We are excited to announce the release of ClickhouseRAG v0.2.0! This update brings significant enhancements and new features to improve your experience with Clickhouse data management and Retrieval-Augmented Generation (RAG) operations.
New Features:
- **Transformers Vectorizer Integration**: Added support for using Transformers models to convert text to vectors, enabling more powerful and flexible data embeddings.
- **Cosine Similarity Search**: Implemented cosine similarity search to find and rank data based on vector similarity.
- **Bulk Data Operations**: Added methods to insert and vectorize bulk data entries efficiently.
- **Backup and Restore**: Introduced functionalities to backup the database to a JSON file and restore it, ensuring data safety and ease of transfer.
Improvements:
- **Enhanced Exception Handling**: Improved error logging and handling across all modules for better debugging and reliability.
- **Optimized Query Execution**: Refactored query execution methods for better performance and security.
- **Detailed Logging**: Added comprehensive logging throughout the package to help track operations and diagnose issues.
Bug Fixes:
- **Resolved SQL Injection Risks**: Fixed potential SQL injection vulnerabilities in query construction.
- **Data Validation**: Improved data validation mechanisms to ensure data integrity before performing operations.
Installation:
You can install the latest version via pip:
sh
pip install clickhouserag
Usage:
Check out the detailed usage examples in our updated [README](https://github.com/RebelRaider/ClickhouseRAG#usage) to get started with the new features.
For any questions or issues, please open an issue on our [GitHub repository](https://github.com/RebelRaider/ClickhouseRAG/issues) or contact Leonid Chesnikov at leonid.chesnikovgmail.com.
Thank you for using ClickhouseRAG! We appreciate your feedback and contributions.
**Full Changelog**: https://github.com/RebelRaider/ClickhouseRAG/commits/v0.2.0