chore: dictionary curation

[carlosroe]

Issues

N/A

Description

This is my first time trying to commit words to Harper. Any help and criticism is highly appreciated. I am neither an expert in linguistics nor a native English speaker, just someone who would like to help by adding words I frequently use in my research!

+ACC
+CBASP
+Glu
+MRS
+rsFC
+glutamate
+pregenual
+racemate
+racemic

How Has This Been Tested?

cargo test

AI Disclosure

• I am a human and didn't use any AI.
• I used LLM features of my editor, but not an agent.
• I used an AI agent interactively.
• I am an agent or I got an agent to do the work autonomously.
• I used LLM to find matching annotations for my words

Checklist

• I have performed a self-review of my own code
• I have added tests to cover my changes
• I have considered splitting this into smaller pull requests.

[hippietrail]

Some notes on the sort order of the dictionary being case sensitive and obscure/specialized words have a dedicated section.

[hippietrail]

This section is for all-uppercase. It sounds silly but when I tried to re-sort in a case insensitive ordering it revealed a latent bug where order of words in the dictionary makes a difference and made passing tests fail. A PR covering all aspects of case problems in the dictionary is being worked on.

[hippietrail]

Same here. Easiest way to get them in the right spot is to turn on case sensitive search on your editor. Around line 6267 is where you want to look.

[hippietrail]

Some of these look very specialized. The top/main part of the dictionary is for regular words. Specialized words are in a separate section so that they'll be easy to find if and when we decide to split them out into focussed Weir Packs. The specalized section starts around line 53,834 and 54,573 or so should be where a word starting with lowercase pg should go.

[hippietrail]

Looks good. Just one more to move. Thanks.

[hippietrail]

You've got failing "snapshot" tests. Your new words probably change the top 3 spelling suggestions for other unknown words in the old texts we test against. The way to fix this is to run cargo test again. Usually only once but occasionally twice. Then commit the changed snapshots with the PR.