Benjamin_Boenisch
  • Joined on 2026-02-07
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-compliance 2026-03-28 16:24:30 +00:00
712fa8cb74 feat: Pass 0b quality — negative actions, container detection, session object classes
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-compliance 2026-03-28 11:47:49 +00:00
447ec08509 Add migration 082: widen source_article to TEXT, fix pass0b query filters
8cb1dc1108 Fix pass0b queries to skip deprecated/duplicate controls
f8d9919b97 Improve object normalization: shorter keys, synonym expansion, qualifier stripping
Compare 3 commits »
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-lehrer 2026-03-28 09:54:52 +00:00
21b69e06be Fix cross-column word assignment by splitting OCR merge artifacts
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-lehrer 2026-03-27 16:44:19 +00:00
0168ab1a67 Remove Hauptseite/Box tabs from Kombi pipeline
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-lehrer 2026-03-27 15:47:54 +00:00
925f4356ce Use spellchecker instead of pyphen for pipe autocorrect validation
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-lehrer 2026-03-27 15:33:47 +00:00
cc4cb3bc2f Add pipe auto-correction and graphic artifact filter for grid builder
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-lehrer 2026-03-27 14:54:39 +00:00
0685fb12da Fix Bug 3: recover OCR-lost prefixes via overlap merge + chain merging
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-lehrer 2026-03-27 14:36:34 +00:00
96ea23164d Fix word-gap merge: add missing pronouns to stop words, reduce threshold
a8773d5b00 Fix 4 Grid Editor bugs: syllable modes, heading detection, word gaps
9f68bd3425 feat: Implement page-split step with auto-detection and sub-session naming
469f09d1e1 fix: Redesign StepUpload for manual step control
3bb04b25ab fix: OCR Kombi upload race condition — openSession was resetting step to 0
Compare 5 commits »
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-compliance 2026-03-27 07:38:52 +00:00
fb2cf29b34 fix: Pass 0b — Duplicate Guard, Severity-Kalibrierung, Title-Truncation
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-compliance 2026-03-26 19:13:13 +00:00
f39e5a71af feat: Obligation-Deduplizierung — 34.617 Duplikate als 'duplicate' markiert
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-compliance 2026-03-26 16:35:55 +00:00
ac42a0aaa0 fix: Faceted Counts — NULL-Werte einbeziehen + AbortController fuer Race Conditions
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-lehrer 2026-03-26 15:09:47 +00:00
85fe0a73d6 docs: Add OCR Kombi Pipeline to MkDocs and cross-reference from OCR Pipeline
eaade3cad2 feat: Maschinenbau-Branche + INDUSTRY_REGULATION_MAP erweitert
Compare 2 commits »
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-lehrer 2026-03-26 14:55:44 +00:00
d26a9f60ab Add OCR Kombi Pipeline: modular 11-step architecture with multi-page support
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-compliance 2026-03-26 14:00:50 +00:00
52e463a7c8 feat: Faceted Search — Dropdown-Counts passen sich aktiven Filtern an
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-compliance 2026-03-26 13:34:24 +00:00
2dee62fa6f feat: Eigenentwicklung-Filter im Typ-Dropdown mit Counts
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-lehrer 2026-03-26 10:21:57 +00:00
d26233b5b3 Add page number display to StepGridReview summary bar
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-compliance 2026-03-26 10:13:42 +00:00
3fb07e201f fix: V1 Enrichment Threshold auf 0.70 gesenkt (typische Top-Scores 0.70-0.77)
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-compliance 2026-03-26 09:52:52 +00:00
81c9ce5de3 fix: V1 Enrichment — Qdrant Collection + Parent-Resolution fuer regulatorische Matches
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-compliance 2026-03-26 09:32:26 +00:00
db7c207464 feat: V1 Control Enrichment — Eigenentwicklung-Label, regulatorisches Matching & Vergleichsansicht
Benjamin_Boenisch pushed to main at Benjamin_Boenisch/breakpilot-lehrer 2026-03-26 07:52:19 +00:00
e019dde01b Extract page number as metadata instead of silently removing it