.gitignore

node_modules/
wiktionary_pron/test-results/
wiktionary_pron/playwright-report/
wiktionary_pron/coverage/
__pycache__/
*.pyc
# 30MB of Lewis & Short JSON + WORDS data — build input for build_glosses.cjs, not shipped
wiktionary_pron/utils/ext_tmp/
# gloss-extraction scratch probes + audit samples (reference implementation is
# utils/build_glosses.cjs, which is tracked)
wiktionary_pron/_probe*.cjs
wiktionary_pron/_audit_*.cjs
wiktionary_pron/_audit_sample*.tsv
wiktionary_pron/_audit_*.json
wiktionary_pron/_audit_*.tsv
# derived from macrons.txt — rebuilt by `node utils/test_gloss_regression.cjs --rebuild-cache`
wiktionary_pron/utils/gloss_lemma_cache.json.gz
# gloss-build scratch probes (reference implementation is utils/build_glosses.cjs, tracked)
wiktionary_pron/utils/_*.json
wiktionary_pron/utils/_*.cjs
# session scratch — probes, dumps, audit reports (moved out of utils/ per user)
wiktionary_pron/tmp/