[i18n] Translation gaps: 3 new keys missing from every non-English locale

Every non-English locale is currently missing the same 3 translation keys, added by two recent English-only PRs. Native/fluent speakers - PRs updating just your locale file are very welcome, no need to wait for someone to cover every language at once. ## Missing keys | Key | English text | | --- | --- | | `entity.sensor.nozzle_type.state.tpu_high_flow` | `TPU High-Flow` | | `config.error.tfa_csrf_fallback` | `Bambu Lab's two factor authentication endpoint is currently rejecting requests. A verification code has been sent to your email or phone instead - please enter it below.` | | `options.error.tfa_csrf_fallback` | (same text as above) | Each key lives in [`custom_components/bambu_lab/translations/.json`](../tree/main/custom_components/bambu_lab/translations), nested under those paths (e.g. `entity` -> `sensor` -> `nozzle_type` -> `state` -> `tpu_high_flow`) - see [`en.json`](../blob/main/custom_components/bambu_lab/translations/en.json) for the exact structure to match. ## Locales needing an update - [ ] `ca.json` - Catalan - [ ] `cs.json` - Czech - [ ] `da.json` - Danish - [ ] `de.json` - German - [ ] `el.json` - Greek - [ ] `es.json` - Spanish - [ ] `fr.json` - French - [ ] `it.json` - Italian - [ ] `ko.json` - Korean - [ ] `nl.json` - Dutch - [ ] `pl.json` - Polish - [ ] `pt-br.json` - Portuguese (Brazil) - [ ] `pt.json` - Portuguese (Portugal) - [ ] `sk.json` - Slovak - [ ] `th.json` - Thai - [ ] `zh-Hans.json` - Chinese (Simplified) ## How to help Open a PR adding those 3 keys to your locale's JSON file, matching the nesting shown in `en.json`. A new [CI check](../actions/workflows/translation-parity.yml) will flag on your PR (as a comment, non-blocking) if any locale still doesn't match `en.json` after your change - handy for double-checking you got the structure right. Alternatively, `python3 scripts/auto_translate.py` can machine-translate these automatically for all locales at once (uses Google Translate, no API key needed) - useful as a starting point, though a native speaker's correction is always preferred over raw machine translation.