Neo4j message history

CHANGELOG.md

@@ -8,6 +8,9 @@
 - Support for effective_search_ratio parameter in vector and hybrid searches.
 - Introduced upsert_vectors utility function for batch upserting embeddings to vector indexes.
 - Introduced `extract_cypher` function to enhance Cypher query extraction and formatting in `Text2CypherRetriever`.
+- Introduced Neo4jMessageHistory and InMemoryMessageHistory classes for managing LLM message histories.
+- Added examples and documentation for using message history with Neo4j and in-memory storage.
+- Updated LLM and GraphRAG classes to support new message history classes.
 
 ### Changed
 

docs/source/api.rst

@@ -403,6 +403,15 @@ Database Interaction
 .. autofunction:: neo4j_graphrag.schema.format_schema
 
 
+***************
+Message History
+***************
+
+.. autoclass:: neo4j_graphrag.message_history.InMemoryMessageHistory
+
+.. autoclass:: neo4j_graphrag.message_history.Neo4jMessageHistory
+
+
 ******
 Errors
 ******

docs/source/index.rst

@@ -148,6 +148,7 @@ Note that the below example is not the only way you can upsert data into your Ne
 
 
 .. code:: python
+
     from neo4j import GraphDatabase
     from neo4j_graphrag.indexes import upsert_vectors
     from neo4j_graphrag.types import EntityType

docs/source/user_guide_rag.rst

@@ -917,6 +917,7 @@ Populate a Vector Index
 ==========================
 
 .. code:: python
+
     from random import random
 
     from neo4j import GraphDatabase

examples/README.md

@@ -52,7 +52,7 @@ are listed in [the last section of this file](#customize).
 
 - [End to end GraphRAG](./answer/graphrag.py)
 - [GraphRAG with message history](./question_answering/graphrag_with_message_history.py)
-
+- [GraphRAG with Neo4j message history](./question_answering/graphrag_with_neo4j_message_history.py)
 
 ## Customize
 
@@ -75,6 +75,7 @@ are listed in [the last section of this file](#customize).
 - [Custom LLM](./customize/llms/custom_llm.py)
 
 - [Message history](./customize/llms/llm_with_message_history.py)
+- [Message history with Neo4j](./customize/llms/llm_with_neo4j_message_history.py)
 - [System Instruction](./customize/llms/llm_with_system_instructions.py)
 
 

examples/customize/llms/custom_llm.py

@@ -1,9 +1,10 @@
 import random
 import string
-from typing import Any, Optional
+from typing import Any, List, Optional, Union
 
 from neo4j_graphrag.llm import LLMInterface, LLMResponse
 from neo4j_graphrag.llm.types import LLMMessage
+from neo4j_graphrag.message_history import MessageHistory
 
 
 class CustomLLM(LLMInterface):
@@ -15,7 +16,7 @@ def __init__(
     def invoke(
         self,
         input: str,
-        message_history: Optional[list[LLMMessage]] = None,
+        message_history: Optional[Union[List[LLMMessage], MessageHistory]] = None,
         system_instruction: Optional[str] = None,
     ) -> LLMResponse:
         content: str = (
@@ -26,7 +27,7 @@ def invoke(
     async def ainvoke(
         self,
         input: str,
-        message_history: Optional[list[LLMMessage]] = None,
+        message_history: Optional[Union[List[LLMMessage], MessageHistory]] = None,
         system_instruction: Optional[str] = None,
     ) -> LLMResponse:
         raise NotImplementedError()

examples/customize/llms/llm_with_neo4j_message_history.py

@@ -0,0 +1,59 @@
+"""This example illustrates the message_history feature
+of the LLMInterface by mocking a conversation between a user
+and an LLM about Tom Hanks.
+
+Neo4j is used as the database for storing the message history.
+
+OpenAILLM can be replaced by any supported LLM from this package.
+"""
+
+import neo4j
+from neo4j_graphrag.llm import LLMResponse, OpenAILLM
+from neo4j_graphrag.message_history import Neo4jMessageHistory
+
+# Define database credentials
+URI = "neo4j+s://demo.neo4jlabs.com"
+AUTH = ("recommendations", "recommendations")
+DATABASE = "recommendations"
+INDEX = "moviePlotsEmbedding"
+
+# set api key here on in the OPENAI_API_KEY env var
+api_key = None
+
+llm = OpenAILLM(model_name="gpt-4o", api_key=api_key)
+
+questions = [
+    "What are some movies Tom Hanks starred in?",
+    "Is he also a director?",
+    "Wow, that's impressive. And what about his personal life, does he have children?",
+]
+
+driver = neo4j.GraphDatabase.driver(
+    URI,
+    auth=AUTH,
+    database=DATABASE,
+)
+
+history = Neo4jMessageHistory(session_id="123", driver=driver, window=10)
+
+for question in questions:
+    res: LLMResponse = llm.invoke(
+        question,
+        message_history=history,
+    )
+    history.add_message(
+        {
+            "role": "user",
+            "content": question,
+        }
+    )
+    history.add_message(
+        {
+            "role": "assistant",
+            "content": res.content,
+        }
+    )
+
+    print("#" * 50, question)
+    print(res.content)
+    print("#" * 50)

