Closed Kirchheim closed 3 days ago
How are you thinking this should be grouped? Maybe one group for each 1000 characters? For example, kanji grid includes a grouping for the JPDB kanji frequency list which uses this way to group. https://github.com/Kuuuube/kanjigrid/blob/master/data/jpdb_kanji_frequency_list.json
Group per 1000 sounds good to me
Am 14.11.2024 um 04:41 schrieb Kuuuube @.***>:
How are you thinking this should be grouped? Maybe one group for each 1000 characters? For example, kanji grid includes a grouping for the JPDB kanji frequency list which uses this way to group. https://github.com/Kuuuube/kanjigrid/blob/master/data/jpdb_kanji_frequency_list.json
— Reply to this email directly, view it on GitHub https://github.com/Kuuuube/kanjigrid/issues/18#issuecomment-2475348774, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACPHBHMUGG23JNNNMY4DKL32AQLVBAVCNFSM6AAAAABRX5H652VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINZVGM2DQNZXGQ. You are receiving this because you authored the thread.
Added! I set the language for this to zh-Hant
(traditional characters), please let me know if that is incorrect.
The Cantonese ranking Words.hk http://words.hk/ and Humanum are all based on Hong Kong, and Hong Kong uses Traditional Characters only
https://www.wikidata.org/wiki/Q18130932
so zh-Hant is correct.
Thanks.
Am 14.11.2024 um 05:39 schrieb Kuuuube @.***>:
Added! I set the language for this to zh-Hant (traditional characters), please let me know if that is incorrect.
— Reply to this email directly, view it on GitHub https://github.com/Kuuuube/kanjigrid/issues/18#issuecomment-2475399792, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACPHBHJ2AIG5EP4ATXC6AQD2AQSQJAVCNFSM6AAAAABRX5H652VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINZVGM4TSNZZGI. You are receiving this because you authored the thread.
Group characters by frequency rank from this site https://humanum.arts.cuhk.edu.hk/Lexis/lexi-can/ I cannot get direct link, but if you click menu left "Ranking Chart", you see the table I refer to
I have downloaded that table manually and converted it to attached csv file Humanum-Hanzi-Rank-comma-delimited.csv