autogen.agentchat.contrib.graph_rag.neo4j_native_graph_query_engine.Neo4jNativeGraphQueryEngine

Neo4jNativeGraphQueryEngine

Neo4jNativeGraphQueryEngine(
    host: str = 'neo4j://localhost',
    port: int = 7687,
    username: str = 'neo4j',
    password: str = 'password',
    embeddings: ForwardRef('Embedder') | None = None,
    embedding_dimension: int | None = 3072,
    llm: ForwardRef('LLMInterface') | None = None,
    query_llm: ForwardRef('LLMInterface') | None = None,
    entities: List[str] | None = None,
    relations: List[str] | None = None,
    potential_schema: List[tuple[str, str, str]] | None = None
)

A graph query engine implemented using the Neo4j GraphRAG SDK.
Provides functionality to initialize a knowledge graph, create a vector index, and query the graph using Neo4j and LLM.
Initialize a Neo4j graph query engine.

Parameters:

Name	Description
host	Type: str Default: ‘neo4j
port	Type: int Default: 7687
username	Type: str Default: ‘neo4j’
password	Type: str Default: ‘password’
embeddings	Type: ForwardRef(‘Embedder’) | None Default: None
embedding_dimension	Type: int | None Default: 3072
llm	Type: ForwardRef(‘LLMInterface’) | None Default: None
query_llm	Type: ForwardRef(‘LLMInterface’) | None Default: None
entities	Type: List[str] | None Default: None
relations	Type: List[str] | None Default: None
potential_schema	Type: List[tuple[str, str, str]] | None Default: None

Instance Methods

add_records

add_records(self, new_records: list[Document]) -> bool

Add new records to the Neo4j database.

Parameters:

Name	Description
new_records	List of new Documents to be added Type: list[Document]

Returns:

Type	Description
bool	bool: True if records were added successfully, False otherwise.

init_db

init_db(self, input_doc: list[Document] | None = None) -> None

Initialize the Neo4j graph database using the provided input doc.
Currently this method only supports single document input (only reads the first doc).
This method supports both text and PDF documents. It performs the following steps:
1. Clears the existing database.
2. Extracts graph nodes and relationships from the input data to build a knowledge graph.
3. Creates a vector index for efficient retrieval.

Parameters:

Name	Description
input_doc	Input documents for building the graph. Type: list[Document] | None Default: None

query

query(
    self,
    question: str,
    n_results: int = 1,
    **kwargs: Any
) -> GraphStoreQueryResult

Query the Neo4j database using a natural language question.

Parameters:

Name	Description
question	The question to be answered by querying the graph. Type: str
n_results	Type: int Default: 1
**kwargs	Type: Any

Returns:

Type	Description
GraphStoreQueryResult	GraphStoreQueryResult: The result of the query.