Sdl-Community
Sdl-Community copied to clipboard
DeepL MT performance improvements
The plugin does invoke translate endpoint per each segment. This approach is not optimal. DeepL API allow to translate up to 50 segments at once. Translating more segments at once reduces pre-translation time by up to 60% (based on my tests). In addition, to make translation process more reliable I've added translation retrying routine in case of some http status codes received from DeepL Service.
Bump. Is this repo still alive?