- README.md: bilingual card (English + `<a id="zh">` Chinese) -- what it is,
what it is for, how it is built, files, data format, per-language table
(code / language / native name / first line / words translated / hits /
entries, with a total row), regenerating, toolchain, license, About VocabRadar.
Front matter carries `language` (all 42 codes), `source_datasets:
vocabradar/wordfreq` and `task_categories: translation`.
- data/: 162 gzip JSONL packs `<src>_<dst>.jsonl.gz` (en/zh <-> 40 partner
languages), one fixed-field object per line -- `{"word": ..., "translations":
[...]}`; tracked through Git LFS, with byte sizes and SHA-256 in files.json.
- meta.json: per direction, the source, caps and per-language statistics.
- tools/wt.py: probe / fill / pack in one file; tools/migrate_fixed_schema.py
and tools/migrate_flat_layout.py: one-off, re-entrant migrations.
- docs/: planning notes.
- .gitignore keeps the build inputs out of the repo (build/ source dumps and
caches, …
