v1.3.0

Added FalkorDB graph store. FalkorDB currently only works with the SemanticGuidedRetriever, not the TraversalBasedRetriever.

What's Changed

• src: added FalkorDB GraphStore implementation i.e code, requirements, and docs by @kingtroga in #8

New Contributors

• @kingtroga made their first contribution in #8

Full Changelog: v1.2.1...v1.3.0

v1.2.1

Updated dependencies in requirements.txt

v1.2.0

This release introduces a SourceDocument object that groups chunks belonging to a single source as they pass through the extract and build pipelines.

v1.1.3

Updated requirements.txt

Full Changelog: v1.1.2...v1.1.3

v1.1.2

Improve source metadata in embeddings

v1.1.1

Updated LlamaIndex LLM dependencies

Full Changelog: v1.1.0...v1.1.1

v1.1.0

Added support for batch extraction using Amazon Bedrock batch inference
Updated LlamaIndex packages so as to support latest Bedrock FMs supported in llama-index-llms-bedrock
Updated extraction prompts so that they work better with a wider range of FMs
Update to Neptune 1.4.1.0 in quick start CloudFormation template
Added ability to configure number of threads per worker in extract stage
Logging config is now only applied if you call set_logging_config (allows toolkit to be incorporated into applications with their own logging configurations)

Full Changelog: v1.0.1...v1.1.0

v1.0.1

What's Changed

Fixed issues:

• Fixed GPU allocation issue in BGEReranker
• Improved error reporting for model invocation failures
• Improved checkpointing frequency

Full Changelog: v1.0.0...v1.0.1

v1.0.0

Initial release of graphrag-toolkit