For a personal project I want to create a (huge) graph that contains some (as many as possible) chinese characters along the paths of their corresponding keyboard input using the 5 stroke classes horizontal(一), vertical (丨), left falling (丿), right falling/dot (丶), and turning (乛/乙). I didn't have a problem to find a database that included a list of some thousand characters transcribed for their use in the Wubihua input method, but in this method only the 4 first strokes plus the last stroke is used, as opposed to a transcript of all the strokes building up that character.

For example, for 毓 you only type 丿一乛乛乛.

However, I am looking for a transcript like this: 果 = 丨乛一一一丨丿丶

The strokes used can be encoded in whatever way, I just need the full information. For example, I found an open source project for the Wubihua method (= exactly the 4+1 way I am not looking for) that encoded them like this:

nmmn. 即 .mmm/ 許 m/nm/ 區 m/m., 報 /nmmm 量 m//mm 基 m/mm/ 打 m//mm 聯 m/n,n 西 .mmm/ 該 m/mmn 指

with ., /, ,, n and m representing the 5 stroke classes mentioned above. I've been searching in GitHub repos for hours, but was only able to find the short 4+1 strokes version used in Wubihua.

I would be very grateful for a hint or link to a free-to-use database containing the necessary information for, say, at least some hundred of the most frequent characters.

Thanks in advance!

  • Have you tried zdic? See this
    – dROOOze
    Commented Jun 2, 2018 at 15:32

1 Answer 1


The keyword to search for is 筆順編號.

This github respository contains stroke order sequences for 29685 characters, coded as numbers 1-5. From the readme:


“开”字,按笔顺“一、一、丿、丨”,编码为1132; “我”字为31;“向”字为325;“力”为53; 注意“万”为153,“方”为4153,“忄”为442。

其中有些笔画容易被误解: “提”归为“一”:如“氵、扌”中的最后一笔;有些电脑字体繁体字的“雨字头”四点显示为四小横,皆按国标笔顺归为四点;




Bolded parts are easy to misinterpret strokes - but they've grouped it under the same 5 classifications as you've mentioned in the question details, so I guess this wasn't a problem for you.

  • Thank you so much! That's exactly what I was looking for!
    – user19808
    Commented Jun 3, 2018 at 13:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.