Neo4j-graphrag

IMPORTANT NOTICE
- The `neo4j-genai` package is now deprecated. Users are advised to switch to the new package `neo4j-graphrag`.
Added
- Ability to visualise pipeline with `my_pipeline.draw("pipeline.png")`.
- `LexicalGraphBuilder` component to create the lexical graph without entity-relation extraction.

Fixed
- Pipelines now return correct results when the same pipeline is run in parallel.

Changed
- Pipeline run method now return a PipelineResult object.
- Improved parameter validation for pipelines (124). Pipeline now raise an error before a run starts if:
- the same parameter is mapped twice
- or a parameter is defined in the mapping but is not a valid component input

Added
- PDF-to-graph pipeline for knowledge graph construction in experimental mode
- Introduced support for Component/Pipeline flexible architecture.
- Added new components for knowledge graph construction, including text splitters, schema builders, entity-relation extractors, and Neo4j writers.
- Implemented end-to-end tests for the new knowledge graph builder pipeline.

Changed
- When saving the lexical graph in a KG creation pipeline, the document is also saved as a specific node, together with relationships between each chunk and the document they were created from.

Fixed
- Corrected the hybrid retriever query to ensure proper normalization of scores in vector search results.

Added
- Add optional custom_prompt arg to the Text2CypherRetriever class.

Changed
- `GraphRAG.search` method first parameter has been renamed `query_text` (was `query`) for consistency with the retrievers interface.
- Made `GraphRAG.search` method backwards compatible with the query parameter, raising warnings to encourage using query_text instead.

Fixed
- Corrected initialization to allow specifying the embedding model name.
- Removed sentence_transformers from embeddings/__init__.py to avoid ImportError when the package is not installed.

Added
- Stopped embeddings from being returned when searching with `VectorRetriever`. Added `nodeLabels` and `id` to the metadata of `VectorRetriever` results.
- Added `upsert_vector` utility function for attaching vectors to node properties.
- Introduced `Neo4jInsertionError` for handling insertion failures in Neo4j.
- Included Pinecone and Weaviate retrievers in neo4j_graphrag.retrievers.
- Introduced the GraphRAG object, enabling a full RAG (Retrieval-Augmented Generation) pipeline with context retrieval, prompt formatting, and answer generation.
- Added PromptTemplate and RagTemplate for customizable prompt generation.
- Added LLMInterface with implementation for OpenAI LLM.
- Updated project configuration to support multiple Python versions (3.8 to 3.12) in CI workflows.
- Improved developer experience by copying the docstring from the `Retriever.get_search_results` method to the `Retriever.search` method
- Support for specifying database names in index handling methods and retrievers.
- User Guide in documentation.
- Introduced result_formatter argument to all retrievers, allowing custom formatting of retriever results.

Changed
- Refactored import paths for retrievers to neo4j_graphrag.retrievers.
- Implemented exception chaining for all re-raised exceptions to improve stack trace readability.
- Made error messages in `index.py` more consistent.
- Renamed `Retriever._get_search_results` to `Retriever.get_search_results`
- Updated retrievers and index handling methods to accept optional database names.

Fixed

- Removed Pinecone and Weaviate retrievers from **init**.py to prevent ImportError when optional dependencies are not installed.
- Moved few-shot examples in `Text2CypherRetriever` to the constructor for better initialization and usage. Updated unit tests and example script accordingly.
- Fixed regex warnings in E2E tests for Weaviate and Pinecone retrievers.
- Corrected HuggingFaceEmbeddings import in E2E tests.