poprawki do parsowania autora

VERIFICATION.md

@@ -0,0 +1,112 @@
+# Verification Workflow
+
+Use `data/verified_author_overrides.tsv` for manual metadata corrections.
+
+## Using `generate_abs_mock_report.py`
+
+The script generates a non-destructive TSV report with proposed Audiobookshelf paths.
+It does not rename or move files.
+
+What it does:
+- scans the audiobook library tree
+- detects audiobook roots based on audio files
+- tries to infer author, title, series, sequence, year, and narrator from folder names and sidecar OPF files
+- applies manual corrections from `data/verified_author_overrides.tsv`
+- writes a TSV report with proposed target paths for Audiobookshelf
+
+What it does not do:
+- does not rename files
+- does not move directories
+- does not modify the library itself
+
+Basic usage:
+
+```bash
+python3 generate_abs_mock_report.py
+```
+
+Default behavior:
+- reads the library from `/mnt/nextcloudExtDS/Ksiazki/Audiobooki`
+- writes the report to `reports/audiobookshelf_mock_report.tsv`
+- applies manual corrections from `data/verified_author_overrides.tsv`
+
+Available options:
+
+```bash
+python3 generate_abs_mock_report.py --help
+```
+
+```text
+--root ROOT           Path to the current audiobook library
+--output OUTPUT       TSV output path
+--overrides OVERRIDES Optional TSV with verified metadata overrides
+```
+
+Examples:
+
+```bash
+python3 generate_abs_mock_report.py \
+  --root /mnt/nextcloudExtDS/Ksiazki/Audiobooki
+```
+
+```bash
+python3 generate_abs_mock_report.py \
+  --root /path/to/library \
+  --output reports/custom_report.tsv \
+  --overrides data/verified_author_overrides.tsv
+```
+
+Typical workflow:
+1. Run `python3 generate_abs_mock_report.py`.
+2. Open `reports/audiobookshelf_mock_report.tsv`.
+3. Review rows with `status=review` first, then ambiguous `unverified` rows.
+4. Add confirmed metadata to `data/verified_author_overrides.tsv`.
+5. Run the script again to regenerate the report with overrides applied.
+
+What the script prints after completion:
+- `library_root` used for the scan
+- `report` path to the generated TSV
+- `books` number of detected audiobook roots
+- `ready` rows with enough metadata to propose a target path
+- `review` rows that still need manual verification
+
+Main output file:
+- `reports/audiobookshelf_mock_report.tsv`
+
+Important columns in the TSV:
+- `status`
+- `current_path`
+- `author`
+- `series`
+- `sequence`
+- `title`
+- `proposed_abs_path`
+- `notes`
+- `verification_status`
+- `verification_source`
+
+How to read the main status fields:
+- `status=ready` means the row has enough metadata to build a proposed target path.
+- `status=review` means the row still needs manual verification.
+- `verification_status=unverified` means no manual override was applied yet.
+- `verification_status=verified_web` means the row was corrected or confirmed from a web source stored in `verification_source`.
+
+Notes about paths:
+- `current_path` is the detected source folder in the current library.
+- `proposed_abs_path` is the suggested logical Audiobookshelf path relative to the author/series/title structure.
+- The script creates the parent directory for the output TSV automatically if it does not exist.
+
+Source preference:
+- Prefer a direct audiobook/store/catalog page when it clearly confirms the metadata.
+- `lubimyczytac.pl` is an approved auxiliary source for verifying author, title, and series/cycle names.
+- Use `lubimyczytac.pl` especially when path-derived guesses are ambiguous or when storefront metadata is incomplete.
+
+Recommended fields to confirm:
+- author
+- title
+- series
+- sequence
+
+When adding an override:
+- Put the confirming page URL in `verification_source`.
+- Keep the note in `verification_note` short and only add it when it explains a correction or ambiguity.