- Added support for an `include_paths` parameter to filter files based on specified prefixes, taking priority over `exclude_paths`.
- Updated the `parse_xliff` and `process_language` functions to accommodate the new filtering logic.
- Enhanced command-line argument parsing to allow users to specify files for translation directly, improving flexibility in translation management.
- Removed the default output directory constant and updated the logic to set the output directory based on the provided argument or default to a "translated" subdirectory within the bundle directory.
- Improved help text for the output directory argument to clarify its default behavior.
- Introduced a new variable to track successfully translated units, improving clarity and efficiency in the processing logic.
- Updated the logic for combining existing and new translation units, ensuring accurate output and preventing unnecessary uploads when no new translations are available.
- Implemented a new function to remove duplicate paragraphs from cached translations, addressing issues with the API returning repeated content.
- Updated the caching logic to apply deduplication both when loading from cache and after generating new translations, ensuring cleaner and more accurate cached data.
- Enhanced the `find_xliff_files` function to preserve the order of locales when mapping Crowdin locales to XLIFF paths.
- Updated the logic to return all available locales when no specific locales are provided, ensuring better handling of XLIFF files.
- Simplified the iteration over XLIFF files by removing unnecessary sorting in the final loop.
- Introduced a new Python script for automating the translation of Crowdin XLIFF bundles using the Alibaba Cloud Qwen-MT API.
- Updated .gitignore to include the Crowdin translation working directory, ensuring that temporary files are not tracked.
- The script supports downloading, parsing, translating, and uploading XLIFF files, streamlining the localization process.
