- Add original_text field to LLM output for better translation accuracy
- Implement validation to check if LLM returned original text matches input text
- Adjust importance weights to better balance accuracy and fluency (1:0.6)
- Implement concurrent translation processing for improved performance
- Add --max-workers option to control the number of concurrent translation requests
- Add thread locks to prevent race conditions in debug and warning outputs
- Add context limit handling to detect and warn about truncated outputs
- Add warnings for translation failures and count mismatches
- Update README files with new features and context limit handling information