bvandeusen 3b3e7565fb fix(ml): align tagger + downloader with Camie v2 actual layout (model.onnx -> camie-tagger-v2.onnx + JSON metadata + ImageNet preprocessing + sigmoid on refined output) ...

The HF repo Camais03/camie-tagger-v2 has camie-tagger-v2.onnx (789 MB)
+ camie-tagger-v2-metadata.json (7.77 MB) at root, NOT model.onnx +
selected_tags.csv. Tags ship as nested JSON (dataset_info.tag_mapping)
not CSV. Per the published onnx_inference.py reference: input is NCHW
not NHWC, normalize with ImageNet mean/std, pad-square color (124,116,
104), sigmoid the second output (refined predictions) not the first.

Operator hit this during the IR migration ML backfill — download_models
silently fetched only 3 json files (allow_patterns matched nothing
useful), tagger.load() then raised RuntimeError. Fetched the actual
v2 layout via WebFetch, rewrote tagger to match published reference.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-05-25 02:25:30 -04:00

..

__init__.py

feat: add Quart app factory, config loader, and /api/health endpoint

2026-05-14 07:32:17 -04:00

conftest.py

fix(integration): lazy seed-snapshot (unblock unit job); isolate recover test from eager import

2026-05-15 22:43:46 -04:00

test_api_admin.py

fix(fc3k): add origin=imported_filesystem to test ImageRecord ctors (second NOT NULL column after mime)

2026-05-25 01:28:54 -04:00

test_api_aliases.py

feat(fc2b): add /api/allowlist and /api/aliases endpoints

2026-05-15 07:46:38 -04:00

test_api_allowlist.py

feat(fc2b): add /api/allowlist and /api/aliases endpoints

2026-05-15 07:46:38 -04:00

test_api_artist.py

feat(fc2c-i): artist overview + paged images endpoints

2026-05-15 15:52:02 -04:00

test_api_artists_create.py

feat(fc3a): /api/artists POST + autocomplete

2026-05-20 12:55:51 -04:00

test_api_artists_directory.py

fc3f: GET /api/artists/directory — cursor + q + platform filters with validation

2026-05-21 22:14:42 -04:00

test_api_attachments.py

feat(attachments): provenance payload attachments + download route

2026-05-19 11:16:01 -04:00

test_api_bulk_tags.py

fix(bulk): node-safe toast (globalThis.window) + common-tags test matches non-empty spec guard

2026-05-16 18:15:23 -04:00

test_api_credentials.py

feat(fc3b): /api/credentials blueprint (CRUD + X-Extension-Key auth)

2026-05-20 18:35:54 -04:00

test_api_downloads.py

feat(fc3c): /api/downloads blueprint (list + detail w/ keyset pagination)

2026-05-20 20:44:28 -04:00

test_api_extension_api_key.py

feat(fc3b): extension_api_key endpoints + create_app bootstraps credential key

2026-05-20 18:36:28 -04:00

test_api_extension.py

fc3g: integration tests for /api/extension/* and /extension/<filename>

2026-05-23 23:40:00 -04:00

test_api_gallery.py

fix(fc2a): align CI with FabledRulebook — lint + short unit tests only

2026-05-14 15:23:07 -04:00

test_api_import_admin.py

feat(integrity): /trigger accepts mode=verify → enqueues verify_integrity

2026-05-19 17:52:18 -04:00

test_api_migrate.py

fc3h(tests): drop pinned migration-backup tests (retired surface; coverage moved to test_backup_service.py + test_api_system_backup.py)

2026-05-24 23:03:57 -04:00

test_api_ml_admin.py

feat(fc2b): add tag rename (PATCH /api/tags/<id>) + /api/ml admin

2026-05-15 07:54:12 -04:00

test_api_platforms.py

feat(fc3b): /api/platforms blueprint (read-only registry view)

2026-05-20 18:34:23 -04:00

test_api_posts.py

fc3e: /api/posts blueprint — list (with cursor + filters) + detail

2026-05-21 18:55:18 -04:00

test_api_provenance.py

feat(provenance): /api/provenance image + post endpoints

2026-05-18 18:07:06 -04:00

test_api_series.py

feat(series): /api/series pages/add/remove/reorder/cover routes

2026-05-16 23:35:12 -04:00

test_api_settings_downloader.py

feat(fc3c): expose download_rate_limit_seconds + download_validate_files in /api/settings/import

2026-05-20 20:44:58 -04:00

test_api_settings.py

feat(ui): bump thumbnail sizes 1.5x, make Settings fluid, add subscription_count stat

