Skip to content

08. Translating Scan Results

tetherscript edited this page Feb 16, 2025 · 6 revisions

Translating

  • Reads target\Strings\en-US\Resources.resw and calls a Profile for each item with a hint token in the comments field.
  • Saves the translation in the cache, and then saves the translations and Specials in your non-en-US target\Strings\???\Resources.resw files, where ??? = de-DE, ar-SA, etc.

Caching Translations

Translations are cached in CacheEx.json for re-use. This saves cost and makes the translation process faster so you don't need to call the translation function as often.

You might do a translation, test it, approve and deploy it. Then for the next release, you translate again but this time the profile ex. OpenAI might use a slightly different model, returning a slightly different translations. These models will improve over time, but you don't want it re-translating things you have already approved. So translations that are cached, will not be replaced by the translation function.

Sometimes, the profile just doesn't get it right, and you need to manually edit the cache so the correct translation is used. You can do this in the Cache Editor.

Important

You can also run translations locally and for free, depending on the model/API/hardware you use. AI API specs/prompt designs are wildly non-standard and changing frequently as of this time, so the app's code is flexible, allowing you to code your own profile using the API of your choice.

Here's what translating the sample app Sample-Packaged looks like.

translate1

Summary

  • Your translations are complete, and .resw files are updated.
  • Start the target app, and you should see the translations.
Clone this wiki locally