ASSERT(hca_state_kv); + GGML_ASSERT(hca_state_score); + + hca_state_kv = dsv4_with_zero_dep(ctx0, hca_state_kv, hca_state_dep); + hca_state_score = dsv4_with_zero_dep(ctx0, hca_state_score, hca_state_dep); + + hca_state_kv = inp_dsv4->mctx->get_hca_state()->cpy_kv(ctx0, + hca_state_kv, inp_dsv4->get_hca().state_idxs, il); + hca_state_score = inp_dsv4->mctx->get_hca_state()->cpy_score(ctx0, + hca_state_score, inp_dsv4->get_hca().state_idxs, il); + + ggml_build_forward_expand(gf, hca_state_kv); + ggml_build_forward_expand(gf, hca_state_score); + } + + ggml_tensor * out = nullptr; + const bool use_csa = + ratio == DSV4_CSA_RATIO && + kv_comp_csa && + kv_comp_csa->ne[2] == nt/DSV4_CSA_RATIO && + !inp_dsv4->get_csa().state_write_idxs; + if (ratio == DSV4_CSA_RATIO && + inp_dsv4->get_csa().kq_mask && + inp_dsv4->get_lid().kq_mask && + inp_dsv4->get_lid().k_rot && + inp_attn->self_k_rot_swa == nullptr && + inp_attn->self_v_rot_swa == nullptr) { + out = build_csa_lid_attention(model, inp_dsv4, inp_attn, q, kv, qr, cur, inp_pos, layer.attn_sinks, + 1.0f/sqrtf(float(n_embd_head)), il); + } else if (use_csa) { + // Keep the raw SWA cache populated for the first generated token; the dense CSA output below is used now. + ggml_tensor * raw_swa = build_attn(inp_attn, + nullptr, nullptr, nullptr, + q, kv, kv, nullptr, layer.attn_sinks, nullptr, + 1.0f/sqrtf(float(n_embd_head)), il); + cb(raw_swa, "attn_raw_swa_cache", il); + + ggml_tensor * kv_all = ggml_concat(ctx0, kv, kv_comp_csa, 2); + cb(kv_all, "csa_kv_all", il); + + ggml_tensor * csa_mask = build_csa_mask(inp_pos, kv_all->ne[2], nt); + cb(csa_mask, "csa_mask", il); + + out = build_attn_mha(q, kv_all, kv_all, nullptr, csa_mask, layer.attn_sinks, nullptr, + 1.0f/sqrtf(float(n_embd_head)), il); + cb(out, "attn_csa", il); + } else if (ratio == DSV4_HCA_RATIO && + inp_dsv4->get_hca().kq_mask && + inp_attn->self_k_rot_swa == nullptr && + inp_attn->self_v_rot_swa == nullptr) { + out = build_hca_attention(inp_dsv4, inp_attn, q, kv, layer.attn_sinks, + 1.0f/sqrtf(float(n_embd_head)), il); + } else { + out = build_attn(inp_attn, + nullptr, nullptr, nullptr, + q, kv, kv, nullptr, layer.attn_sinks, nullptr, + 1.0f/sqrtf(float(n_embd_head)), il); + cb(out, "attn_raw", il); + } out = ggml_reshape_3d(ctx0, out, n_embd_head, n_head, nt); ggml_tensor * out_nope = ggml_view_3d(ctx0, out, n_embd_head_nope, n_head, nt, @@ -439,7 +1249,8 @@ llama_model_deepseek_v4_flash::graph::graph(const llama_model & model, const llm ggml_tensor * inp = build_inp_embd(model.tok_embd); ggml_tensor * inp_pos = build_inp_pos(); ggml_tensor * inp_out_ids = build_inp_out_ids(); - llm_graph_input_attn_no_cache * inp_attn = build_attn_inp_no_cache(); + llm_graph_input_dsv4 * inp_dsv4 = build_inp_dsv4(); + llm_graph_input_attn_kv_iswa * inp_attn = inp_dsv4->get_raw(); ggml_build_forward_expand(gf, inp_attn->self_kq_mask); if (inp_attn->self_kq_mask_swa) { ggml_build_forward_expand(gf, inp_attn->self_kq_mask_swa); @@ -465,7 +1276,7 @@ llama_model_deepseek_v4_flash::graph::graph(const llama_model & model, const llm cur = build_norm(cur, model.layers[il].attn_norm, nullptr, LLM_NORM_RMS, il); cb(cur, "attn_norm", il); - cur = build_attention(model, inp_attn, cur, inp_pos, il); + cur = build_attention(model, inp_dsv4, cur, inp_pos, il); inpL = build_hc_post(cur, residual, post, comb, il); cb(inpL, "hc_attn_post", il); diff --git a/src/models/models.h b/src/models/models.h index afe38c2b3a41..9154be60e2ac 100644 --- a/src/models/models.h +++ b/src/models/models.h @@ -1087,11 +1087,86 @@ struct llama_model_deepseek_v4_flash : public llama_model_base { ggml_tensor * build_attention( const llama_model & model, - llm_graph_input_attn_no_cache * inp_attn, + llm_graph_input_dsv4 * inp_dsv4, ggml_tensor * cur, ggml_tensor * inp_pos, int il) const; + ggml_tensor * build_compressed_kv( + ggml_tensor * cur, + ggml_tensor * comp_pos, + ggml_tensor * wkv, + ggml_tensor * wgate, + ggml_tensor * ape, + ggml_tensor * norm, + int64_t ratio, + int64_t n_embd_head, + bool overlap, + const char * name, + int il) const; + + ggml_tensor * build_hca_compressed_kv_from_state( + ggml_tensor * kv_state, + ggml_tensor * score_state, + ggml_tensor * state_read_idxs, + ggml_tensor * comp_pos, + ggml_tensor * norm, + int64_t n_embd_head, + const char * name, + int il) const; + + ggml_tensor * build_overlap_compressed_kv_from_state( + ggml_tensor * kv_state, + ggml_tensor * score_state, + ggml_tensor * state_read_idxs, + ggml_tensor * comp_pos, + ggml_tensor * norm, + int64_t ratio, + int64_t n_embd_head, + const char * name, + int il) const; + + ggml_tensor * build_csa_mask( + ggml_tensor * inp_pos, + int64_t n_kv, + int64_t n_tokens) const; + + ggml_tensor * build_lid_top_k( + const llama_model & model, + llm_graph_input_dsv4 * inp_dsv4, + ggml_tensor * qr, + ggml_tensor * cur, + ggml_tensor * inp_pos, + int il) const; + + ggml_tensor * build_top_k_mask( + ggml_tensor * kq_mask, + ggml_tensor * top_k, + const char * name, + int il) const; + + ggml_tensor * build_csa_lid_attention( + const llama_model & model, + llm_graph_input_dsv4 * inp_dsv4, + llm_graph_input_attn_kv_iswa * inp_attn, + ggml_tensor * q, + ggml_tensor * kv, + ggml_tensor * qr, + ggml_tensor * cur, + ggml_tensor * inp_pos, + ggml_tensor * sinks, + float kq_scale, + int il) const; + + ggml_tensor * build_hca_attention( + llm_graph_input_dsv4 * inp_dsv4, + llm_graph_input_attn_kv_iswa * inp_attn, + ggml_tensor * q, + ggml_tensor * kv, + ggml_tensor * sinks, + float kq_scale, + int il) const; + ggml_tensor * build_hc_weighted_sum( ggml_tensor * x, ggml_tensor * weights) const; From 2170238524af9d22d54b8a49b5c636b5d9efb627 Mon Sep 17 00:00:00 2001 From: Aman Gupta Date: Wed, 3 Jun 2026 13:10:25 +0200 Subject: [PATCH 06/13] add save-load state --- src/llama-kv-cache-dsv4.cpp | 319 +++++++++++++++++++++++++++++++++++- src/llama-kv-cache-dsv4.h | 5 + src/llama-kv-cache.cpp | 17 ++ src/llama-kv-cache.h | 3 + 4 files changed, 342 insertions(+), 2 deletions(-) diff --git a/src/llama-kv-cache-dsv4.cpp b/src/llama-kv-cache-dsv4.cpp index e0de2d91d728..753c3249402e 100644 --- a/src/llama-kv-cache-dsv4.cpp +++ b/src/llama-kv-cache-dsv4.cpp @@ -3,6 +3,7 @@ #include "ggml-backend.h" #include "llama-impl.h" #include "llama-batch.h" +#include "llama-io.h" #include "llama-model.h" #include @@ -17,10 +18,181 @@ static constexpr uint32_t DSV4_CSA_RATIO = 4; static constexpr uint32_t DSV4_HCA_RATIO = 128; +static constexpr uint32_t DSV4_STATE_MAGIC = 0x34565344; // DSV4 +static constexpr uint32_t DSV4_STATE_VERSION = 1; +static constexpr uint32_t DSV4_K_CACHE_STATE_VER = 1; +static constexpr uint32_t DSV4_COMP_STATE_VER = 1; + static uint32_t dsv4_comp_size(uint32_t kv_size, uint32_t ratio) { return std::max(1, (kv_size + ratio - 1)/ratio); } +static void dsv4_state_src_stream_range( + uint32_t n_stream, + llama_seq_id seq_id, + uint32_t & s0, + uint32_t & ns) { + if (seq_id >= 0 && n_stream > 1) { + if ((uint32_t) seq_id >= n_stream) { + throw std::runtime_error("DSV4 state sequence id out of stream range"); + } + + s0 = (uint32_t) seq_id; + ns = 1; + return; + } + + s0 = 0; + ns = seq_id >= 0 ? 1 : n_stream; +} + +static void dsv4_state_dst_stream_range( + uint32_t n_stream, + llama_seq_id seq_id, + uint32_t ns, + uint32_t & s0) { + if (seq_id >= 0) { + if (ns != 1) { + throw std::runtime_error("DSV4 sequence state stream count mismatch"); + } + if (n_stream > 1 && (uint32_t) seq_id >= n_stream) { + throw std::runtime_error("DSV4 state sequence id out of stream range"); + } + + s0 = n_stream > 1 ? (uint32_t) seq_id : 0; + return; + } + + if (ns != n_stream) { + throw std::runtime_error("DSV4 full state stream count mismatch"); + } + + s0 = 0; +} + +static void dsv4_state_write_tensor_streams( + llama_io_write_i & io, + ggml_tensor * tensor, + uint32_t n_rows, + uint32_t s0, + uint32_t ns) { + const int32_t type_i = (int32_t) tensor->type; + const uint64_t ne0 = tensor->ne[0]; + const uint64_t rows = n_rows; + const uint64_t row_size = ggml_row_size(tensor->type, tensor->ne[0]); + + io.write(&type_i, sizeof(type_i)); + io.write(&ne0, sizeof(ne0)); + io.write(&rows, sizeof(rows)); + io.write(&row_size, sizeof(row_size)); + + const size_t offset = (size_t) s0*n_rows*row_size; + const size_t size = (size_t) ns*n_rows*row_size; + + io.write_tensor(tensor, offset, size); +} + +static void dsv4_state_read_tensor_streams( + llama_io_read_i & io, + ggml_tensor * tensor, + uint32_t n_rows, + uint32_t s0, + uint32_t ns) { + int32_t type_i_ref; + uint64_t ne0_ref; + uint64_t rows_ref; + uint64_t row_size_ref; + + io.read(&type_i_ref, sizeof(type_i_ref)); + io.read(&ne0_ref, sizeof(ne0_ref)); + io.read(&rows_ref, sizeof(rows_ref)); + io.read(&row_size_ref, sizeof(row_size_ref)); + + const int32_t type_i = (int32_t) tensor->type; + const uint64_t ne0 = tensor->ne[0]; + const uint64_t rows = n_rows; + const uint64_t row_size = ggml_row_size(tensor->type, tensor->ne[0]); + + if (type_i != type_i_ref || ne0 != ne0_ref || rows != rows_ref || row_size != row_size_ref) { + throw std::runtime_error("DSV4 state tensor metadata mismatch"); + } + + const size_t offset = (size_t) s0*n_rows*row_size; + const size_t size = (size_t) ns*n_rows*row_size; + + io.read_tensor(tensor, offset, size); +} + +static void dsv4_state_write_k_cache( + llama_io_write_i & io, + const llama_kv_cache * kv, + llama_seq_id seq_id, + llama_state_seq_flags flags) { + GGML_UNUSED(flags); + + uint32_t s0; + uint32_t ns; + dsv4_state_src_stream_range(kv->get_n_stream(), seq_id, s0, ns); + + const uint32_t version = DSV4_K_CACHE_STATE_VER; + const uint32_t kv_size = kv->get_size(); + const auto layer_ids = kv->get_layer_ids(); + const uint32_t n_layer = layer_ids.size(); + + io.write(&version, sizeof(version)); + io.write(&kv_size, sizeof(kv_size)); + io.write(&ns, sizeof(ns)); + io.write(&n_layer, sizeof(n_layer)); + + for (uint32_t il : layer_ids) { + io.write(&il, sizeof(il)); + dsv4_state_write_tensor_streams(io, kv->get_k_storage(il), kv_size, s0, ns); + } +} + +static