v0.1.2: Parquet Exports and Architecture Documentation

🚀 v0.1.2 Release

This release introduces compressed Parquet support for more efficient data handling and a significant restructuring of the project documentation.

✨ Highlights

• Parquet Export Support: The sentiment pipeline now exports results as compressed .parquet.gzip files by default, providing better performance and smaller storage footprints than CSV.
• New Architecture Documentation: Added docs/architecture.md, consolidating technical details, file structures, and LaTeX formulas for sentiment aggregation methods.
• Improved Metadata: Added authors field and public API type hints for better downstream library usage.

📝 Full Changelog

Added

• Created docs/architecture.md combining architecture and file structure details.
• Added LaTeX formulas for posneg and allwords aggregation in the architecture documentation.
• Added input file paths to the configuration overview table in README.md.
• Added maintainer info and public API type hints to pyproject.toml.

Changed

• Pipeline: Updated AutoEconSentiment and TextCleaner to export results as compressed Parquet files.
• Documentation: Migrated detailed library components from the README to specialized docs.
• Notebooks: Updated autoecon_demo.ipynb and demo_cb_speechs.ipynb to reflect the new Parquet output workflow.
• Dependencies: Tightened upper bounds on matplotlib and seaborn to ensure stability.

---

Full Diff: v0.1.1...v0.1.2