Conversation
* use OneOf instead of list syntax sugar for confuse * update deps * let yaml have raw yaml datetimes instead of strings * use simplejson decode errors
commit 1fb394e109263188a16c4addeaab87bbdfdf882e
Author: AJ Slater <aj@slater.net>
Date: Wed Feb 11 17:09:25 2026 -0800
generate-schema scripts
commit fc9b4f5c27db827ae1592010b01708865cf3733e
Author: AJ Slater <aj@slater.net>
Date: Wed Feb 11 17:09:08 2026 -0800
format schemas
commit 9ccdf70d8c2318220c443714e509b6746f19a90e
Author: AJ Slater <aj@slater.net>
Date: Wed Feb 11 16:39:04 2026 -0800
fix schema
commit 1a082c52887571cd258ebbc467846461c8e9686f
Author: AJ Slater <aj@slater.net>
Date: Wed Feb 11 13:29:02 2026 -0800
add marshmallow jsonschema
mokkari 4.2.0 added average_rating and rating_count to the Issue model. Map them into the native community_rating block. Unrated issues (null average, zero count) emit nothing, and a zero count is never passed through to the minimum=1 native field where it would clamp up to 1. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
write_metadata() and BulkWriteItem accept delete_keys glom key paths, layered onto general.delete_keys. Empty patch values are pruned on schema load, so update-mode writes could never clear an existing tag; delete_keys is the explicit clear mechanism. A pure-clear write may pass an empty patch. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Extracted pdf pages were named from the archive type and only renamed if the pdf reader reported the format it served. A reader that hands back a page's embedded image instead of a one page pdf without reporting an extension left jpeg data on disk named ".pdf". Sniff the data's magic bytes at the pdf read boundary to fill in a format the reader didn't report, so pages are named for what they contain. A reported format still wins, so nothing that names correctly today changes. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
--extract-pages RANGE with --pdf-pages pdf (or the default pdf format) now writes one merged pdf spanning the range instead of a file per page. The --pdf-pages help table is derived from the installed pdffile's PageFormat enum, so it can never advertise a value the CLI rejects, and all five modes are documented. Also stop counting a pdf's embedded metadata (e.g. an embedded ComicInfo.xml) as a page, which inflated the page count and left a stray file when extracting ranges. Requires comicbox-pdffile 0.6.3 for the read_pdf page-range argument. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- RUF036: order None last in union return types - PLC0414: keep the explicit re-export aliases with a targeted noqa - ISC004: parenthesize an implicit string concatenation - PLR0917: extend the TestParser noqa to cover its positional-arg count - Split bin/roman.py _walk_tree's scan loop into a _scan_dir helper to drop its cognitive complexity below threshold Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Default conversion wrote raw .ppm pixmaps that no comic reader (including comicbox itself) recognizes as pages, yielding a zero-page book — and --delete-orig then removed the working original. Conversion now defaults to pixmap_jpeg whole-page renders, which also apply pdf display rotation. --pdf-pages image wrote rotated pages as stored — sideways or upside down for scans relying on /Rotate (or a rotated content-stream placement) for display. Rotated image-dominant pages are now rendered to match the displayed orientation (pdffile 0.6.3 classify); unrotated pages still extract their original bytes untouched. Also store converted page images uncompressed: the compression decision ran before pdf pages gained their image suffix, so page images were pointlessly deflated. Ships as v4.5.1 alongside the pdf range extraction and page naming fixes already in NEWS; version stays 4.5.1. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The release ships as v4.5.0; both unreleased sections merge into one. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
radon counts each assert as a branch, so the 10-assert core-fields test ranked C (11). Group the scalar and nested-name checks into single dict comparisons — rank B (6), same coverage, clearer failure diagnostics. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
YesNoEnum was missing its YES member, so EnumBooleanField's truthy coercion produced the string "Yes", failed the enum lookup, and got swallowed by TrapExceptionsMeta. BlackAndWhite was dead in both directions: monochrome could never be read from a comic nor written to one, leaving the field impossible to set or clear. The member was dropped in 04342a3 when YesNoField was rebased from BooleanField onto EnumBooleanField. Restore the original semantics: deserialize to a bool, with the schema's Unknown meaning no value. An unset monochrome omits the tag rather than writing Unknown, matching how every other absent field is treated. Also fix the shared enum-boolean base, which mishandled False for Manga as well: str(False) matched no enum member, and _serialize raised AttributeError on any bool. Coercion now consults the enum vocabulary first so multi-valued enums keep their extra members, then falls back to reading the value as a boolean. The tag had no test coverage, which is how this shipped. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
mokkari 4.3.0 added `alt_names` to its Series and IssueSeries schemas. Map them to comicbox reprints, the same representation MetronInfo.xml's Series/AlternativeNames already round-trip through. ComicVine has exposed volume `aliases` all along and comicbox ignored them. Map those too: `get()` injects them from the volume it already fetches for the publisher, so no extra API call. Alternative names also feed online matching. `s_series` scores best-of over the candidate's primary name and its aliases, and the pre-call volume filter keeps a volume when any alias clears the threshold, so a comic filed under a localized title matches its canonical volume. Aliases only ever widen the gate; they cannot drop a volume the primary name would have kept. Metron's search results carry no alternative names — mokkari's BasicSeries lacks the field, unlike the IssueSeries on issue detail — so its candidates score exactly as before. Reading them at match time would cost an API call per candidate. The contributing-signals gate counts alternative names as candidate series data. Since s_series can now score from them alone, omitting them would drop W_SERIES from the renormalization denominator and inflate the remaining signals. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Keys like "series:178012" mirror the urn form comicbox writes to notes/GTIN, but were interpolated verbatim into issue-shaped URL templates, producing broken links for ComicVine, Metron, Grand Comics Database, and League of Comic Geeks. A new normalize_key() strips urn/source/type prefixes (a type prefix overrides the positional id type) and runs in create_identifier and the computed url pass, which now also emits corrected keys, so native YAML/JSON/CLI input gets comicvine long-code normalization too. Also: source:type:key tag strings no longer truncate to the type word, every notes urn is read instead of only the first, online-tagged story arcs get arc-typed urls, and keys with unrecognized prefixes emit no url rather than a broken one. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
mokkari 4.4.0 added an `api_token` parameter that authenticates with a
Bearer token and takes precedence over username/passwd. Comicbox
hard-coded "user AND password" for Metron in three places, so token
holders could not authenticate.
Reuse the generic `key` credential field rather than adding a new
`token` field: `--auth metron:key=`, `COMICBOX_METRON_KEY`, and
`online.auth.metron.key` already flowed through the CLI, env, confuse,
and redacting-repr plumbing; only the consumers needed changing.
The shared-session cache key grows to (user, password, key) so
token-only sessions do not all collide on ("", ""), which would have
made every distinct token share one account's rate-limit state.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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
2 participants
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.
with
--auth metron:key=TOKEN, theCOMICBOX_METRON_KEYenvironmentvariable, or the
online.auth.metron.keyconfig key. The token wins ifboth are set.