Indox

We're excited to announce the release of version 0.1.31, bringing **Memgraph support** and **LLM bug fixes** to further enhance our platform!

🚀 New Features

- **Graph Database Integration**
- **Memgraph Support**: In addition to Neo4j, we've now added **Memgraph** support, providing seamless integration for high-performance, in-memory graph databases.

🛠 Improvements

- **LLM Flexibility**
- **Model Customization**: Improved fine-tuning and customization options for LLM models, allowing better adaptability to specific use cases.
- **Dynamic Model Loading**: New feature enabling on-the-fly model switching and loading based on task requirements.
- **Expanded Model Support**: Compatibility with a wider range of LLM architectures and versions.

- **Performance Optimizations**
- Enhanced performance for multi-vector retrievals across all supported databases.
- Improved error handling and logging for better debugging and monitoring.

🐛 Bug Fixes

- Fixed minor issues with vector similarity calculations in certain edge cases.
- Resolved a rare concurrency issue in the MultiQueryRetrieval system.
- Addressed bugs related to LLM switching and model performance.

📚 Documentation

- Added guides for setting up and using **Memgraph** with our platform.
- Updated documentation to reflect LLM customization and dynamic model loading options.

As always, we're committed to continually improving our platform and value your feedback. If you encounter any issues or have suggestions, please don't hesitate to open an issue.

Thank you for your continued support!

---

We're excited to announce the release of version 0.1.29, bringing powerful new database support and enhanced LLM flexibility to our platform!

🚀 New Features

Graph Database Integration
* **Neo4j Support**: Seamless integration with Neo4j, enabling powerful graph-based data modeling and querying capabilities.

Enhanced LLM Flexibility
* **Model Customization**: Improved options for fine-tuning and customizing LLM models to better suit specific use cases.
* **Dynamic Model Loading**: New feature allowing for on-the-fly model switching and loading based on task requirements.
* **Expanded Model Support**: Added compatibility with a wider range of LLM architectures and versions.

🛠 Improvements
* Optimized performance for multi-vector retrievals across all supported databases.
* Enhanced error handling and logging for better debugging and monitoring.
* Improved documentation covering new graph database features and LLM customization options.

🐛 Bug Fixes
* Resolved minor issues with vector similarity calculations in certain edge cases.
* Fixed a rare concurrency issue in the MultiQueryRetrieval system.

📚 Documentation
* Added comprehensive guides for setting up and using Neo4j with our platform.
* Updated LLM configuration documentation to reflect new flexibility options.


---

We're constantly working to improve our platform and welcome your feedback. If you encounter any issues or have suggestions, please don't hesitate to [open an issue](link-to-issues).

Thank you for your continued support!

- **Couchbase DB Integration**: Added support for Couchbase as a flexible, scalable NoSQL database for high-performance data storage and retrieval.
- **Qdrant Integration**: Seamless integration with Qdrant, a vector search engine optimized for neural networks, enhancing the performance of vector search and retrieval.
- **VESPA Integration**: Support for Vespa.ai, enabling state-of-the-art retrieval and serving of large-scale data and vectors.
- **Apache Cassandra Integration**: Integration with Apache Cassandra for distributed, scalable database management.
- **RedisDB Integration**: Added support for RedisDB, allowing fast in-memory data storage and retrieval, perfect for real-time applications.
- **LanternDB Integration**: New integration with LanternDB, providing efficient storage and fast search capabilities for vector data.
- **SingleStoreVectorDB Integration**: Integration with SingleStoreVectorDB, designed for high-performance analytical processing and vector-based retrieval.
- **DeepLake Expansion**: Enhanced support for DeepLake, ensuring even better management of large vector data stores.

🛠 **Advanced Retrieval Systems**:
- **MultiQueryRetrieval**: New feature allowing multiple query retrievals simultaneously, enabling more complex and flexible search strategies.
- **MultiVectorRetriever**: Added support for retrieving vectors from multiple sources simultaneously, enhancing performance and flexibility for diverse use cases.

🛠 General Improvements:
- Streamlined multi-vector store options, improving the overall performance and adaptability of the platform.
- Enhanced data processing capabilities across all modules for optimized handling of large datasets.

🚀 New Features:

- **Azure OpenAI Embeddings Model**: Support for Azure OpenAI embeddings for enhanced NLP tasks.
- **DuckDB for Vector Store**: Integration of DuckDB as a high-performance vector store for analytical workloads.
- **DeepLake for Vector Store**: DeepLake integration for scalable and efficient vector data management.

🛠️ General Improvements:

- Enhanced data management with multiple vector store options.
- Improved performance and flexibility across modules.

Stay tuned for more updates on the Indox platform!