Bulk-translate


What's new

:fire: **Major**: Support schema! You no longer need to group your JSONL content afterwards! See: 19 and 20

See `--schema` parameter below, which supports passing JSON for mapping output columns:
bash
python -m bulk_translate.translate \
--src "<YOUR-INPUT-CSV-JSONL-DATA>" \
--schema '{"translated":"{text}"}' \
--adapter "dynamic:googletrans_310a.py:GoogleTranslateModel" \
--batch-size 10 \
%%m \
--src "auto" \
--dest "ru"


Models are available at `nlp-thirdgate` :milky_way:
https://github.com/nicolay-r/nlp-thirdgate?tab=readme-ov-file

**Full Changelog**: https://github.com/nicolay-r/bulk-translate/compare/0.25.1...0.25.2


What's new

See the related issue:
https://github.com/nicolay-r/bulk-translate/issues/17

**Full Changelog**: https://github.com/nicolay-r/bulk-translate/compare/0.25.0...0.25.1


Major Updates

👉 This is the main page for API usage example: https://github.com/nicolay-r/bulk-translate/wiki
- 🔥 Handly API support 6
- 🔥 We have to pick the practice of the bulk-chain for batching organization 7
- 🔥 Support complex prompts for combining several lists in one 12
- API: support spans as lists in list of text parts 9
- parse-entities flag is no longer applicable 14
- AREkit embed into project and remove parent_ctx 10
- Move AREkit components into the repo 15

Minor Fixes
- 🔧 AREkit API feedback and necessity of the IDLE mode. 5
- Refactoring Entities annotation on index-based SPANS 3
- csv fix the problem of passing extra parameters, separately from translator model 11

**Full Changelog**: https://github.com/nicolay-r/bulk-translate/compare/0.24.1...0.25.0