Host-only data files for WM Keyboard, an Android keyboard app I'm currently developing. These are the large source/dictionary files the app downloads on demand (they are deliberately not bundled in the APK, to keep it small), so this repository is pushed first, on its own, for testing the app's download paths before the app itself goes public.
This is not the addon repository (themes, layouts, snippets, emoji packs). That lives separately.
| Folder | What | Source |
|---|---|---|
data/ |
Per-language word-frequency lists (<lang>/<lang>_full.txt.gz), optional offensive lists (<lang_offensive.txt.gz), and emoji search dictionaries (<lang>/<lang>_emoji.json.gz) across 160+ languages |
Various (see data/README.md) |
cjk/ |
Chinese Pinyin (pinyin.tsv), Japanese kana (ja_kana.tsv) and Chinese stroke (stroke.tsv) conversion tables, plus the build scripts under cjk/tools/ — see cjk/README.md |
CC-CEDICT, Mozc, BSD 2-Clause |
The data/ word lists and emoji dictionaries are gzip-compressed (<lang>_full.txt.gz, <lang>_offensive.txt.gz, and <lang>_emoji.json.gz) so every
file stays under GitHub's 100 MB limit — gunzip to get the plain text/JSON data. The cjk/ tables are stored uncompressed.
-
data/— the frequency word lists, offensive word lists, and emoji dictionaries are sourced from various open-source projects including Hermit Dave's FrequencyWords, the Leipzig Corpora Collection, KDE kemoji / Unicode CLDR, and others. Frequency lists are licensed under CC BY-SA 4.0, MIT, Public Domain, or CC BY-SA 3.0. Emoji dictionaries are derived from Unicode CLDR annotations & Unicode Emoji Data under the Unicode License Agreement (v3) / CC0 1.0 Universal. Seedata/README.mdfor the per-language attribution table and full licensing details. -
cjk/—pinyin.tsvfrom CC-CEDICT (CC BY-SA 4.0);ja_kana.tsvfrom the Google Mozc OSS dictionary (BSD 3-Clause);stroke.tsvfrom yefeijiang/Chinese-characters-code-table (BSD 2-Clause). Per-file formats, checksums, sources and rebuild steps are incjk/README.md; the build scripts are incjk/tools/.