examples/question_answering/graphrag_with_neo4j_message_history.py

@@ -0,0 +1,87 @@
+"""End to end example of building a RAG pipeline backed by a Neo4j database,
+simulating a chat with message history which is also stored in Neo4j.
+
+Requires OPENAI_API_KEY to be in the env var.
+"""
+
+import neo4j
+from neo4j_graphrag.embeddings.openai import OpenAIEmbeddings
+from neo4j_graphrag.generation import GraphRAG
+from neo4j_graphrag.llm import OpenAILLM
+from neo4j_graphrag.message_history import Neo4jMessageHistory
+from neo4j_graphrag.retrievers import VectorCypherRetriever
+
+# Define database credentials
+URI = "neo4j+s://demo.neo4jlabs.com"
+AUTH = ("recommendations", "recommendations")
+DATABASE = "recommendations"
+INDEX = "moviePlotsEmbedding"
+
+
+driver = neo4j.GraphDatabase.driver(
+    URI,
+    auth=AUTH,
+)
+
+embedder = OpenAIEmbeddings()
+
+retriever = VectorCypherRetriever(
+    driver,
+    index_name=INDEX,
+    retrieval_query="""
+        WITH node as movie, score
+        CALL(movie) {
+            MATCH (movie)<-[:ACTED_IN]-(p:Person)
+            RETURN collect(p.name) as actors
+        }
+        CALL(movie) {
+            MATCH (movie)<-[:DIRECTED]-(p:Person)
+            RETURN collect(p.name) as directors
+        }
+        RETURN movie.title as title, movie.plot as plot, movie.year as year, actors, directors
+    """,
+    embedder=embedder,
+    neo4j_database=DATABASE,
+)
+
+llm = OpenAILLM(model_name="gpt-4o", model_params={"temperature": 0})
+
+rag = GraphRAG(
+    retriever=retriever,
+    llm=llm,
+)
+
+history = Neo4jMessageHistory(session_id="123", driver=driver, window=10)
+
+questions = [
+    "Who starred in the Apollo 13 movies?",
+    "Who was its director?",
+    "In which year was this movie released?",
+]
+
+for question in questions:
+    result = rag.search(
+        question,
+        return_context=False,
+        message_history=history,
+    )
+
+    answer = result.answer
+    print("#" * 50, question)
+    print(answer)
+    print("#" * 50)
+
+    history.add_message(
+        {
+            "role": "user",
+            "content": question,
+        }
+    )
+    history.add_message(
+        {
+            "role": "assistant",
+            "content": answer,
+        }
+    )
+
+driver.close()

src/neo4j_graphrag/generation/graphrag.py

@@ -16,7 +16,7 @@
 
 import logging
 import warnings
-from typing import Any, Optional
+from typing import Any, List, Optional, Union
 
 from pydantic import ValidationError
 
@@ -28,6 +28,7 @@
 from neo4j_graphrag.generation.types import RagInitModel, RagResultModel, RagSearchModel
 from neo4j_graphrag.llm import LLMInterface
 from neo4j_graphrag.llm.types import LLMMessage
+from neo4j_graphrag.message_history import MessageHistory
 from neo4j_graphrag.retrievers.base import Retriever
 from neo4j_graphrag.types import RetrieverResult
 
@@ -84,7 +85,7 @@ def __init__(
     def search(
         self,
         query_text: str = "",
-        message_history: Optional[list[LLMMessage]] = None,
+        message_history: Optional[Union[List[LLMMessage], MessageHistory]] = None,
         examples: str = "",
         retriever_config: Optional[dict[str, Any]] = None,
         return_context: bool | None = None,
@@ -102,7 +103,8 @@ def search(
 
         Args:
             query_text (str): The user question.
-            message_history (Optional[list]): A collection previous messages, with each message having a specific role assigned.
+            message_history (Optional[Union[List[LLMMessage], MessageHistory]]): A collection previous messages,
+                with each message having a specific role assigned.
             examples (str): Examples added to the LLM prompt.
             retriever_config (Optional[dict]): Parameters passed to the retriever.
                 search method; e.g.: top_k
@@ -127,7 +129,9 @@ def search(
             )
         except ValidationError as e:
             raise SearchValidationError(e.errors())
-        query = self.build_query(validated_data.query_text, message_history)
+        if isinstance(message_history, MessageHistory):
+            message_history = message_history.messages
+        query = self._build_query(validated_data.query_text, message_history)
         retriever_result: RetrieverResult = self.retriever.search(
             query_text=query, **validated_data.retriever_config
         )
@@ -147,12 +151,14 @@ def search(
             result["retriever_result"] = retriever_result
         return RagResultModel(**result)
 
-    def build_query(
-        self, query_text: str, message_history: Optional[list[LLMMessage]] = None
+    def _build_query(
+        self,
+        query_text: str,
+        message_history: Optional[List[LLMMessage]] = None,
     ) -> str:
         summary_system_message = "You are a summarization assistant. Summarize the given text in no more than 300 words."
         if message_history:
-            summarization_prompt = self.chat_summary_prompt(
+            summarization_prompt = self._chat_summary_prompt(
                 message_history=message_history
             )
             summary = self.llm.invoke(
@@ -162,10 +168,9 @@ def build_query(
             return self.conversation_prompt(summary=summary, current_query=query_text)
         return query_text
 
-    def chat_summary_prompt(self, message_history: list[LLMMessage]) -> str:
+    def _chat_summary_prompt(self, message_history: List[LLMMessage]) -> str:
         message_list = [
-            ": ".join([f"{value}" for _, value in message.items()])
-            for message in message_history
+            f"{message['role']}: {message['content']}" for message in message_history
         ]
         history = "\n".join(message_list)
         return f"""