This public snapshot shares the successful portion of the AKS Yoksa classical text pipeline in a GitHub-friendly structure inspired by legalize-kr.
- Source books: 439
- Source volumes: 1,325
- Source articles with translatable body text: 112,287
- Successful translations: 28,079
- Failed first-pass attempts retained locally for retry: 79,451
- Pending/running: 4,697 / 60
_index/translation_status.json: machine-readable progress and security notes_index/translations_success.jsonl.gz: successful Korean reading/translation/meaning records_index/articles_success.jsonl.gz: successful translations joined with cleaned Hanmun source and provenance URL_index/translations_success_sample.jsonl: quick review sample of successful rowskr/: path-based source tree sample with bibliography, haje, article source, and translation fields
API keys, tokens, raw HTML cache, local SQLite databases, logs, virtual environments, process files, and failed-attempt logs are excluded. Canonical article bodies and the public kr/ tree were cleaned to remove AKS viewer navigation artifacts such as ????, ????, and ? before this snapshot.
The Hanmun source text remains canonical. Korean translation, Korean reading, semantic summary, and entity candidates are derivative layers that can be regenerated with MiniMax, MiMo, or other models as credits and quality controls improve.
? ???? AKS ??? ??? ?????? legalize-kr ???? ?? ?? ??? ????, ??? ?? ?? ??? ?? ?? ??? ??? ???? ?? ?? ??????.
?? ??? canonical? ????, ?? ???????? ?????? ??? ?? ???? ???. MiMo ?? ?? ? ?? ?? ?? ??? ?? ??? ??? ? ??? ?? ??? ?? ??? ??? ?????.