Wordward Mandarin Chinese vocabulary and pronunciation data

Primary source: CC-CEDICT, maintained by the CC-CEDICT community and published by
MDBG. This dictionary continues CEDICT, Copyright (C)1997,1998 Paul Andrew Denisowski.
https://www.mdbg.net/chinese/dictionary?page=cc-cedict
Pinned official export:2026-10-23T23:15:44Z,125314entries,version1,subversion0.
https://www.mdbg.net/chinese/export/cedict/cedict_1_0_ts_utf-8_mdbg.txt.gz

CC-CEDICT is licensed under Creative Commons Attribution-ShareAlike4.0 International:
https://creativecommons.org/licenses/by-sa/4.0/
The full legal text is in chinese-source/LICENSE-CC-BY-SA-4.0.txt.
Its original header and source are preserved in chinese-source/cedict_ts.u8.

Transformations by this project: four-basic-Han filtering; whole-word Simplified
identities; source-wide Traditional/variant alias extraction; fixed whole-word
Mandarin reading selection; documented citation corrections; explicit editorial
answer selection; initial/final/tone metadata and validation reports. The derived
dictionary, answer data and metadata are distributed under CC BY-SA4.0. This notice
does not state that unrelated game code, art, or services are dictionary derivatives.

Bounded supplements:
34explicit ordinary gaps from CedPane by Silas S. Brown, pinned commit
5fe54638334f2d572c8ad2efd6fe5508e6c966c2. The author dedicates this data to the public
domain; the repository includes the Unlicense. Preserve its original notices:
chinese-source/cedpane/LICENSE and chinese-source/cedpane/README.md.
https://github.com/ssb22/CedPane/tree/5fe54638334f2d572c8ad2efd6fe5508e6c966c2
https://ssb22.user.srcf.net/cedpane/

10ordinary expressions and2citation-reading corrections from English Wiktionary's
Chinese entries, credited to the Wiktionary contributors under CC BY-SA4.0. Exact
page/revision links, extracted Chinese source sections, and hashes are preserved in
chinese-source/supplement.json, reading-corrections.json and wiktionary/.
Those links also provide contributor histories. No other language section is used.

One finite factual exception, 网上购物, records its spelling and whole-expression
pronunciation from the TrainChinese and Collins sources cited in
chinese-source/factual-exceptions.txt. Its short gloss is original to this project;
no publisher database, extended definition, or example collection was copied.

Optional corpus QA credits wordfreq by Robyn Speer, using its large_zh data under
CC BY-SA 4.0. Its full metadata and source acknowledgments are retained in
chinese-source/qa-notices/wordfreq-3.1.1-METADATA.txt, including Google Books,
Leeds, Wikipedia, ParaCrawl, OPUS/OpenSubtitles and SUBTLEX authors. SUBTLEX data
remain freely available; Chinese frequency research includes Cai, Q. and
Brysbaert, M. (2010), SUBTLEX-CH, https://doi.org/10.1371/journal.pone.0010729.
Wordfreq code uses Apache 2.0; the independent pypinyin checker uses MIT. Their
notices are preserved separately. Optional runtimes are not part of this package.
Wordfreq citation: Robyn Speer (2022), https://doi.org/10.5281/zenodo.7199437.

Changes do not imply endorsement by MDBG, CC-CEDICT contributors, Silas S. Brown,
Wiktionary, TrainChinese, Collins, or any other source author.

Editorial/source verification is by an AI agent. Native Mandarin review is pending.
Source dictionaries and accepted guesses include uncommon words, names and technical
terms. The answer selection is separate and excludes name-only entries and initial
erhua puzzles. Coverage remains incomplete; detailed omissions are retained.