photoanalyzer/legacy_cli_archive/donor_ledger.yaml

# Donor ledger — US01-01 (INTEGRATED_PIPELINE_CONCEPT.md §3 "Donor-first CLI migration")
#
# Maps every relevant legacy module/function to its classification, rationale,
# target module in the future photo_pipeline package, and characterization
# test IDs. Linted by tests/characterization/test_donor_ledger.py.
#
# classification: reuse    — take as-is
#                 extract  — move behind a thin adapter, behavior unchanged
#                 refactor — extract while changing structure, behavior preserved
#                            unless an intentional delta is stated
#                 replace  — new implementation; rationale documents why and what
#                            (if anything) carries over
# Every row needs either `tests` (existing test IDs, module::function) or
# `pending_story` (the backlog story that will characterize/deliver it).
# status: characterized — donor behavior pinned by characterization tests
#         resolved      — replacement shipped; `parity` names the tests that prove
#                         it, and `delta` states every intentional difference
#         pending       — not migrated yet; `pending_story` says which story will
# parity: test ids (path::function) in any suite, proving the replacement
# delta:  what the replacement deliberately does differently, or not at all
#
# Archived by US07-01: the sources referenced below now live beside this file in
# src/ and are frozen (see README.md). Nothing in photo_pipeline imports them.

rows:
  # ── photo_analyzer.py ──────────────────────────────────────────────────────
  - id: pa-discovery
    area: discovery
    source: {file: photo_analyzer.py, symbols: [discover_photos, SUPPORTED_EXTENSIONS]}
    classification: extract
    rationale: >
      Proven recursive discovery with _IGNORE/ and .@__thumb exclusion, sorted +
      deduplicated. Becomes the one shared discovery used by every stage; the
      exclusion predicate moves to path_policy.py.
    target: photo_pipeline/services/inventory.py + photo_pipeline/path_policy.py
    tests:
      - test_pa_discovery::test_discover_excludes_ignore_and_thumbs
      - test_pa_discovery::test_discover_finds_supported_including_uppercase
      - test_pa_discovery::test_discover_sorted_and_deduplicated
      - test_pa_discovery::test_supported_extensions_contract
    status: characterized

  - id: pa-purge-excluded
    area: discovery
    source: {file: photo_analyzer.py, symbols: [purge_excluded]}
    classification: refactor
    rationale: >
      SQL LIKE patterns duplicate the path policy; refactor to call the shared
      path_policy predicate so exclusion has exactly one definition.
    target: photo_pipeline/services/inventory.py
    tests: [test_pa_db::test_purge_excluded_removes_ignore_and_thumb_rows]
    status: characterized

  - id: pa-prune-missing
    area: database
    source: {file: photo_analyzer.py, symbols: [prune_missing]}
    classification: replace
    rationale: >
      Deletes rows for missing files — incompatible with stable asset identity;
      the concept keeps the asset and sets missing_at/availability_state.
      The unmounted-library guard (never prune when the root is absent) carries
      over as behavior.
    target: photo_pipeline/services/inventory.py
    tests: [test_pa_db::test_prune_missing_guards_unmounted_library]
    status: characterized

  - id: pa-variants
    area: discovery
    source: {file: photo_analyzer.py, symbols: [_strip_variant_markers, _is_variant, _variant_key, build_variant_groups]}
    classification: extract
    rationale: >
      Filename-based variant grouping (pixel-size suffix, 'bearbeitet' edit
      marker) with unmarked-largest primary selection; feeds duplicate/variant
      clustering evidence.
    target: photo_pipeline/services/duplicates.py
    tests:
      - test_pa_variants::test_strip_variant_markers_goldens
      - test_pa_variants::test_is_variant
      - test_pa_variants::test_variant_key_folder_scoped_case_insensitive
      - test_pa_variants::test_build_variant_groups_primary_is_largest_unmarked
    status: characterized