2026-05-22 23:18:22 -04:00

test_api_showcase.py

feat(fc2c-i): showcase random-sample service and endpoint

2026-05-15 15:49:25 -04:00

test_api_sources_check.py

feat(fc3c): POST /api/sources/<id>/check — enqueue download

2026-05-20 20:43:48 -04:00

test_api_sources.py

fc3d: api_sources tests cover consecutive_failures + next_check_at fields

2026-05-21 16:03:48 -04:00

test_api_suggestions.py

feat(fc2b): add /api/images/<id>/suggestions endpoints

2026-05-15 07:45:55 -04:00

test_api_system_activity.py

fix(fc3i): single-line celery.signals import + INT32 bounds on target_id + dict.fromkeys + rewrite retry test as direct _finalize unit

2026-05-24 21:22:42 -04:00

test_api_system_backup.py

fix(fc3h): split semicolon-stacked statements (E702) and bridge v-dialog v-model to avoid prop write

2026-05-24 23:37:06 -04:00

test_api_tag_merge.py

feat(api): rich 409 merge hint and POST tags merge endpoint with backfill enqueue

2026-05-16 13:40:37 -04:00

test_api_tags_directory.py

fix(tags): invalid 'copyright' kind 500 — KINDS uses real enum 'fandom'; directory 400s on bad kind

2026-05-16 15:51:27 -04:00

test_api_tags.py

fix(fc2a): align CI with FabledRulebook — lint + short unit tests only

2026-05-14 15:23:07 -04:00

test_archive_extractor.py

feat(attachments): fail-soft archive_extractor (zip/cbz/rar/7z)

2026-05-19 11:11:37 -04:00

test_artist_directory_service.py

fc3f: ArtistDirectoryService — cursor-paginated artists directory mirroring TagDirectoryService

2026-05-21 22:14:15 -04:00

test_artist_service_find_or_create.py

feat(fc3a): ArtistService.find_or_create + autocomplete

2026-05-20 12:54:47 -04:00

test_artist_service.py

fix(test): artist_service _fixture sets artist_id (post-vii-c ArtistService reads it)

2026-05-18 21:47:45 -04:00

test_attachment_store.py

feat(attachments): sha-addressed AttachmentStore

2026-05-19 11:11:07 -04:00

test_backfill_phash.py

feat(deep-scan): keyset-paginated backfill_phash task

2026-05-19 14:59:04 -04:00

test_backup_service.py

fix(fc3h): split semicolon-stacked statements (E702) and bridge v-dialog v-model to avoid prop write

2026-05-24 23:37:06 -04:00

test_bulk_tag_service.py

feat(bulk): BulkTagService (common_tags/bulk_add/bulk_remove) + 3 image bulk-tag routes

2026-05-16 18:02:01 -04:00

test_celery_smoke.py

feat: configure Celery with per-feature queue lanes and a smoke task

2026-05-14 07:34:47 -04:00

test_cleanup_old_download_events.py

fc3d: cleanup_old_download_events daily task (terminal-only, configurable retention)

2026-05-21 16:02:36 -04:00

test_cleanup_service.py

fix(fc3k): add origin=imported_filesystem to test ImageRecord ctors (second NOT NULL column after mime)

2026-05-25 01:28:54 -04:00

test_credential_crypto.py

fix(fc3b): ruff I001 (blank-line) + F841 (unused crypto binding)

2026-05-20 18:46:59 -04:00

test_credential_service.py

feat(fc3b): CredentialService — CRUD + on-demand cookies.txt materialisation

2026-05-20 18:35:07 -04:00

test_download_models.py

feat(fc2b): importer enqueues tag_and_embed + ml-worker model self-heal

2026-05-15 07:45:09 -04:00

test_download_service.py

fc3d: DownloadService finalize hook — Source.consecutive_failures/last_error/last_checked_at

2026-05-21 16:00:40 -04:00

test_download_source_task.py

fix(fc3c): explicit import of download task module in registration test

2026-05-20 21:20:30 -04:00

test_extensions_session.py

refactor(fc2c-i): single shared async engine with lifecycle dispose

2026-05-15 13:53:42 -04:00

test_file_validator.py

feat(fc3c): FileValidator — magic-byte head/tail check for fresh writes

2026-05-20 20:39:17 -04:00

test_gallery_artist_payload.py

feat(provenance): _artists_for resolves via artist_id (closes vii-b gap)

2026-05-18 21:36:16 -04:00

test_gallery_dl_service.py

feat(fc3c): GalleryDLService — subprocess wrap + error classification + validate/quarantine

2026-05-20 20:41:42 -04:00

test_gallery_provenance_filter.py

fix(test): ruff F841 — don't bind unmatched seed images (tuple-unpack flagged)

2026-05-18 18:29:18 -04:00

test_gallery_service.py

feat(integrity): image-detail payload includes integrity_status

2026-05-19 17:53:54 -04:00

test_gs_ingest.py

fix(fc5): ruff I001 — consolidate unaliased imports + tests/test_gs_ingest credential_type assertion

