Kanji entry methodology

sschmidTU / mr-kanji-search-wtk

WTK-Search is a Kanji search engine using (multiple) Wanikani radicals or RTK names, on a RTK element dataset of 3000+ Kanji

https://sschmidtu.github.io/mr-kanji-search-wtk/

7 stars 2 forks source link

Kanji entry methodology #30

Closed sschmidTU closed 1 month ago

sschmidTU commented 2 years ago

a07dee819fd57f59f8212bf5b36f73683c14f8d9 added a script to generate a new kanji page with stroke number etc filled out: npm run newkanji [kanji] (or npm run newkanji -- [kanji] if you want to be verbose/explicit) (or with node only: node _tools/createNewKanjiPage.mjs [kanji])

if you added or changed elements in elements_data.txt, create the new elements json: npm run elements (or: node _tools/elementsDataToJson.js)

It creates a file [number].md with the new info in rtk3-remain/.

Now we only need to add variants, elementsTree, etc. (and check the data fetched from kanjiapi.dev)

sschmidTU commented 2 years ago

In elements_data.txt, subelement synonyms can now be listed like this: wand = magic wand = augury instead of: wand, magic wand, augury; wand=magic wand=augury

this saves a column, typing, redundance, etc. Also, all rows in the file can now be checked to have exactly 4 columns, not possibly 5 or more (>= 4).

sschmidTU commented 2 years ago

[X] TODO createNewKanjiPage.mjs: check if the "new" kanji already exists in .md files. I accidentally added 3089.md 迄, which already existed as 2421.md. fixed in e5fadaf3d6632947eeabf7ab3a8778f0442de2de, now 3089.md is 囘 (actually new kanji) (edit: script now also checks for duplicate)
[X] add separate script checking for duplicates in all existing kanji? -> npm test / npm run duplicates / checkKanjiDuplicates.mjs

sschmidTU commented 2 years ago

We should add positioning data to elements_data.txt now, instead of just unordered elements.

E.g. for summer (夏), instead of one, ceiling, drop, eye, walking legs, we should give t(ceiling, drop, eye, walking legs). This will allow us to also show the structure of sub elements, e.g. for 嗄, which has elementsTree: l(mouth, summer). Currently, it just adds the elements for summer without positioning info.

This will require some refactoring of how we handle elements_data.txt. (for now, just remove all l(, t(, o(, c(, tl( etc. and ) substrings, like in search.js) Also, we will have to again add subelements or synonyms from this, e.g. one from ceiling.

-> split into sub-issue #32

sschmidTU commented 1 month ago

Done.