  - id: pa-variant-propagate
    area: database
    source: {file: photo_analyzer.py, symbols: [propagate_variants, _row_to_result]}
    classification: refactor
    rationale: >
      Copying a primary's analysis into variant rows survives, but keyed by
      asset_id and recorded as stage state instead of raw row copies.
    target: photo_pipeline/services/duplicates.py
    parity:
      - tests/integration/test_duplicate_engine.py::test_exact_copies_form_auto_decided_cluster
      - tests/integration/test_duplicate_engine.py::test_perceptual_variant_is_review_only
    delta: >
      The donor propagated variant links implicitly while writing rows; the replacement makes the canonical link a reviewable, reversible cluster decision, so a propagated link can always be undone.
    status: resolved

  - id: pa-hashing
    area: hashing
    source: {file: photo_analyzer.py, symbols: [_sha1_file, _phash_image]}
    classification: extract
    rationale: >
      Streamed SHA-1 plus 64-bit DCT phash (imagehash.phash recipe on
      numpy/scipy, no extra dependency). Hash outputs are goldens — a change
      means the decoder/algorithm changed and hash_version must bump.
    target: photo_pipeline/services/inventory.py
    tests:
      - test_pa_hashing_dedup::test_phash_goldens
      - test_pa_hashing_dedup::test_phash_survives_resize_and_recompress
      - test_pa_hashing_dedup::test_phash_format_16_hex_chars
      - test_pa_hashing_dedup::test_sha1_matches_hashlib
      - test_pa_hashing_dedup::test_sha1_unreadable_returns_none
    status: characterized

  - id: pa-ensure-hashes
    area: hashing
    source: {file: photo_analyzer.py, symbols: [ensure_hashes]}
    classification: extract
    rationale: >
      Resume-safe incremental hashing (skip already-hashed, COALESCE so a failed
      hash never clobbers a stored one, periodic commits, honors stop event).
    target: photo_pipeline/services/inventory.py
    tests: [test_pa_hashing_dedup::test_ensure_hashes_skips_hashed_and_is_resume_safe]
    status: characterized

  - id: pa-cluster
    area: hashing
    source: {file: photo_analyzer.py, symbols: [cluster_duplicates]}
    classification: extract
    rationale: >
      Union-find over the vectorized Hamming graph; largest-first ordering.
      Known ceiling (documented in-source): O(n²) scan, BK-tree past ~100k.
    target: photo_pipeline/services/duplicates.py
    tests: [test_pa_hashing_dedup::test_cluster_and_mark_duplicates_largest_is_canonical]
    status: characterized

  - id: pa-mark-duplicates
    area: database
    source: {file: photo_analyzer.py, symbols: [mark_duplicates]}
    classification: refactor
    rationale: >
      Intentional delta — the concept forbids silently marking fuzzy matches:
      largest-file canonical becomes a *recommendation*; the decision lands in
      reviewable duplicate_clusters, auto-resolve only for exact matches.
      The status='duplicate' exclusion from analysis/upload carries over.
    target: photo_pipeline/services/duplicates.py
    tests: [test_pa_hashing_dedup::test_cluster_and_mark_duplicates_largest_is_canonical]
    status: characterized

  - id: pa-list-duplicates
    area: ui
    source: {file: photo_analyzer.py, symbols: [list_duplicates]}
    classification: replace
    rationale: Console report; superseded by the duplicate-review API/UI (US01-06).
    target: photo_pipeline/api/routes + frontend duplicate review
    parity:
      - tests/integration/test_review_api.py::test_clusters_list_and_detail
      - tests/e2e/test_review_ui.py::test_fuzzy_decision_requires_confirmation
    delta: >
      The text listing became the paged cluster API and the comparison UI; no textual report is produced.
    status: resolved

