Enable string extraction tooling (e.g. i18n-extract) to surface missing or new keys to translate

[churik]

Problem

There’s no automated way to detect new or missing translation keys, which makes it easy to overlook untranslated content. A string extraction tool will help maintain complete and consistent language files.

Requires DevOps input.

Implementation

• Integrate i18n-extract (or equivalent) into the build or CI process
• Configure it to compare source code usage vs. existing translation files
• Output a report of missing or unused keys

Acceptance Criteria

• Tool can detect missing keys based on source usage
• Tool does not introduce breaking changes to build
• Report can be used to update en.json and sync with Crowdin

Notes

• Keep the tool optional in local dev flow but required in CI.
• Useful as a QA and dev support tool before translation syncs.

Future Steps

• Automate pushing extracted source keys to Crowdin
• Use the report to flag regressions in i18n coverage

[ilmotta]

There’s no automated way to detect new or missing translation keys

We already have a linter written in CLJS to detect missing translation keys in the file scripts/lint/translations.clj but it focuses on the en.json file. There's no 100% deterministic way to detect all keys because IIRC there are a few cases where the keys are dynamic and can't be detected by the simple :t/* keyword convention, so this requirement deserves to be evaluated because it may not be feasible without significant changes.

That CLJS script could, in theory, evolve to cover our new needs.