Conversation
This fills in the gaps that CLDR has.
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: d902cbfdaf
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| populations = dict(reader) | ||
|
|
||
| return [code for code in languages if int(populations.get(code, "0")) == 0] |
There was a problem hiding this comment.
load_zero_population_codes() derives its worklist from the current population.csv, but this commit now populates population.csv with values from population-fallback.csv; as a result, rerunning this helper excludes previously backfilled languages and can produce a truncated fallback file that drops existing rows. If that output is committed and population.csv is regenerated, many languages revert to zero again, so the refresh workflow is not idempotent. The candidate set should come from unsupplemented CLDR output (or otherwise ignore existing fallback contributions) before querying Wikidata.
Useful? React with 👍 / 👎.
1 participant
This fills in the gaps that CLDR has.