  - id: pa-reconcile
    area: database
    source: {file: photo_analyzer.py, symbols: [reconcile_moved]}
    classification: refactor
    rationale: >
      SHA-1-keyed move/rename reconciliation preserving analysis + EXIF state.
      Refactor to update assets.current_path by asset_id and append an
      asset_paths history row instead of rewriting the path in place.
    target: photo_pipeline/services/inventory.py
    tests: [test_pa_hashing_dedup::test_reconcile_moved_preserves_row_by_sha1]
    status: characterized

  - id: pa-db-schema
    area: database
    source: {file: photo_analyzer.py, symbols: [SCHEMA, get_db, _migrate_schema]}
    classification: refactor
    rationale: >
      photos table + FTS5 + idempotent ALTER migration become the migration
      baseline; new schema adds assets/asset_paths/stage state per concept.
      Existing data must migrate losslessly (US01-02).
    target: photo_pipeline migrations (Alembic)
    tests: [test_pa_db::test_schema_and_migration_idempotent]
    status: characterized

  - id: pa-db-status
    area: error
    source: {file: photo_analyzer.py, symbols: [upsert_pending, mark_analyzed, mark_error, mark_exif_written, get_pending, get_analyzed_no_exif]}
    classification: refactor
    rationale: >
      Status lifecycle (pending→analyzed→exif_written; error rows auto-retried
      on the next run; INSERT OR IGNORE never downgrades; duplicates never
      queued) is the resume-safety contract — preserved as asset_stage_states
      transitions keyed by asset_id.
    target: photo_pipeline/repositories + services/analysis.py
    tests:
      - test_pa_db::test_status_lifecycle_pending_analyzed_exif_written
      - test_pa_db::test_mark_error_and_retry_via_get_pending
      - test_pa_db::test_mark_analyzed_clears_error
      - test_pa_db::test_upsert_pending_never_downgrades
    status: characterized

  - id: pa-fts
    area: database
    source: {file: photo_analyzer.py, symbols: [photos_fts]}
    classification: reuse
    rationale: FTS5 table + sync triggers already work; carried into the new schema.
    target: photo_pipeline migrations (Alembic)
    tests: [test_pa_db::test_fts_kept_in_sync_by_triggers]
    status: characterized

  - id: pa-imaging
    area: imaging
    source: {file: photo_analyzer.py, symbols: [prepare_image, MAX_LONG_EDGE]}
    classification: extract
    rationale: >
      RGB-normalize (drops alpha, converts HEIC), LANCZOS resize to 2048px
      long-edge, JPEG q85 base64 — the provider-input contract. Truncated-image
      tolerance (ImageFile.LOAD_TRUNCATED_IMAGES) carries with it.
    target: photo_pipeline/services/analysis.py
    tests:
      - test_pa_imaging::test_prepare_image_small_passthrough_jpeg
      - test_pa_imaging::test_prepare_image_resizes_to_max_long_edge
      - test_pa_imaging::test_prepare_image_converts_png_alpha_to_rgb_jpeg
    status: characterized

  - id: pa-repair
    area: error
    source: {file: photo_analyzer.py, symbols: [repair_image, _REPAIRABLE_ERRORS, _log_repair]}
    classification: refactor
    rationale: >
      Re-encode structurally broken images via sips (macOS-only — portability
      note), .orig backup, atomic replace, repairs.jsonl audit. Needs corrupt
      fixtures to characterize; lands with media hardening.
    target: photo_pipeline/integrations/exiftool.py
    pending_story: US07-03
    status: pending

  - id: pa-vision
    area: vision
    source: {file: photo_analyzer.py, symbols: [analyze_image, ANALYSIS_PROMPT, LLM_BASE_URL, LLM_MODEL]}
    classification: extract
    rationale: >
      OpenAI-compatible request construction, base64 image_url payload, JSON
      response validation, 429/503 retry with exponential backoff. Prompt and
      model/config version must be persisted per analysis_runs. Characterized
      against a deterministic fake provider when the analysis service is ported.
    target: photo_pipeline/services/analysis.py
    parity:
      - tests/integration/test_safety_analysis.py::test_provider_called_only_for_confirmed_sfw
    delta: >
      The prompt and response schema carry over; the provider is an injected adapter so the privacy gate is testable, and results are keyed to asset ids rather than paths.
    status: resolved