void dsv4_state_read_k_cache( + llama_io_read_i & io, + llama_kv_cache * kv, + llama_seq_id seq_id, + llama_state_seq_flags flags) { + GGML_UNUSED(flags); + + uint32_t version; + uint32_t kv_size_ref; + uint32_t ns; + uint32_t n_layer_ref; + + io.read(&version, sizeof(version)); + io.read(&kv_size_ref, sizeof(kv_size_ref)); + io.read(&ns, sizeof(ns)); + io.read(&n_layer_ref, sizeof(n_layer_ref)); + + if (version != DSV4_K_CACHE_STATE_VER) { + throw std::runtime_error("DSV4 K-cache state version mismatch"); + } + if (kv_size_ref != kv->get_size()) { + throw std::runtime_error("DSV4 K-cache state size mismatch"); + } + + uint32_t s0; + dsv4_state_dst_stream_range(kv->get_n_stream(), seq_id, ns, s0); + + const auto layer_ids = kv->get_layer_ids(); + if (n_layer_ref != layer_ids.size()) { + throw std::runtime_error("DSV4 K-cache layer count mismatch"); + } + + for (uint32_t il : layer_ids) { + uint32_t il_ref; + io.read(&il_ref, sizeof(il_ref)); + if (il_ref != il) { + throw std::runtime_error("DSV4 K-cache layer id mismatch"); + } + + dsv4_state_read_tensor_streams(io, kv->get_k_storage(il), kv->get_size(), s0, ns); + } +} + static std::string dsv4_plan_positions(const std::vector & values) { std::ostringstream ss; ss << "["; @@ -306,6 +478,73 @@ std::map llama_dsv4_comp_state::memory_break return ret; } +void llama_dsv4_comp_state::state_write(llama_io_write_i & io, llama_seq_id seq_id, llama_state_seq_flags flags) const { + GGML_UNUSED(flags); + + uint32_t s0; + uint32_t ns; + dsv4_state_src_stream_range(n_stream, seq_id, s0, ns); + + const uint32_t version = DSV4_COMP_STATE_VER; + const uint32_t n_layer = layers.size(); + + io.write(&version, sizeof(version)); + io.write(&ratio, sizeof(ratio)); + io.write(&state_size, sizeof(state_size)); + io.write(&n_embd_state, sizeof(n_embd_state)); + io.write(&ns, sizeof(ns)); + io.write(&n_layer, sizeof(n_layer)); + + for (const auto & layer : layers) { + io.write(&layer.il, sizeof(layer.il)); + + dsv4_state_write_tensor_streams(io, layer.kv, state_size, s0, ns); + dsv4_state_write_tensor_streams(io, layer.score, state_size, s0, ns); + } +} + +void llama_dsv4_comp_state::state_read(llama_io_read_i & io, llama_seq_id seq_id, llama_state_seq_flags flags) { + GGML_UNUSED(flags); + + uint32_t version; + uint32_t ratio_ref; + uint32_t state_size_ref; + uint32_t n_embd_state_ref; + uint32_t ns; + uint32_t n_layer_ref; + + io.read(&version, sizeof(version)); + io.read(&ratio_ref, sizeof(ratio_ref)); + io.read(&state_size_ref, sizeof(state_size_ref)); + io.read(&n_embd_state_ref, sizeof(n_embd_state_ref)); + io.read(&ns, sizeof(ns)); + io.read(&n_layer_ref, sizeof(n_layer_ref)); + + if (version != DSV4_COMP_STATE_VER) { + throw std::runtime_error("DSV4 compressor state version mismatch"); + } + if (ratio_ref != ratio || state_size_ref != state_size || n_embd_state_ref != n_embd_state) { + throw std::runtime_error("DSV4 compressor state metadata mismatch"); + } + if (n_layer_ref != layers.size()) { + throw std::runtime_error("DSV4 compressor state layer count mismatch"); + } + + uint32_t s0; + dsv4_state_dst_stream_range(n_stream, seq_id, ns, s0); + + for (const auto & layer : layers) { + uint32_t il_ref; + io.read(&il_ref, sizeof(il_ref)); + if (il_ref != layer.il) { + throw std::runtime_error("DSV4 compressor state layer id mismatch"); + } + + dsv4_state_read_tensor_streams(io, layer.kv, state_size, s0, ns); + dsv4_state_read_tensor_streams(io, layer.score, state_size, s0, ns); + } +} + ggml_tensor * llama_dsv4_comp_state::get_kv(ggml_context * ctx, int32_t il) const { const int32_t ids = map_layer_ids.at(il); @@ -540,6 +779,8 @@ bool llama_kv_cache_dsv4::get_can_shift() const { } void llama_kv_cache_dsv4::clear(bool data) { + restored_trim_pos.clear(); + kv_raw->clear(data); clear_compressed(data); } @@ -550,6 +791,20 @@ bool llama_kv_cache_dsv4::seq_rm(llama_seq_id seq_id, llama_pos p0, llama_pos p1 } if (p0 > 0) { + if (seq_id >= 0) { + auto it = restored_trim_pos.find(seq_id); + if (it != restored_trim_pos.end()) { + const llama_pos pos_max = it->second; + restored_trim_pos.erase(it); + + if (p0 >= pos_max) { + return kv_raw->seq_rm(seq_id, p0, p1); + } + + return false; + } + } + // DSV4 compressed cache rows are derived from running compressor state, // so arbitrary rollback is not reconstructible from the raw cache alone. // Allow the common prompt-cache cleanup no-op: remove [end, infinity). @@ -563,6 +818,12 @@ bool llama_kv_cache_dsv4::seq_rm(llama_seq_id seq_id, llama_pos p0, llama_pos p1 const bool res = kv_raw->seq_rm(seq_id, p0, p1); if (res) { + if (seq_id >= 0) { + restored_trim_pos.erase(seq_id); + } else { + restored_trim_pos.clear(); + } + clear_compressed(false); } @@ -570,27 +831,38 @@ bool llama_kv_cache_dsv4::seq_rm(llama_seq_id seq_id, llama_pos p0, llama_pos p1 } void llama_kv_cache_dsv4::seq_cp(llama_seq_id seq_id_src, llama_seq_id seq_id_dst, llama_pos p0, llama_pos p1) { + restored_trim_pos.clear(); + kv_raw->seq_cp(seq_id_src, seq_id_dst, p0, p1); clear_compressed(false); } void llama_kv_cache_dsv4::seq_keep(llama_seq_id seq_id) { + restored_trim_pos.clear(); + kv_raw->seq_keep(seq_id); clear_compressed(false); } void llama_kv_cache_dsv4::seq_add(llama_seq_id seq_id, llama_pos p0, llama_pos p1, llama_pos shift) { + restored_trim_pos.clear(); + kv_raw->seq_add(seq_id, p0, p1, shift); clear_compressed(false); } void llama_kv_cache_dsv4::seq_div(llama_seq_id seq_id, llama_pos p0, llama_pos p1, int d) { + restored_trim_pos.clear(); + kv_raw->seq_div(seq_id, p0, p1, d); clear_compressed(false); } llama_pos llama_kv_cache_dsv4::seq_pos_min(llama_seq_id seq_id) const { - return kv_raw->seq_pos_min(seq_id); + // The raw SWA cache may contain a wider window, but the compressed DSV4 + // state cannot be rolled back to the beginning of that window. Report the + // exact restored boundary so server-context prefers checkpoints. + return kv_raw->seq_pos_max(seq_id); } llama_pos llama_kv_cache_dsv4::seq_pos_max(llama_seq_id seq_id) const { @@ -621,12 +893,55 @@ std::map llama_kv_cache_dsv4::memory_breakdo } void llama_kv_cache_dsv4::state_write(llama_io_write_i & io, llama_seq_id seq_id, llama_state_seq_flags flags) const { + const uint32_t magic = DSV4_STATE_MAGIC; + const uint32_t version = DSV4_STATE_VERSION; + + io.write(&magic, sizeof(magic)); + io.write(&version, sizeof(version)); + kv_raw->state_write(io, seq_id, flags); + + dsv4_state_write_k_cache(io, kv_csa.get(), seq_id, flags); + dsv4_state_write_k_cache(io, kv_hca.get(), seq_id, flags); + dsv4_state_write_k_cache(io, kv_lid.get(), seq_id, flags); + + csa_state->state_write(io, seq_id, flags); + hca_state->state_write(io, seq_id, flags); + lid_state->state_write(io, seq_id, flags); } void llama_kv_cache_dsv4::state_read(llama_io_read_i & io, llama_seq_id seq_id, llama_state_seq_flags flags) { + uint32_t magic; + uint32_t version; + + io.read(&magic, sizeof(magic)); + io.read(&version, sizeof(version)); + + if (magic != DSV4_STATE_MAGIC) { + throw std::runtime_error("DSV4 state magic mismatch"); + } + if (version != DSV4_STATE_VERSION) { + throw std::runtime_error("DSV4 state version mismatch"); + } + + restored_trim_pos.clear(); + kv_raw->state_read(io, seq_id, flags); - clear_compressed(false); + + dsv4_state_read_k_cache(io, kv_csa.get(), seq_id, flags); + dsv4_state_read_k_cache(io, kv_hca.get(), seq_id, flags); + dsv4_state_read_k_cache(io, kv_lid.get(), seq_id, flags); + + csa_state->state_read(io, seq_id, flags); + hca_state->state_read(io, seq_id, flags); + lid_state->state_read(io, seq_id, flags); + + if (seq_id >= 0) { + const llama_pos pos_max = kv_raw->seq_pos_max(seq_id); + if (pos_max >= 0) { + restored_trim_pos[seq_id] = pos_max; + } + } } llama_kv_cache_iswa * llama_kv_cache_dsv4::get_raw() const { diff --git a/src/llama-kv-cache-dsv4.h b/src/llama-kv-cache-dsv4.h index 57bf4caf7019..14fd9793f1ff 100644 --- a/src/llama-kv-cache-dsv4.h +++ b/src/llama-kv-cache-dsv4.h @@ -29,6 +29,9 @@ class llama_dsv4_comp_state { std::map memory_breakdown() const; + void state_write(llama_io_write_i & io, llama_seq_id seq_id, llama_state_seq_flags flags) const; + void state_read (llama_io_read_i & io, llama_seq_id seq_id, llama_state_seq_flags flags); + ggml_tensor * get_kv (ggml_context * ctx, int32_t il) const; ggml_tensor * get_score(ggml_context * ctx, int32_t il) const; @@ -140,6 +143,8 @@ class llama_kv_cache_dsv4 : public llama_memory_i { std::unique_ptr hca_state; std::unique_ptr lid_state; + std::unordered_map restored_trim_pos; + void clear_compressed(bool data); }; diff --git a/src/llama-kv-cache.cpp b/src/llama-kv-cache.cpp index 03bc81935f9d..31b3d3d5844f 100644 --- a/src/llama-kv-cache.cpp +++ b/src/llama-kv-cache.cpp @@ -1133,6 +1133,23 @@ ggml_type llama_kv_cache::type_v() const { return layers[0].v->type; } +std::vector llama_kv_cache::get_layer_ids() const { + std::vector res; + res.reserve(layers.size()); + + for (const auto & layer : layers) { + res.push_back(layer.il); + } + + return res; +} + +ggml_tensor * llama_kv_cache::get_k_storage(int32_t il) const { + const int32_t ikv = map_layer_ids.at(il); + + return layers[ikv].k; +} + uint32_t llama_kv_cache::get_n_kv(const slot_info & sinfo) const { uint32_t result = 0; diff --git a/src/llama-kv-cache.h b/src/llama-kv-cache.h index 649269af6dd4..1bff7628fe2b 100644 --- a/src/llama-kv-cache.h +++ b/src/llama-kv-cache.h @@ -159,6 +159,9 @@ class llama_kv_cache : public llama_memory_i { ggml_type type_k() const; ggml_type type_v() const; + std::vector get_layer_ids() const; + ggml_tensor * get_k_storage(int32_t il) const; + // // graph_build API // From 5534b47fb0c82aa86d53d4f236db7c38df3d3eaa Mon Sep 17 00:00:00 2001 From: Aman Gupta Date: Thu, 4 Jun 2026 06:27:55 +0200 Subject: [PATCH 07/13] add sinkhorn eps - correction by @fairydreaming --- src/models/deepseek-v4.cpp | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) diff --git a/src/models/deepseek-v4.cpp b/src/models/deepseek-v4.cpp