Fix Linear B Wiktionary tr= extraction bug causing 5 wrong IPA values

Add Linear B (Mycenaean Greek) dataset

Add Linear A Phonotactics Validation dataset

Add phonologically-filtered cognate pairs config

Add Phono_Quality column, downgrade contested/doubtful pairs

Add 573K expert cognate pairs from 3 CLDF repos (IE-CoR, Robbeets, Savelyev)

Add docs/changelog/ with retroactive dataset provenance logs

Add Parquet files and YAML dataset card for HuggingFace datasets integration

Track XML/JSON/ZIP with LFS for HuggingFace size limits

Add raw scraped sources, staging intermediates, and missing audit trail

Track PDF files with LFS

Add complete dataset: all sources, metadata, scripts, docs, and phylo enrichment

Fix LFS tracking + add phylogenetic relationship metadata

Update inherited cognate pairs with ACD (22.9M rows)

Update DATABASE_REFERENCE with ACD stats

Add ACD cognate extraction script

Update cognate_pairs_inherited.tsv (21M expert pairs, P0-P2 provenance fixes)

Update cognate pair scripts + borrowing/similarity data (P0-P2 fixes)

Add cognate pairs v2 (21.5M pairs) + Phase 8 audit fixes

Phase 8: Add 24 new ancient/proto-language lexicons (12,911 entries) + scripts

initial commit