  - id: pa-throttle
    area: logging
    source: {file: photo_analyzer.py, symbols: [_record_ratelimit, _ratelimit_recent, write_throttle_summary, _rpd_load, _rpd_increment, _rpd_count]}
    classification: refactor
    rationale: >
      Rolling throttle window + persistent throttle_events.jsonl + RPD day
      counter become job metrics/events on the durable job model.
    target: photo_pipeline/services/jobs.py + photo_pipeline/services/analysis.py
    pending_story: US07-04
    status: pending

  - id: pa-nsfw-filter
    area: nsfw
    source: {file: photo_analyzer.py, symbols: [_rec_has_nsfw, filter_nsfw_tagged]}
    classification: replace
    rationale: >
      Intentional delta — donor fails OPEN (exiftool hiccup → analyze
      everything), which violates the concept privacy invariant; the new gate
      is the DB safety decision and fails CLOSED (only confirmed sfw reaches
      the provider). EXIF keyword reading survives as evidence/verification,
      not as the gate. Current behavior captured for migration.
    target: photo_pipeline/services/safety.py
    tests:
      - test_pa_exif::test_rec_has_nsfw_list_scalar_case
      - test_pa_exif::test_filter_nsfw_tagged_splits_and_normalizes
      - test_pa_exif::test_filter_nsfw_tagged_empty_list_noop
    status: characterized

  - id: pa-exif-caption
    area: exif
    source: {file: photo_analyzer.py, symbols: [build_exif_caption, build_exif_caption_from_result]}
    classification: refactor
    rationale: >
      Caption format ('desc | Tags: … | Mood: … | Location: … | ~year') is
      what 25k+ photos already carry — preserved for compatibility. Intentional
      delta: new writes go into a managed 'AI:' segment so re-analysis can
      replace only its own text (concept EXIF ownership).
    target: photo_pipeline/services/exif projection
    tests:
      - test_pa_exif::test_caption_golden_full
      - test_pa_exif::test_caption_golden_sparse
    status: characterized

  - id: pa-exif-read
    area: exif
    source: {file: photo_analyzer.py, symbols: [read_existing_exif]}
    classification: extract
    rationale: Targeted exiftool JSON read of the four owned fields, 30s timeout, fail-empty.
    target: photo_pipeline/integrations/exiftool.py
    tests: [test_pa_exif::test_write_exif_roundtrip_and_idempotent]
    status: characterized

  - id: pa-exif-write
    area: exif
    source: {file: photo_analyzer.py, symbols: [write_exif]}
    classification: refactor
    rationale: >
      Merge-don't-overwrite semantics are the crown jewels: caption prepended to
      existing text (idempotent — never appended twice), keywords merged +
      order-preserving dedup, -m -overwrite_original, repair-and-retry on
      structurally broken files. Refactored to add the concept's read-back
      verification + non-owned-field snapshot comparison. sys.exit on missing
      exiftool becomes a service error.
    target: photo_pipeline/integrations/exiftool.py + services/exif checkpoints
    tests:
      - test_pa_exif::test_write_exif_roundtrip_and_idempotent
      - test_pa_exif::test_write_exif_preserves_existing_metadata
    status: characterized

  - id: pa-run-loop
    area: cancellation
    source: {file: photo_analyzer.py, symbols: [run_analysis, _run_folder_loop, run_exif_write, _stop, _handle_sigint]}
    classification: refactor
    rationale: >
      Folder-grouped worker pool, per-item DB commits, cooperative stop event
      checked between items, double-SIGINT force quit — becomes the durable
      JobRunner worker loop with the same drain-and-resume semantics.
    target: photo_pipeline/jobs/worker.py
    parity:
      - tests/integration/test_worker.py::test_worker_processes_all_items
      - tests/integration/test_worker.py::test_cooperative_cancellation_leaves_items_resumable
    delta: >
      The in-process folder loop with SIGINT handling became durable jobs claimed by a worker: cancellation is a persisted request, not a signal, and an interrupted run resumes from the database instead of restarting.
    status: resolved