index d67b19fab29b..a9c40668f39f 100644 --- a/src/models/deepseek-v4.cpp +++ b/src/models/deepseek-v4.cpp @@ -226,17 +226,22 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hc_sinkhorn( // row softmax over dst, one column normalization, then repeated row/column normalization. comb = ggml_soft_max(ctx0, comb); + ggml_tensor * eps = ggml_new_tensor_1d(ctx0, GGML_TYPE_F32, 1); + eps = ggml_fill(ctx0, eps, hparams.dsv4_hc_eps); + + comb = ggml_add(ctx0, comb, eps); + auto norm_cols = [&]() { ggml_tensor * comb_src_dst = ggml_cont(ctx0, ggml_permute(ctx0, comb, 1, 0, 2, 3)); ggml_tensor * col_sum = ggml_sum_rows(ctx0, comb_src_dst); - col_sum = ggml_clamp(ctx0, col_sum, hparams.dsv4_hc_eps, INFINITY); + col_sum = ggml_add(ctx0, col_sum, eps); col_sum = ggml_permute(ctx0, col_sum, 1, 0, 2, 3); comb = ggml_div(ctx0, comb, col_sum); }; auto norm_rows = [&]() { ggml_tensor * row_sum = ggml_sum_rows(ctx0, comb); - row_sum = ggml_clamp(ctx0, row_sum, hparams.dsv4_hc_eps, INFINITY); + row_sum = ggml_add(ctx0, row_sum, eps); comb = ggml_div(ctx0, comb, row_sum); }; @@ -281,7 +286,9 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hc_pre( ggml_tensor * pre = dsv4_view_2d(ctx0, mixes, hc, nt, 0); pre = dsv4_hc_affine(ctx0, pre, scale_pre, base_pre); pre = ggml_sigmoid(ctx0, pre); - pre = ggml_clamp(ctx0, pre, hparams.dsv4_hc_eps, INFINITY); + ggml_tensor * eps = ggml_new_tensor_1d(ctx0, GGML_TYPE_F32, 1); + eps = ggml_fill(ctx0, eps, hparams.dsv4_hc_eps); + pre = ggml_add(ctx0, pre, eps); cb(pre, "hc_pre", il); *post = dsv4_view_2d(ctx0, mixes, hc, nt, hc); @@ -344,7 +351,9 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hc_head( ggml_tensor * pre = dsv4_hc_affine(ctx0, mixes, hc_scale, hc_base); pre = ggml_sigmoid(ctx0, pre); - pre = ggml_clamp(ctx0, pre, hparams.dsv4_hc_eps, INFINITY); + ggml_tensor * eps = ggml_new_tensor_1d(ctx0, GGML_TYPE_F32, 1); + eps = ggml_fill(ctx0, eps, hparams.dsv4_hc_eps); + pre = ggml_add(ctx0, pre, eps); cb(pre, "hc_head_pre", -1); return build_hc_weighted_sum(x, pre); From 4e36bd10ee17ac2b0d144c612dfbb2aa87a1f544 Mon Sep 17 00:00:00 2001 From: Aman Gupta Date: Thu, 4 Jun 2026 07:25:31 +0200 Subject: [PATCH 08/13] add rope fix --- src/models/deepseek-v4.cpp | 28 ++++++++++++++++++++-------- 1 file changed, 20 insertions(+), 8 deletions(-) diff --git a/src/models/deepseek-v4.cpp b/src/models/deepseek-v4.cpp index a9c40668f39f..2a93d6ec2023 100644 --- a/src/models/deepseek-v4.cpp +++ b/src/models/deepseek-v4.cpp @@ -11,6 +11,14 @@ static std::string dsv4_kv(const char * suffix) { return std::string("deepseek-v4-flash.") + suffix; } +static float dsv4_rope_attn_factor(float freq_scale, float ext_factor) { + if (ext_factor == 0.0f) { + return 1.0f; + } + + return 1.0f / (1.0f + 0.1f*logf(1.0f/freq_scale)); +} + void llama_model_deepseek_v4_flash::load_arch_hparams(llama_model_loader & ml) { ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); ml.get_key(LLM_KV_ATTENTION_Q_LORA_RANK, hparams.n_lora_q); @@ -54,7 +62,7 @@ void llama_model_deepseek_v4_flash::load_arch_hparams(llama_model_loader & ml) { } hparams.swa_type = LLAMA_SWA_TYPE_STANDARD; - std::fill(hparams.swa_layers.begin(), hparams.swa_layers.begin() + hparams.n_layer, 1); + hparams.set_swa_pattern(0); switch (hparams.n_layer) { case 43: type = LLM_TYPE_UNKNOWN; break; @@ -462,7 +470,8 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_compressed_kv( ggml_row_size(comp->type, n_embd_head_nope)); comp_pe = ggml_rope_ext(ctx0, comp_pe, comp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig, - hparams.dsv4_compress_rope_base, freq_scale, ext_factor, 1.0f, beta_fast, beta_slow); + hparams.dsv4_compress_rope_base, freq_scale, ext_factor, + dsv4_rope_attn_factor(freq_scale, ext_factor), beta_fast, beta_slow); cb(comp_pe, name, il); comp = ggml_concat(ctx0, comp_nope, comp_pe, 0); @@ -519,7 +528,8 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hca_compressed_kv_from ggml_row_size(comp->type, n_embd_head_nope)); comp_pe = ggml_rope_ext(ctx0, comp_pe, comp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig, - hparams.dsv4_compress_rope_base, freq_scale, ext_factor, 1.0f, beta_fast, beta_slow); + hparams.dsv4_compress_rope_base, freq_scale, ext_factor, + dsv4_rope_attn_factor(freq_scale, ext_factor), beta_fast, beta_slow); cb(comp_pe, name, il); comp = ggml_concat(ctx0, comp_nope, comp_pe, 0); @@ -600,7 +610,8 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_overlap_compressed_kv_ ggml_row_size(comp->type, n_embd_head_nope)); comp_pe = ggml_rope_ext(ctx0, comp_pe, comp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig, - hparams.dsv4_compress_rope_base, freq_scale, ext_factor, 1.0f, beta_fast, beta_slow); + hparams.dsv4_compress_rope_base, freq_scale, ext_factor, + dsv4_rope_attn_factor(freq_scale, ext_factor), beta_fast, beta_slow); cb(comp_pe, name, il); comp = ggml_concat(ctx0, comp_nope, comp_pe, 0); @@ -670,7 +681,7 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_lid_top_k( indexer_q_pe = ggml_rope_ext(ctx0, indexer_q_pe, inp_pos, nullptr, n_embd_indexer_head_rope, rope_type, n_ctx_orig, hparams.dsv4_compress_rope_base, freq_scale, - ext_factor, 1.0f, beta_fast, beta_slow); + ext_factor, dsv4_rope_attn_factor(freq_scale, ext_factor), beta_fast, beta_slow); cb(indexer_q_pe, "lid_q_pe", il); indexer_q = ggml_concat(ctx0, indexer_q_nope, indexer_q_pe, 0); @@ -902,6 +913,7 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( const float freq_base_l = use_compress_rope ? hparams.dsv4_compress_rope_base : freq_base; const float freq_scale_l = use_compress_rope ? freq_scale : 1.0f; const float ext_factor_l = use_compress_rope ? ext_factor : 0.0f; + const float attn_factor_l = dsv4_rope_attn_factor(freq_scale_l, ext_factor_l); const float beta_fast_l = use_compress_rope ? beta_fast : 0.0f; const float beta_slow_l = use_compress_rope ? beta_slow : 0.0f; const int32_t n_ctx_orig_l = use_compress_rope ? n_ctx_orig : 0; @@ -926,7 +938,7 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( ggml_row_size(q->type, n_embd_head)*n_head, ggml_row_size(q->type, n_embd_head_nope)); q_pe = ggml_rope_ext(ctx0, q_pe, inp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig_l, - freq_base_l, freq_scale_l, ext_factor_l, 1.0f, beta_fast_l, beta_slow_l); + freq_base_l, freq_scale_l, ext_factor_l, attn_factor_l, beta_fast_l, beta_slow_l); cb(q_pe, "q_pe", il); q = ggml_concat(ctx0, q_nope, q_pe, 0); cb(q, "q", il); @@ -945,7 +957,7 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( ggml_row_size(kv->type, n_embd_head), ggml_row_size(kv->type, n_embd_head_nope)); kv_pe = ggml_rope_ext(ctx0, kv_pe, inp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig_l, - freq_base_l, freq_scale_l, ext_factor_l, 1.0f, beta_fast_l, beta_slow_l); + freq_base_l, freq_scale_l, ext_factor_l, attn_factor_l, beta_fast_l, beta_slow_l); cb(kv_pe, "kv_pe", il); kv = ggml_concat(ctx0, kv_nope, kv_pe, 0); cb(kv, "kv", il); @@ -1234,7 +1246,7 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( ggml_row_size(out->type, n_embd_head)*n_head, ggml_row_size(out->type, n_embd_head_nope)); out_pe = ggml_rope_ext_back(ctx0, out_pe, inp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig_l, - freq_base_l, freq_scale_l, ext_factor_l, 1.0f, beta_fast_l, beta_slow_l); + freq_base_l, freq_scale_l, ext_factor_l, attn_factor_l, beta_fast_l, beta_slow_l); out = ggml_concat(ctx0, out_nope, out_pe, 0); cb(out, "attn_derope", il); From 20616c118b4fafe5c428d4c9fe33740b260eb602 Mon Sep 17 00:00:00 2001 From: Aman Gupta Date: Thu, 4 Jun 2026 07:44:02 +0200 Subject: [PATCH 09/13] cleanup dead code --- src/llama-graph.cpp | 22 +--- src/llama-graph.h | 5 - src/llama-kv-cache-dsv4.cpp | 62 ++++------- src/llama-kv-cache-dsv4.h | 14 --- src/models/deepseek-v4.cpp | 217 ------------------------------------ src/models/models.h | 18 --- 6 files changed, 25 insertions(+), 313 deletions(-) diff --git a/src/llama-graph.cpp b/src/llama-graph.cpp index 5b3dc2722b0f..37f017b70573 100644 --- a/src/llama-graph.cpp +++ b/src/llama-graph.cpp @@ -695,10 +695,6 @@ static void dsv4_set_comp_inputs( const char * name, bool debug, uint32_t n_tokens) { - dsv4_set_i64(inp.write_idxs, plan.write_idxs); - dsv4_set_i32(inp.write_pos, plan.write_pos); - dsv4_set_i32(inp.write_end, plan.write_end); - dsv4_set_i32(inp.pending_end, plan.pending_end); dsv4_set_i32(inp.state_idxs, plan.state_idxs); dsv4_set_i32(inp.state_pos, plan.state_pos); dsv4_set_i32(inp.state_read_idxs, plan.state_read_idxs); @@ -709,11 +705,9 @@ static void dsv4_set_comp_inputs( dsv4_set_kq_mask(inp.kq_mask, plan, n_tokens); if (debug || dsv4_compress_debug()) { - LLAMA_LOG_INFO("%s: %s ratio=%u, n_tokens=%u, write_end=%s, state_write_end=%s, pending_end=%s\n", + LLAMA_LOG_INFO("%s: %s ratio=%u, n_tokens=%u, state_write_end=%s\n", __func__, name, plan.ratio, n_tokens, - dsv4_plan_positions(plan.write_end).c_str(), - dsv4_plan_positions(plan.state_write_end).c_str(), - dsv4_plan_positions(plan.pending_end).c_str()); + dsv4_plan_positions(plan.state_write_end).c_str()); } } @@ -740,13 +734,7 @@ static bool dsv4_can_reuse_comp_input( const llm_graph_input_dsv4::comp_input & inp, const llama_kv_cache_dsv4_context::comp_plan & plan, uint32_t n_tokens) { - const int64_t n_write = plan.write_idxs.size(); - bool res = true; - res &= dsv4_can_reuse_tensor_1d(inp.write_idxs, n_write); - res &= dsv4_can_reuse_tensor_1d(inp.write_pos, n_write); - res &= dsv4_can_reuse_tensor_1d(inp.write_end, n_write); - res &= dsv4_can_reuse_tensor_1d(inp.pending_end, plan.pending_end.size()); res &= dsv4_can_reuse_tensor_1d(inp.state_idxs, plan.state_idxs.size()); res &= dsv4_can_reuse_tensor_1d(inp.state_pos, plan.state_pos.size()); res &= dsv4_can_reuse_tensor_1d(inp.state_read_idxs, plan.state_read_idxs.size()); @@ -780,12 +768,6 @@ static void dsv4_build_comp_inputs( llm_graph_input_dsv4::comp_input & inp, const llama_kv_cache_dsv4_context::comp_plan & plan, const char * name) { - const int64_t n_write = plan.write_idxs.size(); - - inp.write_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I64, n_write, std::string("dsv4_") + name + "_write_idxs"); - inp.write_pos = dsv4_build_input_1d(ctx, GGML_TYPE_I32, n_write, std::string("dsv4_") + name + "_write_pos"); - inp.write_end = dsv4_build_input_1d(ctx, GGML_TYPE_I32, n_write, std::string("dsv4_") + name + "_write_end"); - inp.pending_end = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.pending_end.size(), std::string("dsv4_") + name + "_pending_end"); inp.state_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_idxs.size(), std::string("dsv4_") + name + "_state_idxs"); inp.state_pos = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_pos.size(), std::string("dsv4_") + name + "_state_pos"); inp.state_read_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_read_idxs.size(), std::string("dsv4_") + name + "_state_read_idxs"); diff --git a/src/llama-graph.h b/src/llama-graph.h index 58906534ad7f..d4285811042e 100644 --- a/src/llama-graph.h +++ b/src/llama-graph.h @@ -463,11 +463,6 @@ class llm_graph_input_attn_kv_iswa : public llm_graph_input_i { class llm_graph_input_dsv4 : public llm_graph_input_i { public: struct comp_input { - ggml_tensor * write_idxs = nullptr; // I64 [n_write] - ggml_tensor * write_pos = nullptr; // I32 [n_write] - ggml_tensor * write_end = nullptr; // I32 [n_write] - ggml_tensor * pending_end = nullptr; // I32 [n_pending] - ggml_tensor * state_idxs = nullptr; // I32 [n_state] ggml_tensor * state_pos = nullptr; // I32 [n_state] ggml_tensor * state_read_idxs = nullptr; // I32 [ratio*n_state_write] diff --git a/src/llama-kv-cache-dsv4.cpp b/src/llama-kv-cache-dsv4.cpp index 753c3249402e..603934d41332 100644 --- a/src/llama-kv-cache-dsv4.cpp +++ b/src/llama-kv-cache-dsv4.cpp @@ -210,7 +210,6 @@ static llama_kv_cache_dsv4_context::comp_plan dsv4_build_comp_plan( const llama_ubatch & ubatch, uint32_t ratio, bool overlap, - bool stateful, uint32_t state_size, uint32_t kv_size, uint32_t n_stream) { @@ -256,12 +255,10 @@ static llama_kv_cache_dsv4_context::comp_plan dsv4_build_comp_plan( const llama_seq_id seq_id = ubatch.seq_id[i][0]; - if (stateful) { - const int64_t stream_off = n_stream > 1 ? (int64_t) seq_id*state_size : 0; + const int64_t stream_off = n_stream > 1 ? (int64_t) seq_id*state_size : 0; - plan.state_idxs.push_back((int32_t) (stream_off + pos%state_size)); - plan.state_pos .push_back((int32_t) (pos%ratio)); - } + plan.state_idxs.push_back((int32_t) (stream_off + pos%state_size)); + plan.state_pos .push_back((int32_t) (pos%ratio)); const int64_t n_visible = (int64_t) (pos + 1)/ratio; plan.n_visible[i] = (int32_t) n_visible; @@ -273,36 +270,26 @@ static llama_kv_cache_dsv4_context::comp_plan dsv4_build_comp_plan( const llama_pos source_start = pos + 1 - ratio; - if (stateful) { - const int64_t cache_off = n_stream > 1 ? (int64_t) seq_id*kv_size : 0; + const int64_t cache_off = n_stream > 1 ? (int64_t) seq_id*kv_size : 0; - plan.state_write_idxs.push_back(cache_off + pos/ratio); - plan.state_write_pos .push_back((int32_t) source_start); - plan.state_write_end .push_back((int32_t) pos); + plan.state_write_idxs.push_back(cache_off + pos/ratio); + plan.state_write_pos .push_back((int32_t) source_start); + plan.state_write_end .push_back((int32_t) pos); - if (overlap) { - const llama_pos prev_start = source_start - ratio; + if (overlap) { + const llama_pos prev_start = source_start - ratio; - for (uint32_t j = 0; j < ratio; ++j) { - plan.state_read_idxs.push_back(state_source_idx(seq_id, prev_start + j)); - } - for (uint32_t j = 0; j < ratio; ++j) { - plan.state_read_idxs.push_back(state_source_idx(seq_id, source_start + j)); - } - } else { - for (uint32_t j = 0; j < ratio; ++j) { - plan.state_read_idxs.push_back(state_source_idx(seq_id, source_start + j)); - } + for (uint32_t j = 0; j < ratio; ++j) { + plan.state_read_idxs.push_back(state_source_idx(seq_id, prev_start + j)); + } + for (uint32_t j = 0; j < ratio; ++j) { + plan.state_read_idxs.push_back(state_source_idx(seq_id, source_start + j)); + } + } else { + for (uint32_t j = 0; j < ratio; ++j) { + plan.state_read_idxs.push_back(state_source_idx(seq_id, source_start + j)); } - - continue; } - - const int64_t stream_off = n_stream > 1 ? (int64_t) seq_id*kv_size : 0; - - plan.write_idxs.push_back(stream_off + pos/ratio); - plan.write_pos .push_back((int32_t) (pos + 1 - ratio)); - plan.write_end .push_back((int32_t) pos); } static const bool debug = []() { @@ -311,11 +298,9 @@ static llama_kv_cache_dsv4_context::comp_plan dsv4_build_comp_plan( }(); if (debug) { - LLAMA_LOG_INFO("%s: ratio=%u, n_tokens=%u, write_end=%s, state_write_end=%s, pending_end=%s\n", + LLAMA_LOG_INFO("%s: ratio=%u, n_tokens=%u, state_write_end=%s\n", __func__, ratio, ubatch.n_tokens, - dsv4_plan_positions(plan.write_end).c_str(), - dsv4_plan_positions(plan.state_write_end).c_str(), - dsv4_plan_positions(plan.pending_end).c_str()); + dsv4_plan_positions(plan.state_write_end).c_str()); } return plan; @@ -325,7 +310,6 @@ static std::vector dsv4_build_comp_plans const std::vector & ubatches, uint32_t ratio, bool overlap, - bool stateful, uint32_t state_size, uint32_t kv_size, uint32_t n_stream) { @@ -333,7 +317,7 @@ static std::vector dsv4_build_comp_plans plans.reserve(ubatches.size()); for (const llama_ubatch & ubatch : ubatches) { - plans.push_back(dsv4_build_comp_plan(ubatch, ratio, overlap, stateful, state_size, kv_size, n_stream)); + plans.push_back(dsv4_build_comp_plan(ubatch, ratio, overlap, state_size, kv_size, n_stream)); } return plans; @@ -1023,9 +1007,9 @@ llama_kv_cache_dsv4_context::llama_kv_cache_dsv4_context( slot_info_vec_t sinfos_raw_swa, std::vector ubatches) : ubatches(std::move(ubatches)), - plans_csa(dsv4_build_comp_plans(this->ubatches, DSV4_CSA_RATIO, true, true, + plans_csa(dsv4_build_comp_plans(this->ubatches, DSV4_CSA_RATIO, true, kv->get_csa_state()->get_state_size(), kv->get_csa()->get_size(), kv->get_csa_state()->get_n_stream())), - plans_hca(dsv4_build_comp_plans(this->ubatches, DSV4_HCA_RATIO, false, true, + plans_hca(dsv4_build_comp_plans(this->ubatches, DSV4_HCA_RATIO, false, kv->get_hca_state()->get_state_size(), kv->get_hca()->get_size(), kv->get_hca_state()->get_n_stream())), plans_lid(plans_csa), ctx_raw(new llama_kv_cache_iswa_context(kv->get_raw(), std::move(sinfos_raw_base), std::move(sinfos_raw_swa), this->ubatches)), diff --git a/src/llama-kv-cache-dsv4.h b/src/llama-kv-cache-dsv4.h index 14fd9793f1ff..980852d1b1a6 100644 --- a/src/llama-kv-cache-dsv4.h +++ b/src/llama-kv-cache-dsv4.h @@ -155,20 +155,6 @@ class llama_kv_cache_dsv4_context : public llama_memory_context_i { struct comp_plan { uint32_t ratio = 0; - // Logical compressed row ids written by the current graph. - std::vector write_idxs; - - // Position used for compressor RoPE. For a completed block this is the - // first source-token position of that block. - std::vector write_pos; - - // Position at which the compressed row becomes visible to attention. - std::vector write_end; - - // Completed blocks that could not be planned. This should remain empty - // for the scratch-backed state path. - std::vector pending_end; - // Compressor-state row ids updated by the current graph. std::vector state_idxs; diff --git a/src/models/deepseek-v4.cpp b/src/models/deepseek-v4.cpp index 2a93d6ec2023..749eb678b7ce 100644 --- a/src/models/deepseek-v4.cpp +++ b/src/models/deepseek-v4.cpp @@ -367,119 +367,6 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hc_head( return build_hc_weighted_sum(x, pre); } -ggml_tensor * llama_model_deepseek_v4_flash::graph::build_compressed_kv( - ggml_tensor * cur, - ggml_tensor * comp_pos, - ggml_tensor * wkv, - ggml_tensor * wgate, - ggml_tensor * ape, - ggml_tensor * norm, - int64_t ratio, - int64_t n_embd_head, - bool overlap, - const char * name, - int il) const { - const int64_t n_embd_head_rope = hparams.n_rot(); - const int64_t n_embd_head_nope = n_embd_head - n_embd_head_rope; - const int64_t nt = cur->ne[1]; - const int64_t n_blocks = comp_pos ? comp_pos->ne[0] : 0; - const int64_t n_complete = n_blocks*ratio; - const int64_t coff = overlap ? 2 : 1; - - GGML_ASSERT(n_blocks > 0); - GGML_ASSERT(nt >= n_complete); - GGML_ASSERT(n_embd_head >= n_embd_head_rope); - - ggml_tensor * kv = build_lora_mm(wkv, cur); - kv = ggml_cont(ctx0, ggml_cast(ctx0, kv, GGML_TYPE_F32)); - cb(kv, name, il); - - ggml_tensor * score = build_lora_mm(wgate, cur); - score = ggml_cont(ctx0, ggml_cast(ctx0, score, GGML_TYPE_F32)); - cb(score, name, il); - - if (ape->type != GGML_TYPE_F32) { - ape = ggml_cast(ctx0, ape, GGML_TYPE_F32); - } - - kv = ggml_view_2d(ctx0, kv, coff*n_embd_head, n_complete, kv->nb[1], 0); - kv = ggml_reshape_3d(ctx0, kv, coff*n_embd_head, ratio, n_blocks); - - score = ggml_view_2d(ctx0, score, coff*n_embd_head, n_complete, score->nb[1], 0); - score = ggml_reshape_3d(ctx0, score, coff*n_embd_head, ratio, n_blocks); - score = ggml_add(ctx0, score, ape); - - ggml_tensor * values = kv; - ggml_tensor * scores = score; - if (overlap) { - ggml_tensor * kv_prev_src = ggml_view_3d(ctx0, kv, n_embd_head, ratio, n_blocks, - kv->nb[1], kv->nb[2], 0); - ggml_tensor * kv_cur = ggml_view_3d(ctx0, kv, n_embd_head, ratio, n_blocks, - kv->nb[1], kv->nb[2], ggml_row_size(kv->type, n_embd_head)); - - ggml_tensor * score_prev_src = ggml_view_3d(ctx0, score, n_embd_head, ratio, n_blocks, - score->nb[1], score->nb[2], 0); - ggml_tensor * score_cur = ggml_view_3d(ctx0, score, n_embd_head, ratio, n_blocks, - score->nb[1], score->nb[2], ggml_row_size(score->type, n_embd_head)); - - ggml_tensor * kv_prev_head = ggml_cont(ctx0, ggml_view_3d(ctx0, kv_prev_src, n_embd_head, ratio, 1, - kv_prev_src->nb[1], kv_prev_src->nb[2], 0)); - ggml_tensor * score_prev_head = ggml_cont(ctx0, ggml_view_3d(ctx0, score_prev_src, n_embd_head, ratio, 1, - score_prev_src->nb[1], score_prev_src->nb[2], 0)); - - ggml_tensor * kv_prev_zero = ggml_scale(ctx0, kv_prev_head, 0.0f); - ggml_tensor * score_prev_zero = ggml_scale(ctx0, score_prev_head, 0.0f); - ggml_tensor * score_prev_inf = ggml_scale_bias(ctx0, score_prev_zero, 0.0f, -INFINITY); - - ggml_tensor * kv_prev = kv_prev_zero; - ggml_tensor * score_prev = score_prev_inf; - if (n_blocks > 1) { - ggml_tensor * kv_prev_tail = ggml_view_3d(ctx0, kv_prev_src, n_embd_head, ratio, n_blocks - 1, - kv_prev_src->nb[1], kv_prev_src->nb[2], 0); - ggml_tensor * score_prev_tail = ggml_view_3d(ctx0, score_prev_src, n_embd_head, ratio, n_blocks - 1, - score_prev_src->nb[1], score_prev_src->nb[2], 0); - - kv_prev = ggml_concat(ctx0, kv_prev_zero, kv_prev_tail, 2); - score_prev = ggml_concat(ctx0, score_prev_inf, score_prev_tail, 2); - } - - values = ggml_concat(ctx0, kv_prev, kv_cur, 1); - scores = ggml_concat(ctx0, score_prev, score_cur, 1); - } - - values = ggml_cont(ctx0, ggml_permute(ctx0, values, 1, 0, 2, 3)); - - scores = ggml_cont(ctx0, ggml_permute(ctx0, scores, 1, 0, 2, 3)); - - ggml_tensor * weights = ggml_soft_max(ctx0, scores); - ggml_tensor * comp = ggml_mul(ctx0, values, weights); - comp = ggml_sum_rows(ctx0, comp); - comp = ggml_cont(ctx0, ggml_permute(ctx0, comp, 1, 0, 2, 3)); - cb(comp, name, il); - - comp = build_norm(comp, norm, nullptr, LLM_NORM_RMS, il); - cb(comp, name, il); - - ggml_tensor * comp_nope = ggml_view_3d(ctx0, comp, n_embd_head_nope, 1, n_blocks, - ggml_row_size(comp->type, n_embd_head), - ggml_row_size(comp->type, n_embd_head), - 0); - ggml_tensor * comp_pe = ggml_view_3d(ctx0, comp, n_embd_head_rope, 1, n_blocks, - ggml_row_size(comp->type, n_embd_head), - ggml_row_size(comp->type, n_embd_head), - ggml_row_size(comp->type, n_embd_head_nope)); - - comp_pe = ggml_rope_ext(ctx0, comp_pe, comp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig, - hparams.dsv4_compress_rope_base, freq_scale, ext_factor, - dsv4_rope_attn_factor(freq_scale, ext_factor), beta_fast, beta_slow); - cb(comp_pe, name, il); - - comp = ggml_concat(ctx0, comp_nope, comp_pe, 0); - cb(comp, name, il); - - return comp; -} - ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hca_compressed_kv_from_state( ggml_tensor * kv_state, ggml_tensor * score_state, @@ -620,33 +507,6 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_overlap_compressed_kv_ return comp; } -ggml_tensor * llama_model_deepseek_v4_flash::graph::build_csa_mask( - ggml_tensor * inp_pos, - int64_t n_kv, - int64_t n_tokens) const { - GGML_ASSERT(hparams.n_swa == DSV4_SWA_WINDOW); - GGML_ASSERT(n_tokens <= DSV4_SWA_WINDOW); - GGML_ASSERT(n_kv == n_tokens + n_tokens/DSV4_CSA_RATIO); - - const int64_t n_blocks = n_tokens/DSV4_CSA_RATIO; - - ggml_tensor * pos_f = ggml_cast(ctx0, inp_pos, GGML_TYPE_F32); - ggml_tensor * zero_1d = ggml_scale(ctx0, pos_f, 0.0f); - - ggml_tensor * raw_shape = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_tokens, n_tokens); - ggml_tensor * raw_mask = ggml_repeat(ctx0, zero_1d, raw_shape); - raw_mask = ggml_diag_mask_inf(ctx0, raw_mask, 0); - - ggml_tensor * block_end = ggml_arange(ctx0, float(DSV4_CSA_RATIO - 1), float(n_blocks*DSV4_CSA_RATIO), float(DSV4_CSA_RATIO)); - block_end = ggml_cast(ctx0, block_end, GGML_TYPE_I32); - - ggml_tensor * raw_mask_t = ggml_cont(ctx0, ggml_transpose(ctx0, raw_mask)); - ggml_tensor * comp_mask_t = ggml_get_rows(ctx0, raw_mask_t, block_end); - ggml_tensor * comp_mask = ggml_cont(ctx0, ggml_transpose(ctx0, comp_mask_t)); - - return ggml_concat(ctx0, raw_mask, comp_mask, 0); -} - ggml_tensor * llama_model_deepseek_v4_flash::graph::build_lid_top_k( const llama_model & model, llm_graph_input_dsv4 * inp_dsv4, @@ -986,61 +846,6 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( } - ggml_tensor * kv_comp_csa = nullptr; - if (ratio == DSV4_CSA_RATIO && inp_dsv4->get_csa().write_idxs) { - kv_comp_csa = build_compressed_kv(cur, - inp_dsv4->get_csa().write_pos, - layer.attn_comp_wkv, - layer.attn_comp_wgate, - layer.attn_comp_ape, - layer.attn_comp_norm, - DSV4_CSA_RATIO, - n_embd_head, - true, - "csa_compress", - il); - - ggml_build_forward_expand(gf, inp_dsv4->mctx->get_csa()->cpy_k(ctx0, - kv_comp_csa, inp_dsv4->get_csa().write_idxs, il)); - - if (inp_dsv4->get_lid().write_idxs) { - ggml_tensor * kv_comp_lid = build_compressed_kv(cur, - inp_dsv4->get_lid().write_pos, - layer.indexer_comp_wkv, - layer.indexer_comp_wgate, - layer.indexer_comp_ape, - layer.indexer_comp_norm, - DSV4_CSA_RATIO, - hparams.indexer_head_size, - true, - "lid_compress", - il); - - if (inp_dsv4->get_lid().k_rot) { - kv_comp_lid = ggml_mul_mat(ctx0, inp_dsv4->get_lid().k_rot, kv_comp_lid); - cb(kv_comp_lid, "lid_compress_rot", il); - } - - ggml_build_forward_expand(gf, inp_dsv4->mctx->get_lid()->cpy_k(ctx0, - kv_comp_lid, inp_dsv4->get_lid().write_idxs, il)); - } - } else if (ratio == DSV4_HCA_RATIO && inp_dsv4->get_hca().write_idxs) { - ggml_tensor * kv_comp_hca = build_compressed_kv(cur, - inp_dsv4->get_hca().write_pos, - layer.attn_comp_wkv, - layer.attn_comp_wgate, - layer.attn_comp_ape, - layer.attn_comp_norm, - DSV4_HCA_RATIO, - n_embd_head, - false, - "hca_compress", - il); - - ggml_build_forward_expand(gf, inp_dsv4->mctx->get_hca()->cpy_k(ctx0, - kv_comp_hca, inp_dsv4->get_hca().write_idxs, il)); - } - if (ratio == DSV4_CSA_RATIO && inp_dsv4->get_csa().state_idxs) { ggml_tensor * csa_state_kv = build_lora_mm(layer.attn_comp_wkv, cur); csa_state_kv = ggml_cont(ctx0, ggml_cast(ctx0, csa_state_kv, GGML_TYPE_F32)); @@ -1192,11 +997,6 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( } ggml_tensor * out = nullptr; - const bool use_csa = - ratio == DSV4_CSA_RATIO && - kv_comp_csa && - kv_comp_csa->ne[2] == nt/DSV4_CSA_RATIO && - !inp_dsv4->get_csa().state_write_idxs; if (ratio == DSV4_CSA_RATIO && inp_dsv4->get_csa().kq_mask && inp_dsv4->get_lid().kq_mask && @@ -1205,23 +1005,6 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( inp_attn->self_v_rot_swa == nullptr) { out = build_csa_lid_attention(model, inp_dsv4, inp_attn, q, kv, qr, cur, inp_pos, layer.attn_sinks, 1.0f/sqrtf(float(n_embd_head)), il); - } else if (use_csa) { - // Keep the raw SWA cache populated for the first generated token; the dense CSA output below is used now. - ggml_tensor * raw_swa = build_attn(inp_attn, - nullptr, nullptr, nullptr, - q, kv, kv, nullptr, layer.attn_sinks, nullptr, - 1.0f/sqrtf(float(n_embd_head)), il); - cb(raw_swa, "attn_raw_swa_cache", il); - - ggml_tensor * kv_all = ggml_concat(ctx0, kv, kv_comp_csa, 2); - cb(kv_all, "csa_kv_all", il); - - ggml_tensor * csa_mask = build_csa_mask(inp_pos, kv_all->ne[2], nt); - cb(csa_mask, "csa_mask", il); - - out = build_attn_mha(q, kv_all, kv_all, nullptr, csa_mask, layer.attn_sinks, nullptr, - 1.0f/sqrtf(float(n_embd_head)), il); - cb(out, "attn_csa", il); } else if (ratio == DSV4_HCA_RATIO && inp_dsv4->get_hca().kq_mask && inp_attn->self_k_rot_swa == nullptr && diff --git a/src/models/models.h b/src/models/models.h index 9154be60e2ac..46cdb1dcbdbc 100644 --- a/src/models/models.h +++ b/src/models/models.h @@ -1092,19 +1092,6 @@ struct llama_model_deepseek_v4_flash : public llama_model_base { ggml_tensor * inp_pos, int il) const; - ggml_tensor * build_compressed_kv( - ggml_tensor * cur, - ggml_tensor * comp_pos, - ggml_tensor * wkv, - ggml_tensor * wgate, - ggml_tensor * ape, - ggml_tensor * norm, - int64_t ratio, - int64_t n_embd_head, - bool overlap, - const char * name, - int il) const; - ggml_tensor * build_hca_compressed_kv_from_state( ggml_tensor * kv_state, ggml_tensor * score_state, @@ -1126,11 +1113,6 @@ struct llama_model_deepseek_v4_flash : public llama_model_base { const char * name, int il) const; - ggml_tensor * build_csa_mask( - ggml_tensor * inp_pos, - int64_t n_kv, - int64_t n_tokens) const; - ggml_tensor * build_lid_top_k( const llama_model & model, llm_graph_input_dsv4 * inp_dsv4, From f9c973451670da03484ae5cfa295d9dfc386ba83 Mon Sep 17 00:00:00 2001 From: Aman Gupta Date: Thu, 4 Jun 2026 15:52:41 +0200 Subject: [PATCH 10/13] fix bugs --- ggml/src/ggml-backend.cpp | 2 +- src/llama-graph.cpp | 9 ++++- src/llama-graph.h | 2 + src/llama-kv-cache-dsv4.cpp | 77 ++++++++++++++++++++++++++++++++----- src/llama-kv-cache-dsv4.h | 5 +++ src/models/deepseek-v4.cpp | 21 +++++++--- 6 files changed, 98 insertions(+), 18 deletions(-) diff --git a/ggml/src/ggml-backend.cpp b/ggml/src/ggml-backend.cpp index 87615921c09b..1bd5155d9ebf 100644 --- a/ggml/src/ggml-backend.cpp +++ b/ggml/src/ggml-backend.cpp @@ -754,7 +754,7 @@ static bool ggml_is_view_op(enum ggml_op op) { #endif #ifndef GGML_SCHED_MAX_SPLIT_INPUTS -#define GGML_SCHED_MAX_SPLIT_INPUTS 30 +#define GGML_SCHED_MAX_SPLIT_INPUTS 128 #endif #ifndef GGML_SCHED_MAX_COPIES diff --git a/src/llama-graph.cpp b/src/llama-graph.cpp index 37f017b70573..35cd70e5fdad 100644 --- a/src/llama-graph.cpp +++ b/src/llama-graph.cpp @@ -697,6 +697,8 @@ static void dsv4_set_comp_inputs( uint32_t n_tokens) { dsv4_set_i32(inp.state_idxs, plan.state_idxs); dsv4_set_i32(inp.state_pos, plan.state_pos); + dsv4_set_i32(inp.state_persist_src_idxs, plan.state_persist_src_idxs); + dsv4_set_i32(inp.state_persist_dst_idxs, plan.state_persist_dst_idxs); dsv4_set_i32(inp.state_read_idxs, plan.state_read_idxs); dsv4_set_i64(inp.state_write_idxs, plan.state_write_idxs); dsv4_set_i32(inp.state_write_pos, plan.state_write_pos); @@ -705,8 +707,9 @@ static void dsv4_set_comp_inputs( dsv4_set_kq_mask(inp.kq_mask, plan, n_tokens); if (debug || dsv4_compress_debug()) { - LLAMA_LOG_INFO("%s: %s ratio=%u, n_tokens=%u, state_write_end=%s\n", + LLAMA_LOG_INFO("%s: %s ratio=%u, n_tokens=%u, state_persist_dst=%s, state_write_end=%s\n", __func__, name, plan.ratio, n_tokens, + dsv4_plan_positions(plan.state_persist_dst_idxs).c_str(), dsv4_plan_positions(plan.state_write_end).c_str()); } } @@ -737,6 +740,8 @@ static bool dsv4_can_reuse_comp_input( bool res = true; res &= dsv4_can_reuse_tensor_1d(inp.state_idxs, plan.state_idxs.size()); res &= dsv4_can_reuse_tensor_1d(inp.state_pos, plan.state_pos.size()); + res &= dsv4_can_reuse_tensor_1d(inp.state_persist_src_idxs, plan.state_persist_src_idxs.size()); + res &= dsv4_can_reuse_tensor_1d(inp.state_persist_dst_idxs, plan.state_persist_dst_idxs.size()); res &= dsv4_can_reuse_tensor_1d(inp.state_read_idxs, plan.state_read_idxs.size()); res &= dsv4_can_reuse_tensor_1d(inp.state_write_idxs, plan.state_write_idxs.size()); res &= dsv4_can_reuse_tensor_1d(inp.state_write_pos, plan.state_write_pos.size()); @@ -770,6 +775,8 @@ static void dsv4_build_comp_inputs( const char * name) { inp.state_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_idxs.size(), std::string("dsv4_") + name + "_state_idxs"); inp.state_pos = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_pos.size(), std::string("dsv4_") + name + "_state_pos"); + inp.state_persist_src_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_persist_src_idxs.size(), std::string("dsv4_") + name + "_state_persist_src_idxs"); + inp.state_persist_dst_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_persist_dst_idxs.size(), std::string("dsv4_") + name + "_state_persist_dst_idxs"); inp.state_read_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_read_idxs.size(), std::string("dsv4_") + name + "_state_read_idxs"); inp.state_write_idxs = dsv4_build_input_1d(ctx, GGML_TYPE_I64, plan.state_write_idxs.size(), std::string("dsv4_") + name + "_state_write_idxs"); inp.state_write_pos = dsv4_build_input_1d(ctx, GGML_TYPE_I32, plan.state_write_pos.size(), std::string("dsv4_") + name + "_state_write_pos"); diff --git a/src/llama-graph.h b/src/llama-graph.h index d4285811042e..d7b398c44120 100644 --- a/src/llama-graph.h +++ b/src/llama-graph.h @@ -465,6 +465,8 @@ class llm_graph_input_dsv4 : public llm_graph_input_i { struct comp_input { ggml_tensor * state_idxs = nullptr; // I32 [n_state] ggml_tensor * state_pos = nullptr; // I32 [n_state] + ggml_tensor * state_persist_src_idxs = nullptr; // I32 [n_state_persist] + ggml_tensor * state_persist_dst_idxs = nullptr; // I32 [n_state_persist] ggml_tensor * state_read_idxs = nullptr; // I32 [ratio*n_state_write] ggml_tensor * state_write_idxs = nullptr; // I64 [n_state_write] ggml_tensor * state_write_pos = nullptr; // I32 [n_state_write] diff --git a/src/llama-kv-cache-dsv4.cpp b/src/llama-kv-cache-dsv4.cpp index 603934d41332..a786c732ed41 100644 --- a/src/llama-kv-cache-dsv4.cpp +++ b/src/llama-kv-cache-dsv4.cpp @@ -219,6 +219,24 @@ static llama_kv_cache_dsv4_context::comp_plan dsv4_build_comp_plan( const int64_t state_rows = (int64_t) state_size*n_stream; + struct persist_row { + int32_t dst; + int32_t src; + llama_pos pos; + }; + + std::vector persist_rows; + + // For the overlap compressor, build_overlap_compressed_kv_from_state() consumes + // state_read_idxs as two contiguous halves: the first ratio*n_blocks entries are + // the "previous-window" gather indices for every block, followed by the + // "current-window" indices for every block. Collect them separately here and + // append cur after prev once the loop has visited all completed blocks, instead + // of interleaving [prev, cur] per block (which corrupted every block but the + // last in multi-block ubatches / long-context prefill). + std::vector overlap_prev_reads; + std::vector overlap_cur_reads; + const auto current_token_idx = [&](llama_seq_id seq_id, llama_pos pos) -> int64_t { for (uint32_t i = 0; i < ubatch.n_tokens; ++i) { if (ubatch.pos[i] == pos && ubatch.seq_id[i][0] == seq_id) { @@ -257,9 +275,22 @@ static llama_kv_cache_dsv4_context::comp_plan dsv4_build_comp_plan( const int64_t stream_off = n_stream > 1 ? (int64_t) seq_id*state_size : 0; - plan.state_idxs.push_back((int32_t) (stream_off + pos%state_size)); + const int32_t state_idx = (int32_t) (stream_off + pos%state_size); + + plan.state_idxs.push_back(state_idx); plan.state_pos .push_back((int32_t) (pos%ratio)); + const auto it = std::find_if(persist_rows.begin(), persist_rows.end(), + [state_idx](const persist_row & row) { + return row.dst == state_idx; + }); + if (it == persist_rows.end()) { + persist_rows.push_back({ state_idx, (int32_t) i, pos }); + } else if (pos > it->pos) { + it->src = (int32_t) i; + it->pos = pos; + } + const int64_t n_visible = (int64_t) (pos + 1)/ratio; plan.n_visible[i] = (int32_t) n_visible; plan.n_kv = std::max(plan.n_kv, n_visible); @@ -280,10 +311,10 @@ static llama_kv_cache_dsv4_context::comp_plan dsv4_build_comp_plan( const llama_pos prev_start = source_start - ratio; for (uint32_t j = 0; j < ratio; ++j) { - plan.state_read_idxs.push_back(state_source_idx(seq_id, prev_start + j)); + overlap_prev_reads.push_back(state_source_idx(seq_id, prev_start + j)); } for (uint32_t j = 0; j < ratio; ++j) { - plan.state_read_idxs.push_back(state_source_idx(seq_id, source_start + j)); + overlap_cur_reads.push_back(state_source_idx(seq_id, source_start + j)); } } else { for (uint32_t j = 0; j < ratio; ++j) { @@ -292,14 +323,34 @@ static llama_kv_cache_dsv4_context::comp_plan dsv4_build_comp_plan( } } + if (overlap) { + // [ all blocks' prev-window indices | all blocks' cur-window indices ] + plan.state_read_idxs.reserve(overlap_prev_reads.size() + overlap_cur_reads.size()); + plan.state_read_idxs.insert(plan.state_read_idxs.end(), + overlap_prev_reads.begin(), overlap_prev_reads.end()); + plan.state_read_idxs.insert(plan.state_read_idxs.end(), + overlap_cur_reads.begin(), overlap_cur_reads.end()); + } + + std::sort(persist_rows.begin(), persist_rows.end(), + [](const persist_row & a, const persist_row & b) { + return a.dst < b.dst; + }); + + for (const persist_row & row : persist_rows) { + plan.state_persist_src_idxs.push_back(row.src); + plan.state_persist_dst_idxs.push_back(row.dst); + } + static const bool debug = []() { const char * env = getenv("LLAMA_DSV4_COMPRESS_DEBUG"); return env && atoi(env) > 0; }(); if (debug) { - LLAMA_LOG_INFO("%s: ratio=%u, n_tokens=%u, state_write_end=%s\n", + LLAMA_LOG_INFO("%s: ratio=%u, n_tokens=%u, state_persist_dst=%s, state_write_end=%s\n", __func__, ratio, ubatch.n_tokens, + dsv4_plan_positions(plan.state_persist_dst_idxs).c_str(), dsv4_plan_positions(plan.state_write_end).c_str()); } @@ -668,6 +719,12 @@ llama_kv_cache_dsv4::llama_kv_cache_dsv4( lid_state = std::make_unique( model, offload, unified, n_seq_max, DSV4_CSA_RATIO, 2*DSV4_CSA_RATIO, 2*model.hparams.indexer_head_size, "lid", filter_csa); + + // DSV4 attention reads compressed-K / compressor-state rows that the current + // graph does not necessarily overwrite; uninitialized buffer contents would + // otherwise leak in (instance-specific garbage) and corrupt recall. Zero all + // compressed buffers up front so reads of un-written rows are deterministic. + clear_compressed(true); } llama_memory_context_ptr llama_kv_cache_dsv4::init_batch( @@ -766,7 +823,7 @@ void llama_kv_cache_dsv4::clear(bool data) { restored_trim_pos.clear(); kv_raw->clear(data); - clear_compressed(data); + clear_compressed(true); // DSV4 compressed buffers must never expose stale/uninit rows } bool llama_kv_cache_dsv4::seq_rm(llama_seq_id seq_id, llama_pos p0, llama_pos p1) { @@ -808,7 +865,7 @@ bool llama_kv_cache_dsv4::seq_rm(llama_seq_id seq_id, llama_pos p0, llama_pos p1 restored_trim_pos.clear(); } - clear_compressed(false); + clear_compressed(true); } return res; @@ -818,28 +875,28 @@ void llama_kv_cache_dsv4::seq_cp(llama_seq_id seq_id_src, llama_seq_id seq_id_ds restored_trim_pos.clear(); kv_raw->seq_cp(seq_id_src, seq_id_dst, p0, p1); - clear_compressed(false); + clear_compressed(true); } void llama_kv_cache_dsv4::seq_keep(llama_seq_id seq_id) { restored_trim_pos.clear(); kv_raw->seq_keep(seq_id); - clear_compressed(false); + clear_compressed(true); } void llama_kv_cache_dsv4::seq_add(llama_seq_id seq_id, llama_pos p0, llama_pos p1, llama_pos shift) { restored_trim_pos.clear(); kv_raw->seq_add(seq_id, p0, p1, shift); - clear_compressed(false); + clear_compressed(true); } void llama_kv_cache_dsv4::seq_div(llama_seq_id seq_id, llama_pos p0, llama_pos p1, int d) { restored_trim_pos.clear(); kv_raw->seq_div(seq_id, p0, p1, d); - clear_compressed(false); + clear_compressed(true); } llama_pos llama_kv_cache_dsv4::seq_pos_min(llama_seq_id seq_id) const { diff --git a/src/llama-kv-cache-dsv4.h b/src/llama-kv-cache-dsv4.h index 980852d1b1a6..d62b87f86b54 100644 --- a/src/llama-kv-cache-dsv4.h +++ b/src/llama-kv-cache-dsv4.h @@ -161,6 +161,11 @@ class llama_kv_cache_dsv4_context : public llama_memory_context_i { // APE row ids, i.e. pos % ratio, for the compressor-state updates. std::vector state_pos; + // Current-ubatch source row ids and unique persistent-state + // destination row ids for deterministic ring-state updates. + std::vector state_persist_src_idxs; + std::vector state_persist_dst_idxs; + // Flattened source row ids used for state-backed commits. Source rows // index the graph-local [persistent_state | current_ubatch_scratch] // tensor. For overlapped compression the first half is previous rows diff --git a/src/models/deepseek-v4.cpp b/src/models/deepseek-v4.cpp index 749eb678b7ce..da3536f3747e 100644 --- a/src/models/deepseek-v4.cpp +++ b/src/models/deepseek-v4.cpp @@ -890,10 +890,13 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( csa_state_kv = dsv4_with_zero_dep(ctx0, csa_state_kv, csa_state_dep); csa_state_score = dsv4_with_zero_dep(ctx0, csa_state_score, csa_state_dep); + ggml_tensor * csa_persist_kv = ggml_get_rows(ctx0, csa_state_kv, inp_dsv4->get_csa().state_persist_src_idxs); + ggml_tensor * csa_persist_score = ggml_get_rows(ctx0, csa_state_score, inp_dsv4->get_csa().state_persist_src_idxs); + csa_state_kv = inp_dsv4->mctx->get_csa_state()->cpy_kv(ctx0, - csa_state_kv, inp_dsv4->get_csa().state_idxs, il); + csa_persist_kv, inp_dsv4->get_csa().state_persist_dst_idxs, il); csa_state_score = inp_dsv4->mctx->get_csa_state()->cpy_score(ctx0, - csa_state_score, inp_dsv4->get_csa().state_idxs, il); + csa_persist_score, inp_dsv4->get_csa().state_persist_dst_idxs, il); ggml_build_forward_expand(gf, csa_state_kv); ggml_build_forward_expand(gf, csa_state_score); @@ -946,10 +949,13 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( lid_state_kv = dsv4_with_zero_dep(ctx0, lid_state_kv, lid_state_dep); lid_state_score = dsv4_with_zero_dep(ctx0, lid_state_score, lid_state_dep); + ggml_tensor * lid_persist_kv = ggml_get_rows(ctx0, lid_state_kv, inp_dsv4->get_lid().state_persist_src_idxs); + ggml_tensor * lid_persist_score = ggml_get_rows(ctx0, lid_state_score, inp_dsv4->get_lid().state_persist_src_idxs); + lid_state_kv = inp_dsv4->mctx->get_lid_state()->cpy_kv(ctx0, - lid_state_kv, inp_dsv4->get_lid().state_idxs, il); + lid_persist_kv, inp_dsv4->get_lid().state_persist_dst_idxs, il); lid_state_score = inp_dsv4->mctx->get_lid_state()->cpy_score(ctx0, - lid_state_score, inp_dsv4->get_lid().state_idxs, il); + lid_persist_score, inp_dsv4->get_lid().state_persist_dst_idxs, il); ggml_build_forward_expand(gf, lid_state_kv); ggml_build_forward_expand(gf, lid_state_score); @@ -987,10 +993,13 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( hca_state_kv = dsv4_with_zero_dep(ctx0, hca_state_kv, hca_state_dep); hca_state_score = dsv4_with_zero_dep(ctx0, hca_state_score, hca_state_dep); + ggml_tensor * hca_persist_kv = ggml_get_rows(ctx0, hca_state_kv, inp_dsv4->get_hca().state_persist_src_idxs); + ggml_tensor * hca_persist_score = ggml_get_rows(ctx0, hca_state_score, inp_dsv4->get_hca().state_persist_src_idxs); + hca_state_kv = inp_dsv4->mctx->get_hca_state()->cpy_kv(ctx0, - hca_state_kv, inp_dsv4->get_hca().state_idxs, il); + hca_persist_kv, inp_dsv4->get_hca().state_persist_dst_idxs, il); hca_state_score = inp_dsv4->mctx->get_hca_state()->cpy_score(ctx0, - hca_state_score, inp_dsv4->get_hca().state_idxs, il); + hca_persist_score, inp_dsv4->get_hca().state_persist_dst_idxs, il); ggml_build_forward_expand(gf, hca_state_kv); ggml_build_forward_expand(gf, hca_state_score); From 22676c103ed256c6696dbffa1ff9754f581d658b Mon Sep 17 00:00:00 2001 From: Aman Gupta Date: Fri, 5 Jun 2026 02:57:46 +0200 Subject: [PATCH 11/13] support pro model: added by @fairydreaming --- conversion/deepseek.py | 3 ++- src/models/deepseek-v4.cpp | 2 +- 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/conversion/deepseek.py b/conversion/deepseek.py index bfa7ca36bd89..c4381c1d90e0 100644 --- a/conversion/deepseek.py +++ b/conversion/deepseek.py @@ -774,7 +774,8 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter if tensor_key == gguf.MODEL_TENSOR.FFN_GATE_TID2EID: return [] elif tensor_key == gguf.MODEL_TENSOR.ATTN_OUT_A: - data_torch = data_torch.reshape(self.hparams["o_groups"], self.hparams["o_lora_rank"], self.hparams["hidden_size"]) + attn_out_a_dim = int(self.hparams["num_attention_heads"] * self.hparams["head_dim"] / self.hparams["o_groups"]) + data_torch = data_torch.reshape(self.hparams["o_groups"], self.hparams["o_lora_rank"], attn_out_a_dim) return [(self._format_dsv4_tensor_name(tensor_key, bid, suffix), data_torch)] diff --git a/src/models/deepseek-v4.cpp b/src/models/deepseek-v4.cpp index da3536f3747e..c741132bdcab 100644 --- a/src/models/deepseek-v4.cpp +++ b/src/models/deepseek-v4.cpp @@ -103,7 +103,7 @@ void llama_model_deepseek_v4_flash::load_arch_tensors(llama_model_loader &) { layer.wq_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_B, "weight", i), {q_lora_rank, n_head * n_embd_head}, 0); layer.wkv = create_tensor(tn(LLM_TENSOR_ATTN_KV, "weight", i), {n_embd, n_embd_head}, 0); layer.attn_kv_norm = create_tensor(tn(LLM_TENSOR_ATTN_KV_NORM, "weight", i), {n_embd_head}, 0); - layer.wo_a = create_tensor(tn(LLM_TENSOR_ATTN_OUT_A, "weight", i), {n_embd, o_lora_rank, o_groups}, 0); + layer.wo_a = create_tensor(tn(LLM_TENSOR_ATTN_OUT_A, "weight", i), {n_head * n_embd_head / o_groups, o_lora_rank, o_groups}, 0); layer.wo_b = create_tensor(tn(LLM_TENSOR_ATTN_OUT_B, "weight", i), {o_groups * o_lora_rank, n_embd}, 0); layer.hc_attn_fn = create_tensor(tn(LLM_TENSOR_HC_ATTN_FN, nullptr, i), {hc_dim, hc_mix_dim}, 0); From 9e00db671b4b4273aa2d298ffb683088250a7c2b Mon Sep 17 00:00:00 2001 From: Aman Gupta Date: Fri, 5 Jun 2026 14:29:08 +0800 Subject: [PATCH 12/13] remove redundant V cache --- conversion/deepseek.py | 2 +- src/llama-graph.cpp | 16 ++++++-------- src/llama-kv-cache-dsv4.cpp | 5 ++++- src/llama-kv-cache-dsv4.h | 1 + src/llama-kv-cache-iswa.cpp | 21 +++++++++++++++++- src/llama-kv-cache-iswa.h | 16 ++++++++++++++ src/llama-kv-cache.cpp | 10 +++++---- src/models/deepseek-v4.cpp | 43 +++++++++++++++++++++++++++++++------ src/models/models.h | 8 +++++++ 9 files changed, 99 insertions(+), 23 deletions(-) diff --git a/conversion/deepseek.py b/conversion/deepseek.py index c4381c1d90e0..7715b81ff0b1 100644 --- a/conversion/deepseek.py +++ b/conversion/deepseek.py @@ -516,7 +516,7 @@ def _e8m0_to_float(scale: Tensor) -> Tensor: return scale.float() bits = scale.view(torch.uint8).float() - return torch.pow(torch.tensor(2.0, device=bits.device), bits - 127.0) + return torch.exp2(bits - 127.0) def _collect_source_dtypes(self) -> None: for name, gen in self.model_tensors.items(): diff --git a/src/llama-graph.cpp b/src/llama-graph.cpp index 35cd70e5fdad..59783dbd8dd8 100644 --- a/src/llama-graph.cpp +++ b/src/llama-graph.cpp @@ -566,7 +566,9 @@ void llm_graph_input_attn_kv_iswa::set_input(const llama_ubatch * ubatch) { // base tensors may not be allocated if there are no non-SWA attention layers if (self_k_idxs && self_k_idxs->buffer) { mctx->get_base()->set_input_k_idxs(self_k_idxs, ubatch); - mctx->get_base()->set_input_v_idxs(self_v_idxs, ubatch); + if (self_v_idxs) { + mctx->get_base()->set_input_v_idxs(self_v_idxs, ubatch); + } mctx->get_base()->set_input_kq_mask(self_kq_mask, ubatch, cparams.causal_attn); } @@ -574,7 +576,9 @@ void llm_graph_input_attn_kv_iswa::set_input(const llama_ubatch * ubatch) { // swa tensors may not be allocated if there are no SWA attention layers if (self_k_idxs_swa && self_k_idxs_swa->buffer) { mctx->get_swa()->set_input_k_idxs(self_k_idxs_swa, ubatch); - mctx->get_swa()->set_input_v_idxs(self_v_idxs_swa, ubatch); + if (self_v_idxs_swa) { + mctx->get_swa()->set_input_v_idxs(self_v_idxs_swa, ubatch); + } mctx->get_swa()->set_input_kq_mask(self_kq_mask_swa, ubatch, cparams.causal_attn); } @@ -2947,8 +2951,6 @@ llm_graph_input_dsv4 * llm_graph_context::build_inp_dsv4() const { { inp_raw->self_k_idxs = raw_ctx->get_base()->build_input_k_idxs(ctx0, ubatch); - inp_raw->self_v_idxs = raw_ctx->get_base()->build_input_v_idxs(ctx0, ubatch); - inp_raw->self_kq_mask = build_attn_inp_kq_mask(ctx0, raw_ctx->get_base(), ubatch, cparams); inp_raw->self_kq_mask_cnv = inp_raw->self_kq_mask; } @@ -2957,18 +2959,12 @@ llm_graph_input_dsv4 * llm_graph_context::build_inp_dsv4() const { GGML_ASSERT(hparams.swa_type != LLAMA_SWA_TYPE_NONE && "DSV4 expects SWA raw cache"); inp_raw->self_k_idxs_swa = raw_ctx->get_swa()->build_input_k_idxs(ctx0, ubatch); - inp_raw->self_v_idxs_swa = raw_ctx->get_swa()->build_input_v_idxs(ctx0, ubatch); - inp_raw->self_kq_mask_swa = build_attn_inp_kq_mask(ctx0, raw_ctx->get_swa(), ubatch, cparams); inp_raw->self_kq_mask_swa_cnv = inp_raw->self_kq_mask_swa; } inp_raw->self_k_rot = raw_ctx->get_base()->build_input_k_rot(ctx0); - inp_raw->self_v_rot = raw_ctx->get_base()->build_input_v_rot(ctx0); - inp_raw->self_k_rot_swa = raw_ctx->get_swa()->build_input_k_rot(ctx0); - inp_raw->self_v_rot_swa = raw_ctx->get_swa()->build_input_v_rot(ctx0); - auto inp = std::make_unique(cparams, std::move(inp_raw), mctx_cur); dsv4_build_comp_inputs(ctx0, inp->inp_csa, mctx_cur->get_csa_plan(), "csa"); diff --git a/src/llama-kv-cache-dsv4.cpp b/src/llama-kv-cache-dsv4.cpp index a786c732ed41..1737d62ae2e9 100644 --- a/src/llama-kv-cache-dsv4.cpp +++ b/src/llama-kv-cache-dsv4.cpp @@ -632,6 +632,7 @@ llama_kv_cache_dsv4::llama_kv_cache_dsv4( uint32_t n_pad, const layer_filter_cb & filter, const layer_reuse_cb & reuse) : + hparams_raw(model.hparams), hparams_csa(model.hparams), hparams_hca(model.hparams), hparams_lid(model.hparams) { @@ -646,8 +647,10 @@ llama_kv_cache_dsv4::llama_kv_cache_dsv4( LLAMA_LOG_INFO("%s: creating DSV4 raw KV cache\n", __func__); + dsv4_make_k_only(hparams_raw); + kv_raw = std::make_unique( - model, type_k, type_v, + model, hparams_raw, type_k, type_v, v_trans, offload, swa_full, unified, kv_size, n_seq_max, n_ubatch, n_pad, filter_raw, reuse); diff --git a/src/llama-kv-cache-dsv4.h b/src/llama-kv-cache-dsv4.h index d62b87f86b54..e9cdd8e27efe 100644 --- a/src/llama-kv-cache-dsv4.h +++ b/src/llama-kv-cache-dsv4.h @@ -131,6 +131,7 @@ class llama_kv_cache_dsv4 : public llama_memory_i { llama_dsv4_comp_state * get_lid_state() const; private: + llama_hparams hparams_raw; llama_hparams hparams_csa; llama_hparams hparams_hca; llama_hparams hparams_lid; diff --git a/src/llama-kv-cache-iswa.cpp b/src/llama-kv-cache-iswa.cpp index 9b9f17903637..4c1f9ba82a8d 100644 --- a/src/llama-kv-cache-iswa.cpp +++ b/src/llama-kv-cache-iswa.cpp @@ -24,7 +24,26 @@ llama_kv_cache_iswa::llama_kv_cache_iswa( uint32_t n_ubatch, uint32_t n_pad, const layer_filter_cb & filter, - const layer_reuse_cb & reuse) : hparams(model.hparams), unified(unified) { + const layer_reuse_cb & reuse) : + llama_kv_cache_iswa(model, model.hparams, type_k, type_v, v_trans, offload, swa_full, unified, + kv_size, n_seq_max, n_ubatch, n_pad, filter, reuse) { +} + +llama_kv_cache_iswa::llama_kv_cache_iswa( + const llama_model & model, + const llama_hparams & hparams, + ggml_type type_k, + ggml_type type_v, + bool v_trans, + bool offload, + bool swa_full, + bool unified, + uint32_t kv_size, + uint32_t n_seq_max, + uint32_t n_ubatch, + uint32_t n_pad, + const layer_filter_cb & filter, + const layer_reuse_cb & reuse) : hparams(hparams), unified(unified) { // chain filters const layer_filter_cb filter_base = [&](int32_t il) { diff --git a/src/llama-kv-cache-iswa.h b/src/llama-kv-cache-iswa.h index 70ab22f0d608..306ed44feb93 100644 --- a/src/llama-kv-cache-iswa.h +++ b/src/llama-kv-cache-iswa.h @@ -28,6 +28,22 @@ class llama_kv_cache_iswa : public llama_memory_i { const layer_filter_cb & filter, const layer_reuse_cb & reuse); + llama_kv_cache_iswa( + const llama_model & model, + const llama_hparams & hparams, + ggml_type type_k, + ggml_type type_v, + bool v_trans, + bool offload, + bool swa_full, + bool unified, + uint32_t kv_size, + uint32_t n_seq_max, + uint32_t n_ubatch, + uint32_t n_pad, + const layer_filter_cb & filter, + const layer_reuse_cb & reuse); + ~llama_kv_cache_iswa() = default; // diff --git a/src/llama-kv-cache.cpp b/src/llama-kv-cache.cpp index 31b3d3d5844f..a192e731d318 100644 --- a/src/llama-kv-cache.cpp +++ b/src/llama-kv-cache.cpp @@ -177,10 +177,12 @@ llama_kv_cache::llama_kv_cache( n_embd_head_k_all = -1; } - if (n_embd_head_v_all == 0) { - n_embd_head_v_all = (int32_t) hparams.n_embd_head_v(il); - } else if (n_embd_head_v_all > 0 && n_embd_head_v_all != (int32_t) hparams.n_embd_head_v(il)) { - n_embd_head_v_all = -1; + if (!is_mla) { + if (n_embd_head_v_all == 0) { + n_embd_head_v_all = (int32_t) hparams.n_embd_head_v(il); + } else if (n_embd_head_v_all > 0 && n_embd_head_v_all != (int32_t) hparams.n_embd_head_v(il)) { + n_embd_head_v_all = -1; + } } // [TAG_V_CACHE_VARIABLE] diff --git a/src/models/deepseek-v4.cpp b/src/models/deepseek-v4.cpp index c741132bdcab..3b9e0a6afc76 100644 --- a/src/models/deepseek-v4.cpp +++ b/src/models/deepseek-v4.cpp @@ -651,7 +651,6 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_csa_lid_attention( const llama_kv_cache_context * mctx_swa = inp_attn->mctx->get_swa(); ggml_build_forward_expand(gf, mctx_swa->cpy_k(ctx0, kv, inp_attn->get_k_idxs_swa(), il)); - ggml_build_forward_expand(gf, mctx_swa->cpy_v(ctx0, kv, inp_attn->get_v_idxs_swa(), il)); ggml_tensor * raw_k = mctx_swa->get_k(ctx0, il); if (raw_k->type != GGML_TYPE_F32) { @@ -709,7 +708,6 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hca_attention( const llama_kv_cache_context * mctx_swa = inp_attn->mctx->get_swa(); ggml_build_forward_expand(gf, mctx_swa->cpy_k(ctx0, kv, inp_attn->get_k_idxs_swa(), il)); - ggml_build_forward_expand(gf, mctx_swa->cpy_v(ctx0, kv, inp_attn->get_v_idxs_swa(), il)); ggml_tensor * raw_k = mctx_swa->get_k(ctx0, il); if (raw_k->type != GGML_TYPE_F32) { @@ -748,6 +746,42 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_hca_attention( return out; } +ggml_tensor * llama_model_deepseek_v4_flash::graph::build_raw_attention( + llm_graph_input_attn_kv_iswa * inp_attn, + ggml_tensor * q, + ggml_tensor * kv, + ggml_tensor * sinks, + float kq_scale, + int il) const { + const bool is_swa = hparams.is_swa(il); + + ggml_tensor * k_rot = is_swa ? inp_attn->self_k_rot_swa : inp_attn->self_k_rot; + ggml_tensor * v_rot = is_swa ? inp_attn->self_v_rot_swa : inp_attn->self_v_rot; + GGML_ASSERT(v_rot == nullptr); + + if (k_rot) { + q = ggml_mul_mat(ctx0, k_rot, q); + kv = ggml_mul_mat(ctx0, k_rot, kv); + } + + ggml_build_forward_expand(gf, q); + ggml_build_forward_expand(gf, kv); + + const llama_kv_cache_context * mctx_cur = is_swa ? inp_attn->mctx->get_swa() : inp_attn->mctx->get_base(); + const auto & k_idxs = is_swa ? inp_attn->get_k_idxs_swa() : inp_attn->get_k_idxs(); + + ggml_build_forward_expand(gf, mctx_cur->cpy_k(ctx0, kv, k_idxs, il)); + + const auto & kq_mask = is_swa ? inp_attn->get_kq_mask_swa() : inp_attn->get_kq_mask(); + + ggml_tensor * k = mctx_cur->get_k(ctx0, il); + + ggml_tensor * out = build_attn_mha(q, k, k, nullptr, kq_mask, sinks, nullptr, kq_scale, il); + cb(out, "attn_raw", il); + + return out; +} + ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( const llama_model & model, llm_graph_input_dsv4 * inp_dsv4, @@ -1021,11 +1055,8 @@ ggml_tensor * llama_model_deepseek_v4_flash::graph::build_attention( out = build_hca_attention(inp_dsv4, inp_attn, q, kv, layer.attn_sinks, 1.0f/sqrtf(float(n_embd_head)), il); } else { - out = build_attn(inp_attn, - nullptr, nullptr, nullptr, - q, kv, kv, nullptr, layer.attn_sinks, nullptr, + out = build_raw_attention(inp_attn, q, kv, layer.attn_sinks, 1.0f/sqrtf(float(n_embd_head)), il); - cb(out, "attn_raw", il); } out = ggml_reshape_3d(ctx0, out, n_embd_head, n_head, nt); diff --git a/src/models/models.h b/src/models/models.h index 46cdb1dcbdbc..56c560a4a2da 100644 --- a/src/models/models.h +++ b/src/models/models.h @@ -1149,6 +1149,14 @@ struct llama_model_deepseek_v4_flash : public llama_model_base { float kq_scale, int il) const; + ggml_tensor * build_raw_attention( + llm_graph_input_attn_kv_iswa * inp_attn, + ggml_tensor * q, + ggml_tensor * kv, + ggml_tensor * sinks, + float kq_scale, + int il) const; + ggml_tensor * build_hc_weighted_sum( ggml_tensor * x, ggml_tensor * weights) const; From 94e724b8dbc0e894851eea5eec814af6673fd6d8 Mon Sep 17 00:00:00 2001 From: Piotr Wilkin Date: Fri, 5 Jun 2026 14:34:29 +0200 Subject: [PATCH 13/13] Chat template --- .../templates/deepseek-ai-DeepSeek-V4.jinja | 112 ++++++++++++++++++ 1 file changed, 112 insertions(+) create mode 100644 models/templates/deepseek-ai-DeepSeek-V4.jinja diff --git a/models/templates/deepseek-ai-DeepSeek-V4.jinja b/models/templates/deepseek-ai-DeepSeek-V4.jinja new file mode 100644 index 000000000000..f19f787b1b7e --- /dev/null +++ b/models/templates/deepseek-ai-DeepSeek-V4.jinja @@ -0,0 +1,112 @@ +{%- if not add_generation_prompt is defined -%} + {%- set add_generation_prompt = false -%} +{%- endif -%} +{%- if not thinking is defined -%} + {%- if enable_thinking is defined -%} + {%- set thinking = enable_thinking -%} + {%- else -%} + {%- set thinking = false -%} + {%- endif -%} +{%- endif -%} +{%- set dsml_token = '｜DSML｜' -%} +{%- set thinking_start_token = '' -%} +{%- set thinking_end_token = '' -%} +{%- set tools_header = '## Tools\n\nYou have access to a set of tools to help answer the user\'s question. You can invoke tools by writing a "<' + dsml_token + 'tool_calls>" block like the following:\n\n<' + dsml_token + 'tool_calls>\n<' + dsml_token + 'invoke name="$TOOL_NAME">\n<' + dsml_token + 'parameter name="$PARAMETER_NAME" string="true|false">$PARAMETER_VALUE\n...\n\n<' + dsml_token + 'invoke name="$TOOL_NAME2">\n...\n\n\n\nString parameters should be specified as is and set `string="true"`. For all other types (numbers, booleans, arrays, objects), pass the value in JSON format and set `string="false"`.\n\nIf thinking_mode is enabled (triggered by ' + thinking_start_token + '), you MUST output your complete reasoning inside ' + thinking_start_token + '...' + thinking_end_token + ' BEFORE any tool calls or final response.\n\nOtherwise, output directly after ' + thinking_end_token + ' with tool calls or final response.\n\n### Available Tool Schemas\n\n' -%} +{%- set tools_footer = '\nYou MUST strictly follow the above defined tool name and parameter schemas to invoke tool calls.\n' -%} +{%- set ns = namespace(system_prompt='', is_first_sp=true) -%} +{%- for message in messages -%} + {%- if message['role'] == 'system' -%} + {%- if ns.is_first_sp -%} + {%- set ns.system_prompt = ns.system_prompt + (message['content'] or '') -%} + {%- set ns.is_first_sp = false -%} + {%- else -%} + {%- set ns.system_prompt = ns.system_prompt + '\n\n' + (message['content'] or '') -%} + {%- endif -%} + {%- endif -%} +{%- endfor -%} +{%- if tools is defined and tools -%} + {%- set ts = namespace(schemas='') -%} + {%- for tool in tools -%} + {%- if tool['type'] == 'function' -%} + {%- set ts.schemas = ts.schemas + (tool['function'] | tojson) + '\n' -%} + {%- endif -%} + {%- endfor -%} + {%- if ns.system_prompt -%} + {%- set ns.system_prompt = ns.system_prompt + '\n\n' + tools_header + ts.schemas + tools_footer -%} + {%- else -%} + {%- set ns.system_prompt = tools_header + ts.schemas + tools_footer -%} + {%- endif -%} +{%- endif -%} +{{- bos_token -}} +{{- ns.system_prompt -}} +{%- set last_user_idx = namespace(value=-1) -%} +{%- for message in messages -%} + {%- if message['role'] == 'user' or message['role'] == 'developer' or message['role'] == 'tool' -%} + {%- set last_user_idx.value = loop.index0 -%} + {%- endif -%} +{%- endfor -%} +{%- set state = namespace(in_user=false) -%} +{%- for message in messages -%} + {%- if message['role'] == 'user' or message['role'] == 'developer' -%} + {%- if state.in_user -%} + {{- '\n\n' -}} + {%- else -%} + {{- '<｜User｜>' -}} + {%- set state.in_user = true -%} + {%- endif -%} + {{- message['content'] or '' -}} + {%- elif message['role'] == 'tool' -%} + {%- if state.in_user -%} + {{- '\n\n' -}} + {%- else -%} + {{- '<｜User｜>' -}} + {%- set state.in_user = true -%} + {%- endif -%} + {{- '' + (message['content'] or '') + '' -}} + {%- elif message['role'] == 'assistant' -%} + {%- set state.in_user = false -%} + {{- '<｜Assistant｜>' -}} + {%- set is_after_last_user = loop.index0 > last_user_idx.value -%} + {%- if is_after_last_user and thinking -%} + {{- thinking_start_token -}} + {%- if message['reasoning_content'] is defined and message['reasoning_content'] -%} + {{- message['reasoning_content'] -}} + {%- endif -%} + {{- thinking_end_token -}} + {%- else -%} + {{- thinking_end_token -}} + {%- endif -%} + {%- if message['content'] is defined and message['content'] -%} + {{- message['content'] -}} + {%- endif -%} + {%- if message['tool_calls'] -%} + {{- '\n\n<' + dsml_token + 'tool_calls>\n' -}} + {%- for tool in message['tool_calls'] -%} + {%- set func = tool['function'] -%} + {{- '<' + dsml_token + 'invoke name="' + func['name'] + '">\n' -}} + {%- set args = func['arguments'] -%} + {%- if args is string -%} + {%- set args = args | from_json -%} + {%- endif -%} + {%- for key, val in args.items() -%} + {%- if val is string -%} + {{- '<' + dsml_token + 'parameter name="' + key + '" string="true">' + val + '\n' -}} + {%- else -%} + {{- '<' + dsml_token + 'parameter name="' + key + '" string="false">' + (val | tojson) + '\n' -}} + {%- endif -%} + {%- endfor -%} + {{- '\n' -}} + {%- endfor -%} + {{- '' -}} + {%- endif -%} + {{- '<｜end▁of▁sentence｜>' -}} + {%- endif -%} +{%- endfor -%} +{%- if add_generation_prompt -%} + {{- '<｜Assistant｜>' -}} + {%- if thinking -%} + {{- thinking_start_token -}} + {%- else -%} + {{- thinking_end_token -}} + {%- endif -%} +{%- endif -%} \ No newline at end of file