2026-05-22 13:29:09 -04:00

test_health.py

fix(fc2a): three unit-test failures — async fixture decorator, name convention, lazy celery includes

2026-05-14 15:40:47 -04:00

test_html_sanitize.py

feat(provenance): nh3 HTML sanitizer for scraped post descriptions

2026-05-18 18:05:45 -04:00

test_import_file_attached.py

feat(attachments): scan enumerates all; import_file maps 'attached' + member ML

2026-05-19 11:14:49 -04:00

test_importer_archive.py

fix(test): archive members need structured imgs+threshold0; provenance payload gained attachments key

2026-05-19 11:22:17 -04:00

test_importer_artist_id.py

feat(provenance): importer sets image_record.artist_id, stops minting artist tags

2026-05-18 21:33:55 -04:00

test_importer_attach_in_place.py

fix(fc3c): origin='downloaded' (existing enum value), not invented 'imported_download'

2026-05-20 22:01:26 -04:00

test_importer_attachments.py

feat(attachments): importer dispatch — archive extract + non-media capture

2026-05-19 11:13:36 -04:00

test_importer_deep.py

feat(deep-scan): Importer(deep=) re-derive on sha-match; import_file wires batch.scan_mode

2026-05-19 15:32:26 -04:00

test_importer.py

feat(attachments): importer dispatch — archive extract + non-media capture

2026-05-19 11:13:36 -04:00

test_ir_ingest.py

fc5: ir_ingest — Tag rows (skip artist/post) + manifest JSON for sha256-keyed tag_apply

2026-05-22 09:00:29 -04:00

test_maintenance.py

fix(maintenance): recover_interrupted_tasks also sweeps pending/queued orphans (>30 min) to failed

2026-05-25 02:00:29 -04:00

test_migration_0002.py

feat(fc2a): schema migration 0002 — tag kinds + fandom hierarchy + import lifecycle

2026-05-14 12:03:41 -04:00

test_migration_0003.py

feat(fc2b): schema migration 0003 — ML pipeline tables

2026-05-15 07:35:22 -04:00

test_migration_0004.py

feat(fc2c-i): migration 0004 enables tsm_system_rows

2026-05-15 15:48:29 -04:00

test_migration_0007.py

feat(post): add nullable description + attachment_count (migration 0007)

2026-05-18 09:38:43 -04:00

test_migration_0008.py

feat(provenance): image_record.artist_id + 0008 backfill/drop-artist-tags

2026-05-18 21:33:11 -04:00

test_migration_0009.py

feat(attachments): PostAttachment model + 0009 (table + batch counter)

2026-05-19 11:10:53 -04:00

test_migration_0010.py

feat(fc3a): migration 0010 — unique(source.artist_id, platform, url)

2026-05-20 12:53:34 -04:00

test_migration_0011.py

feat(fc3b): migration 0011 — credential schema aligned with GS wire fields

2026-05-20 18:32:52 -04:00

test_migration_0012.py

feat(fc3b): migration 0012 + AppSetting model

2026-05-20 18:33:14 -04:00

test_migration_0013.py

feat(fc3c): migration 0013 — download_event.metadata + import_settings downloader fields

2026-05-20 20:36:19 -04:00

test_migration_verify.py

fc5: ml_queue (fire tag_and_embed for unprocessed) + verify (row counts + sha256 sample)

2026-05-22 09:01:46 -04:00

test_ml_aliases.py

feat(fc2b): add AliasService — (name, category) -> canonical tag

2026-05-15 07:37:13 -04:00

test_ml_allowlist.py

feat(fc2b): add AllowlistService + TagService.rename

2026-05-15 07:39:33 -04:00

test_ml_artist_retired.py

feat(provenance): ML stops surfacing the artist category

2026-05-18 21:36:43 -04:00

test_ml_centroids.py

feat(fc2b): add CentroidService — per-tag SigLIP centroids + similarity

2026-05-15 07:37:53 -04:00

test_ml_embedder.py

feat(fc2b): add SigLIP embedder wrapper

2026-05-15 07:36:17 -04:00

test_ml_suggestions.py

feat(fc2b): add SuggestionService — alias-resolved, threshold-filtered, ranked

2026-05-15 07:38:33 -04:00

test_ml_tagger.py

fix(ml): align tagger + downloader with Camie v2 actual layout (model.onnx -> camie-tagger-v2.onnx + JSON metadata + ImageNet preprocessing + sigmoid on refined output)

2026-05-25 02:25:30 -04:00

test_models.py

feat: define unified schema (Artist, Source, Post, ImageRecord, etc.) and initial migration