  - id: pa-ui-terminal
    area: ui
    source: {file: photo_analyzer.py, symbols: [_Dashboard, _LiveRenderer, _PlainRenderer, _make_progress, _stats_panel, _feed_panel, _tokens_panel, _menu_panel, _read_key, _key_listener, print_stats, fit_label]}
    classification: replace
    rationale: >
      Rich terminal dashboard (Live layout, hotkey menu, termios key reader) is
      superseded by the web Workflow/Analyze views. The *data* it renders
      (album progress, feed, tokens, ETA) survives via webapp.runner.progress
      (see wa-runner). fit_label/print_stats die with it.
    target: frontend Analyze view (Phase B)
    tests: [test_pa_album::test_fit_label_left_truncates]
    status: characterized

  - id: pa-album-label
    area: ui
    source: {file: photo_analyzer.py, symbols: [album_label]}
    classification: extract
    rationale: >
      Album = leaf folder relative to library, '(root)' at root, bare parent
      outside the library. The single album-identity rule for stats, proposals
      and folder-as-album upload. webapp.query.album_of is a copy — consolidate
      to one implementation (see wa-album-of). Extracted in US03-01 as
      photo_pipeline.services.albums.album_label (parity test below); the album
      evidence aggregator uses it. LibraryService._album_of still keeps its own
      leaf-name grouping — fold it into this rule when the Library view is next
      touched.
    target: photo_pipeline/services/albums.py
    tests:
      - test_pa_album::test_album_label_leaf_folder_relative
      - test_pa_album::test_webapp_album_of_mirrors_album_label
      - test_pa_album::test_albums_service_album_label_matches_donor
    status: characterized

  - id: pa-config
    area: configuration
    source: {file: photo_analyzer.py, symbols: [load_env_file, _env_str, _env_int, _env_bool, ENV_FILE]}
    classification: refactor
    rationale: >
      photo_analyzer.env parsing (shell env wins, export prefix, quoted values,
      inline-comment rules) feeds the typed config module; same file keeps
      working so the transition needs no re-setup.
    target: photo_pipeline/config.py
    tests:
      - test_pa_config::test_load_env_file_parsing
      - test_pa_config::test_env_helpers
    status: characterized

  - id: pa-logging
    area: logging
    source: {file: photo_analyzer.py, symbols: [log_history, _history_handler, _rich_handler]}
    classification: replace
    rationale: >
      Module-level triple-handler logging (console + info log + debug log) and
      the JSONL history logger become structured JSON logging with job_id/
      asset_id and job_events rows; per-photo history maps to job events.
    target: photo_pipeline structured logging + jobs/job_events
    parity:
      - tests/integration/test_jobs.py::test_enqueue_persists_items_and_event
      - tests/integration/test_jobs_sse.py::test_sse_streams_all_events_then_closes
    delta: >
      The JSONL history file and rich console handler are replaced by structured JSON logs plus durable job_events; the browser reads events over SSE rather than tailing a file.
    status: resolved

  - id: pa-balance
    area: vision
    source: {file: photo_analyzer.py, symbols: [fetch_balance, query_balance, check_quota]}
    classification: extract
    rationale: >
      Provider balance/quota probes (report 'unsupported' on providers without
      the endpoint). Network-bound; characterized against the fake provider.
    target: photo_pipeline/services/analysis.py
    parity:
      - tests/integration/test_safety_analysis.py::test_provider_called_only_for_confirmed_sfw
    delta: >
      Not carried over: balance/quota polling was provider-specific (Gemini/OpenAI billing endpoints) and key-scoped. Cost reporting, when a story asks for it, comes from the per-run usage recorded with each analysis result rather than from a vendor endpoint.
    status: resolved

