Commit graph

12 commits

Author SHA1 Message Date
Ed Page
8b4d196a52 feat(dict): September updates
- The correction to `usefull` and `becuase` has existed from the initial implementation,
  as far as I could tell.
- `busses` is a valid variation of `buses`
- Had to put in a hack so that `aligning` would be preferred over
  `alining` because of how the preference algorithm works

Fixes #823
2023-10-02 08:18:50 -05:00
Ed Page
b798780a13 refactor(dict): Make codegen more specific 2023-09-25 12:13:53 -05:00
Ed Page
0c05b217d4 style: Make clippy happy 2023-09-01 10:20:03 -05:00
Ed Page
a3bf84ade6 chore(dict): Don't clear disallowed words 2023-08-07 16:24:51 -05:00
Martin Fischer
8d026ac23e feat(dict): Preserve correction order
We want to be able to recommend more likely corrections first,
e.g. for "poped" we want to recommend "popped" before "pooped".
2023-06-26 22:13:47 +02:00
Martin Fischer
89d5a97a8a test: Add some tests for dict processing logic 2023-06-26 19:23:27 +02:00
Martin Fischer
49a0eaab7b refactor: Make dict processing logic testable
Previously all the dictionary cleanup logic was in the function:

    fn generate<W: std::io::Write>(file: &mut W, dict: &[u8])

which parsed the provided buffer as CSV and also took care of writing
the processed dictionary back as CSV.  This commit factors out the CSV
handling, leaving a `process` function behind so that it can be easily
tested in the following commit.
2023-06-26 19:22:04 +02:00
Ed Page
a9d9fc03a2 test(dict): Report more cases to user 2023-06-08 09:23:10 -05:00
Ed Page
09f096a096 chore(dict): Automate more cleanup 2023-06-08 08:54:36 -05:00
Ed Page
c963f68083 fix(dict): Remove nilable
See conversation in #613
2022-12-06 10:47:14 -06:00
Ed Page
c49aff00be test: Make platform agnostic 2022-09-01 07:15:42 -05:00
Ed Page
a8599f6a19 test: Move codegen to tests 2022-08-03 09:07:04 -05:00