5e07978171
xref.py: parser fixes (space-initial authors, hyphen-split surnames, multi-author/multi-work splits, 07-08 cards in pool, min-set scoring)
...
- AUTHOR_LINE: 'Kalsched D.' (space-initial) now recognized as entry start
- hyphen-split surname across lines: 'Schwartz-' + 'Salant, N.'
- multi-author entries split across lines (Hersh/Caligor/Yeomans)
- multi-work author blocks now split per book (Jacoby sec08: Shame is separate)
- card_entries: glob 0[1-9] (was 0[1-6]) — sec07/08 cards now matchable
(fixed 5 matcher misses: Bowlby I/II, Winnicott, Jacoby, Knox)
- score_pair: min-set overlap ratio + eps by inter size (subset tie-break: Vol.I vs Vol.II)
- curly quotes excluded from new-book split trigger
sec09 table: 19 auto-xref (18 true + 1 fragment), 19 NEW
2026-09-27 11:43:53 +03:00
285b485377
sec08 stage 0: RAW normalized (multi-author fixes), xref parser patch (4 authors), INDEX (46 items: 41 NEW + 5 xref)
...
- RAW: Mahler/Pine/Stone 1975 vs Stone/Church 1968 (2 separate books), Grossmann×2/Waters
joined; xref.py AUTHOR_LINE now supports up to 4 author groups + continuation-year lines
- xref map: 51 raw entries → 46 unique books (5 B-list duplicates merged); 5 xrefs
(CW8→sec01#03, CW9/I→sec01#04, CW17→sec02#09, Wilkinson→sec01#55, Children's Dreams→sec02#11)
- largest section so far (46 vs 28 sec06)
2026-09-26 16:23:05 +03:00
b47ff7234c
sec07 stage 0: normalized RAW (parseable by xref), verbatim to data/isap-raw, INDEX.md 27 items
...
- RAW re-scraped from official ISAP PDF (2025 ed., 2pp): multi-author items combined
(Solms/Panksepp, Alcaro/Carta, Panksepp/Biven), same-author repeats for separate works
(Kast x3, Krieger x3)
- xref.py AUTHOR_LINE: initials blob now handles 'C. A.'-style multi-initials
(08-12: +1 recovered item each)
- MASTER-LIST sec07: 24 NEW (16 books + 8 articles) + 3 xref (CW8->sec01#03,
Jacobi->sec01#20, Kast Dynamics->sec04#11)
2026-09-26 11:45:14 +03:00
7d3794560e
cross-section dedup system (user tasks 2+3, 2026-09-25):
...
- tools/xref.py: master book list for ALL 12 sections (RAW 07-12 scraped: sections/07/RAW.md
+ data/isap-raw/), cross-ref matcher (CW gate + distinctive tokens + years + Cyrillic
transliteration), stale MANIFEST 'Not downloadable' checker; data/MASTER-LIST.md generated
(17 book-dupes in 01-06 found, e.g. sec01#51=sec06#14 Solomon)
- file-level dedup: 17 md5-identical files removed from later sections (kept in first),
MANIFESTs/cards -> cross-ref rows, totals updated (sec01 102 / 02 37 / 03 125 / 04 38 /
05 38 / 06 37), dedup rule in AGENTS.md + MANIFEST Notes
- sec05 MANIFEST phantom line '03 Aion' removed (Aion not in sec05 list; files in sec01)
- sec03 #44 : 1968 Bodley file was #41 's book (content 'English Fairy Tales') — removed,
#44 keeps ABC-CLIO 2002
2026-09-25 23:25:05 +03:00