  - id: pa-cli
    area: configuration
    source: {file: photo_analyzer.py, symbols: [main]}
    classification: replace
    rationale: >
      argparse surface is superseded by the API; flags map to job configs
      (documented in WEBAPP_CONCEPT.md §8 parity table). Transitional CLI calls
      the shared services until archival (E07).
    target: photo_pipeline/__main__.py (serve | worker | migrate | import-legacy-scores) + /api/v1
    parity:
      - tests/integration/test_app_lifecycle.py::test_restart_preserves_data_and_reruns_migrations
      - tests/unit/test_legacy_archive.py::test_production_code_never_imports_an_archived_module
    delta: >
      The argparse surface is not reproduced. Every flag that drove work became an API command or a job configuration; the CLI keeps only what an application needs to be operated (serve, worker, migrate) plus the one-off legacy CSV import.
    status: resolved

  # ── nsfwtag/ ───────────────────────────────────────────────────────────────
  - id: nt-discovery
    area: discovery
    source: {file: nsfwtag/scoring.py, symbols: [discover_images]}
    classification: refactor
    rationale: >
      Merged into the shared discovery. Known divergence captured: nsfwtag EXTS
      lacks .tiff/.tif, is non-recursive by default, and dedupes by resolved
      path (symlink guard) — the shared service adopts the superset + symlink
      dedup, always recursive, plus _IGNORE policy (donor has none).
    target: photo_pipeline/services/inventory.py
    tests:
      - test_nsfwtag::test_exts_contract_differs_from_photo_analyzer
      - test_nsfwtag::test_discover_images_flat_recursive_limit
      - test_nsfwtag::test_discover_images_dedupes_symlinked_paths
    status: characterized

  - id: nt-score-cache
    area: nsfw
    source: {file: nsfwtag/scoring.py, symbols: [load_cache, _save_cache, CSV_NAME]}
    classification: replace
    rationale: >
      nsfw_scores.csv stops being the source of truth (concept: DB state).
      Format characterized (4-decimal scores, bad rows dropped) because the
      existing CSV must migrate into assets.safety_score.
    target: photo_pipeline/services/legacy_import.py (one-off import into safety_reviews)
    tests: [test_nsfwtag::test_score_cache_roundtrip_and_tolerance]
    parity:
      - tests/integration/test_legacy_import.py::test_scores_are_imported_onto_asset_identity
      - tests/integration/test_legacy_import.py::test_a_reviewed_asset_is_never_overwritten_by_the_csv
    delta: >
      The CSV is no longer read at runtime at all: it is imported once into
      scored-but-unreviewed safety_reviews rows and left on disk untouched. A path
      that matches nothing is reported, never turned into an asset, and a human
      decision always outranks an imported score.
    status: resolved

  - id: nt-score-model
    area: nsfw
    source: {file: nsfwtag/scoring.py, symbols: [score_images, MODEL_ID, BATCH]}
    classification: extract
    rationale: >
      On-device ViT scoring: MPS/CPU pick, model-defined input size (not
      hardcoded — survives model swap), batch inference where one bad batch
      never discards prior scores, periodic checkpoint every 500. Cache-hit
      short-circuit (model never loads when nothing is new) characterized now;
      inference path needs the local model + deterministic fake.
    target: photo_pipeline/integrations/nsfw_model.py
    tests: [test_nsfwtag::test_score_images_cache_hit_skips_model]
    delta: >
      The donor set Pillow's process-global ImageFile.LOAD_TRUNCATED_IMAGES so a
      partially downloaded file still scored. Here the same process also hashes
      files and renders previews, and those must keep refusing a truncated file
      rather than silently working on half of one; scoring opens images through
      the bounded photo_pipeline.imaging door instead and skips the ones it cannot
      read, leaving them unscored and visibly undecided (US07-03).
    status: characterized