2026-05-14 07:34:29 -04:00

test_paths.py

fix(fc2a): apply ruff autofix + skip vue-tsc check (no tsconfig)

2026-05-14 14:24:25 -04:00

test_patreon_resolver.py

feat(fc3c): PatreonResolver — vanity → campaign_id via requests

2026-05-20 20:39:50 -04:00

test_phash_dedup.py

fix(test): threshold test needs structured images (solid colors phash-collapse)

2026-05-17 22:25:48 -04:00

test_phash_util.py

fix(test): phash 'far' case needs structured images — solid colors phash-collapse

2026-05-17 22:18:07 -04:00

test_platforms_registry.py

feat(fc3b): platforms.py — GS-6 registry (single source of truth)

2026-05-20 18:34:02 -04:00

test_post_feed_service.py

fc3e: PostFeedService — cursor-paginated unified posts feed mirroring GalleryService.scroll

2026-05-21 18:54:49 -04:00

test_provenance_service.py

fix(test): archive members need structured imgs+threshold0; provenance payload gained attachments key

2026-05-19 11:22:17 -04:00

test_scan_enumerates_all.py

feat(attachments): scan enumerates all; import_file maps 'attached' + member ML

2026-05-19 11:14:49 -04:00

test_scheduler_service.py

fc3d: scheduler_service — pure logic for due-source selection + backoff

2026-05-21 15:59:35 -04:00

test_series_model.py

feat(series): SeriesPage model + 0005 migration (unique image, ordered pages)

2026-05-16 23:33:12 -04:00

test_series_service.py

feat(series): list_pages returns full image_url for the reader

2026-05-17 16:28:20 -04:00

test_showcase_service.py

feat(fc2c-i): showcase random-sample service and endpoint

2026-05-15 15:49:25 -04:00

test_sidecar_import.py

fix(test): sidecar idempotency needs phash_threshold=0 (orthogonal splits collapse at default)

2026-05-18 15:13:48 -04:00

test_sidecar_util.py

feat(sidecar): minimal gallery-dl sidecar find + parse util

2026-05-18 14:44:37 -04:00

test_slug.py

feat(fc2a): add slug and paths utilities

2026-05-14 12:04:05 -04:00

test_source_service.py

fix(fc3b): replace FC-3a KNOWN_PLATFORMS w/ registry; drop /api/sources/platforms

2026-05-20 18:37:17 -04:00

test_suggestions_bulk.py

feat(bulk): SuggestionService.for_selection consensus + POST /api/suggestions/bulk

2026-05-16 18:03:16 -04:00

test_tag_apply.py

fc-3g-ext: IR Post/Provenance restore (tag_apply phase 4) + modal artist fallback

2026-05-24 14:08:18 -04:00

test_tag_directory_service.py

feat(fc2c-i): tag directory service and /api/tags/directory

2026-05-15 15:50:41 -04:00

test_tag_merge.py

feat(series): TagService.merge repoints series_page.series_tag_id

2026-05-16 23:36:04 -04:00

test_tag_rename.py

feat(fc2b): add AllowlistService + TagService.rename

2026-05-15 07:39:33 -04:00

test_tag_service.py

fix(fc2a): align CI with FabledRulebook — lint + short unit tests only

2026-05-14 15:23:07 -04:00

test_task_run_signals.py

fix(fc3i): single-line celery.signals import + INT32 bounds on target_id + dict.fromkeys + rewrite retry test as direct _finalize unit

2026-05-24 21:22:42 -04:00

test_tasks_admin.py

fix(fc3k): add origin=imported_filesystem to test ImageRecord ctors (second NOT NULL column after mime)

2026-05-25 01:28:54 -04:00

test_tasks_backup.py

fix(fc3h): split semicolon-stacked statements (E702) and bridge v-dialog v-model to avoid prop write

2026-05-24 23:37:06 -04:00

test_tasks_ml.py

feat(fc2b): add allowlist-apply + centroid recompute tasks + beat

2026-05-15 07:43:39 -04:00

test_tasks_register.py

fix(fc2a): three unit-test failures — async fixture decorator, name convention, lazy celery includes

2026-05-14 15:40:47 -04:00

test_text_utils.py

fc3e: html_to_plain + truncate_at_word utilities (stdlib html.parser, no new deps)

2026-05-21 18:53:35 -04:00

test_thumbnailer.py

feat(fc2a): add Thumbnailer service

2026-05-14 12:05:19 -04:00

test_tick_due_sources.py

fc3d: tick_due_sources Celery task on the scan queue

2026-05-21 16:01:59 -04:00

test_verify_integrity.py

fix(test): verify_integrity sha-mismatch row needs a distinct path (image_record.path is unique)

2026-05-19 18:15:59 -04:00