daniel-wojahn/echr-dissent-matrix

Add accent-normalized deduplication, improved terminators, case reference filtering

Code audit: consolidated BLACKLIST, added Gephi export, ruff formatting

Switch to regex-based extraction, remove spaCy dependency

Add initials filtering, accent deduplication, and ß->ss normalization

Switch to en_core_web_trf model and remove query-specific name mappings

Sync extraction logic with run.py - improved regex fallback

Fix comma-separated names and add global partial name filtering

Add second-pass judge name normalization to fix count

Fix server binding for HF Spaces

Fix Gradio 6.0 theme deprecation warning

initial commit