  - id: nt-exif-keyword
    area: exif
    source: {file: nsfwtag/exif.py, symbols: [write_keyword, remove_keyword]}
    classification: extract
    rationale: >
      Idempotent add/remove via '-Keywords-=x -Keywords+=x' (no duplicates on
      re-run), touching only Keywords+Subject. Basis of the safety EXIF
      checkpoint; the mutual-exclusion write (sfw removed when nsfw added and
      vice versa) is the concept's addition on top.
    target: photo_pipeline/services/safety.py + integrations/exiftool.py
    tests: [test_nsfwtag::test_write_remove_keyword_idempotent]
    status: characterized

  - id: nt-exif-marks
    area: nsfw
    source: {file: nsfwtag/exif.py, symbols: [_read_keywords, read_tagged, read_marks]}
    classification: extract
    rationale: >
      Batched exiftool read via stdin, lowercase normalization, and the safety
      rule that a file carrying both keywords reads as nsfw (never as safe).
    target: photo_pipeline/services/safety.py
    tests:
      - test_nsfwtag::test_write_remove_keyword_idempotent
      - test_nsfwtag::test_read_marks_nsfw_wins_over_sfw
    status: characterized

  - id: nt-exif-view
    area: exif
    source: {file: nsfwtag/exif.py, symbols: [read_exif, _EXIF_NOISE]}
    classification: extract
    rationale: >
      Lightbox EXIF read: filesystem-noise filter, binary/oversize skip, signed
      decimal GPS, Google-Maps link prepended. Backs the asset EXIF endpoint.
    target: photo_pipeline/api (asset EXIF endpoint)
    tests: [test_nsfwtag::test_read_exif_filters_noise_and_prepends_map]
    status: characterized

  - id: nt-apply-list
    area: nsfw
    source: {file: nsfwtag/exif.py, symbols: [apply_list]}
    classification: replace
    rationale: >
      Newline-list bulk tagging (nsfw_confirmed.txt flow) is superseded by DB
      review decisions; the existing list is a one-time migration input.
    target: photo_pipeline/services/safety.py (decision import in US01-02)
    parity:
      - tests/integration/test_safety_parity.py::test_extracted_marks_and_partition_match_donor
    delta: >
      Bulk keyword application from a file list is replaced by decisions against asset ids; the EXIF write itself is the extracted, read-back-verified one.
    status: resolved

  - id: nt-ui
    area: ui
    source: {file: nsfwtag/server.py, symbols: [serve_review]}
    classification: replace
    rationale: >
      stdlib ThreadingHTTPServer + token-injected review.html backend is
      superseded by FastAPI. review.html's design tokens, folder tree,
      threshold/score review flow, lightbox and keyboard model are the frontend
      donor for the Safety view (preserved per concept §10; ported in US02-01).
    target: photo_pipeline/api + frontend Safety view
    parity:
      - tests/e2e/test_workflow_views.py::test_safety_review_decide_persists_across_reload
    delta: >
      The stdlib review server is replaced by the API plus the Safety view; the donor's layout, thresholds, and keyboard flow carry over, its embedded HTML generation does not.
    status: resolved

  - id: nt-bench
    area: nsfw
    source: {file: nsfwtag/bench.py, symbols: [main]}
    classification: replace
    rationale: >
      Dev-only model benchmark; archived without webapp replacement (recorded
      basis of the AdamCodd model choice). No production caller.
    target: none (archive as reference)
    parity:
      - tests/unit/test_legacy_archive.py::test_every_archived_source_matches_its_checksum
    delta: >
      No replacement: a dev-only benchmark whose result (the AdamCodd model choice) is already recorded. Kept in the archive as the basis of that choice.
    status: resolved

