Benjamin Admin dac2a9f685 feat(rag): Legal-Metadaten — article_label + deterministische IDs + chunk_hash ...

Neues pures Modul legal_metadata.py (nur stdlib, lokal+CI testbar): §3-Normalisierung
section->article, strikte Header-Extraktion (Datum/Seiten-Rauschen -> kein Falsch-Zitat),
citation_style pro Regulierung (EU/CH=article, DE=paragraph), Urteil=Aktenzeichen statt §,
camelCase-Klarnamen (ProdHaftG), deterministische uuid5-Point-ID + chunk_hash (sha256).
documents.py verdrahtet build_legal_fields in den Payload-Build + document_version.
10 Tests gruen. Vertrag: rag_reingest_spec.md (§2/§3).

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

2026-06-20 14:35:07 +02:00

..

api

feat(rag): Legal-Metadaten — article_label + deterministische IDs + chunk_hash

2026-06-20 14:35:07 +02:00

tests

feat(rag): Legal-Metadaten — article_label + deterministische IDs + chunk_hash

2026-06-20 14:35:07 +02:00

config.py

…

Dockerfile

…

embedding_client.py

feat(pipeline): structural metadata end-to-end (Blocks D2-D4)

2026-05-01 20:34:00 +02:00

html_utils.py

fix(rag): HTML charset detection + opening block tag newlines

2026-05-02 08:35:47 +02:00

legal_metadata.py

feat(rag): Legal-Metadaten — article_label + deterministische IDs + chunk_hash

2026-06-20 14:35:07 +02:00

main.py

…

minio_client_wrapper.py

feat(pipeline): MC Quality Overhaul — 74.5% → 92.8% accuracy, 5.3K → 13.6K MCs

2026-05-10 15:08:15 +02:00

qdrant_client_wrapper.py

feat(pipeline): MC Quality Overhaul — 74.5% → 92.8% accuracy, 5.3K → 13.6K MCs

2026-05-10 15:08:15 +02:00

requirements.txt

…