Repository navigation
Chore/cleanup refactor - #18
Merged
Merged
Conversation
makeitdouble
commented
Jun 14, 2026
makeitdouble
commented
Owner
- runtime memory split into L1 / L2 / L3 modules
- old ui/static/js/telemetry.js replaced with runtime/*.js
- brain_client.py unloaded: brain_client_utils.py and brain_context_builder.py moved out
- agent pipeline made explicit: translator -> brain -> validator -> translator for TRANSLATE_RESPONSE
- streaming fixed when translating responses: think remains visible, final translation is done in chunks
- removed auto-arm background fact-check after each turn
- remember_session can now be armed from user text and wait for memory update completion
- README expanded with examples of L1/L2/L3 snapshot contracts
- added/updated tests for behavior, routing, runtime actions, memory, and pending usage
- split memory rules from memory utils - move brain client helpers into utils modules - extract brain runtime context builder - move remember event triggers into behavior contract
…o chore/cleanup_refactor
- add memory line property split and hide logic - hide trace property in UI
- scope latest runtime memory to per-tab runtime session ids
- add clearing stale snapshots from other tabs - scope latest runtime memory to each tab's runtimeSessionId - clean noisy/internal logs and markers from visible UI output
- add verbatim user_message tracking to L1 runtime memory - build L2 pattern evidence from repeated user_message snapshots - refresh latest runtime snapshot when L2 evidence is appended - keep L2 evidence metadata in hover while hiding noisy brackets in UI - add runtime memory token metrics to panel title - cover repeated user message evidence with tests - support replacing latest runtime memory snapshot on frontend update
- fix user_idle duplication
- guard L2 occurrence evidence - drop model-made L2 evidence without deterministic support - prevent one-snapshot repeated-message patterns
- keep L2 evidence lines immutable for L1 - require separate L2 evidence status keys - prevent extra text after L2 occurrence metadata - avoid storing raw turn counters in L1 status
- add collapsed state for console and JIN panels
- better chat column centering - add subtle background dimming when panels are collapsed
- limit L1 memory trace payload to filtered hot_traces - stop passing L2 pattern memory and stable/fading trace buckets into L1 prompts - add configurable hot trace exclusions - tighten action guard trigger matching around word boundaries - emit remember_session runtime action when armed - cover memory prompt filtering and remember_session behavior with tests
- remove legacy fact_check logic
- remove legacy runtime key support
- remove legacy embedded runtime memory
- separate memory constants from memory logic - add comment for each constant
- split memory.py, move L1/L2/L3 into dedicated modules - move shared summarizer helpers into memory_common - move runtime memory emitters into memory_events
- tighten countdown recall behavior with additional dynamic rules - update recall detection in recal_word probe test - update memory detection in movie_closure probe test
- update README.md - reorder future features - add L1\L2\L3 memory layers structure examples
- split telemetry.js first iteration - add JinRuntime core, storage, and memory model modules - align session/search/event runtime action rows with chat messages
- split telemetry.js second iteration - add JinRuntime idle module
- split telemetry.js third iteration - add JinRuntime feedback module
- tighten save session flow with more strict rules
- split telemetry.js fourth iteration - add JinRuntime session module
- split telemetry.js fifth iteration - add JinRuntime panel module
- split telemetry.js sixth iteration - add JinRuntime view module
- remove telemetry.js - add JinRuntime facade
- clean up nodes logic - route optional response translation through translator node - keep validator node focused on final answer validation
- clean up legacy deep_thought logic from brain client
- split memory_utils into layered memory utils
- fix tests
- fix chunk streaming for translation pipeline
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.