  # ── webapp/ ────────────────────────────────────────────────────────────────
  - id: wa-query-search
    area: database
    source: {file: webapp/query.py, symbols: [_fts_query, _where, search, _row_to_card, photo, _SORTS]}
    classification: extract
    rationale: >
      Safe FTS5 MATCH building (tokenize → quoted prefix terms; punctuation
      can't crash MATCH), facet WHERE composition, paged search with
      relevance/sort modes, raw_response kept out of list payloads.
    target: photo_pipeline/repositories (search) + api/routes
    tests:
      - test_webapp_query::test_fts_query_sanitization_goldens
      - test_webapp_query::test_where_clause_goldens
      - test_webapp_query::test_search_fts_prefix_match
      - test_webapp_query::test_search_filters_and_browse
      - test_webapp_query::test_search_paging
      - test_webapp_query::test_photo_omits_raw_response
    status: characterized

  - id: wa-query-aggregates
    area: database
    source: {file: webapp/query.py, symbols: [facets, stats, _col_counts, _people_buckets, _top_tags]}
    classification: extract
    rationale: >
      Facet counts, year range, album tree, status/album progress (done =
      analyzed + exif_written) and error listing — the Stats/Workflow data
      model.
    target: photo_pipeline/repositories + services/workflow.py
    tests:
      - test_webapp_query::test_facets_and_albums
      - test_webapp_query::test_stats_album_progress_and_errors
    status: characterized

  - id: wa-album-of
    area: ui
    source: {file: webapp/query.py, symbols: [album_of]}
    classification: replace
    rationale: >
      Verbatim copy of photo_analyzer.album_label (equivalence characterized);
      consolidated into the single albums-service implementation (pa-album-label),
      now photo_pipeline.services.albums.album_label (US03-01, parity test below).
    target: photo_pipeline/services/albums.py
    tests:
      - test_pa_album::test_webapp_album_of_mirrors_album_label
      - test_pa_album::test_albums_service_album_label_matches_donor
    status: characterized

  - id: wa-allowlist
    area: ui
    source: {file: webapp/query.py, symbols: [all_paths]}
    classification: replace
    rationale: >
      Path-set allowlist for /img and /exif is superseded by asset-ID-addressed
      endpoints that never accept a browser-supplied path (concept thumbnail
      rules). The validate-before-serving intent carries over.
    target: photo_pipeline/api (asset_id thumbnail/EXIF endpoints)
    tests: [test_webapp_query::test_all_paths_is_the_image_endpoint_allowlist]
    status: characterized

  - id: wa-runner
    area: cancellation
    source: {file: webapp/runner.py, symbols: [Runner, progress]}
    classification: replace
    rationale: >
      Subprocess-driving-the-CLI job control is superseded by durable DB jobs
      with a worker process. Two ideas carry over: progress derived from DB
      counts (not job-private state) and single-mutating-job enforcement.
    target: photo_pipeline/services/jobs.py + photo_pipeline/jobs/worker.py
    parity:
      - tests/integration/test_jobs.py::test_idempotency_key_returns_same_job
      - tests/integration/test_worker.py::test_handler_failure_fails_the_job
    delta: >
      Subprocess supervision of a CLI is replaced by durable jobs in the same process family: there is no subprocess to supervise, and progress is persisted rather than scraped from stdout.
    status: resolved

  - id: wa-server
    area: ui
    source: {file: webapp/server.py, symbols: [serve]}
    classification: replace
    rationale: >
      stdlib HTTP route ladder → FastAPI typed routes. Carried intents:
      127.0.0.1 bind, path validation against the DB, no secrets to the
      browser. analyzer.html + page.py design (dark OLED tokens, Library/
      Analyze/Stats views) is frontend donor material per concept §10.
    target: photo_pipeline/api/app.py + frontend
    parity:
      - tests/e2e/test_frontend_shell.py::test_shell_loads_assets_without_console_or_network_errors
      - tests/e2e/test_frontend_shell.py::test_deep_link_and_reload_restore_view_and_filters
    delta: >
      The stdlib HTTP routes become the versioned FastAPI API and the static frontend shell; Python no longer interpolates HTML.
    status: resolved