"""
Example: Using MemoryGraph with LangChain
This example demonstrates how to use MemoryGraphMemory with LangChain
to create a conversational agent with persistent memory.
Prerequisites:
pip install memorygraphsdk[langchain]
pip install langchain-openai # or your preferred LLM provider
Environment:
export MEMORYGRAPH_API_KEY="mgraph_..."
export OPENAI_API_KEY="sk-..."
"""
import os
from langchain.chains import ConversationChain
from langchain_openai import ChatOpenAI
from memorygraphsdk.integrations.langchain import MemoryGraphMemory
def basic_conversation_example():
"""Basic conversation with memory persistence."""
print("=" * 60)
print("Example 1: Basic Conversation with Memory")
print("=" * 60)
# Get API key from environment
api_key = os.getenv("MEMORYGRAPH_API_KEY")
if not api_key:
print("Error: MEMORYGRAPH_API_KEY environment variable not set")
return
# Create memory with a specific session ID
memory = MemoryGraphMemory(
api_key=api_key,
session_id="example-session-1",
return_messages=False, # Return as string format
)
# Create LLM and chain
llm = ChatOpenAI(temperature=0.7, model="gpt-3.5-turbo")
chain = ConversationChain(llm=llm, memory=memory, verbose=True)
# Have a conversation
print("\n1. First message:")
response = chain.run("Hi, I'm working on a Redis timeout issue in my Python application.")
print(f"Response: {response}\n")
print("2. Follow-up question:")
response = chain.run("What did I just say I was working on?")
print(f"Response: {response}\n")
print("Memory is automatically saved to MemoryGraph!")
print("You can retrieve it in future sessions with the same session_id.\n")
def message_format_example():
"""Example using message format instead of string format."""
print("=" * 60)
print("Example 2: Using Message Format")
print("=" * 60)
api_key = os.getenv("MEMORYGRAPH_API_KEY")
if not api_key:
print("Error: MEMORYGRAPH_API_KEY environment variable not set")
return
# Create memory with return_messages=True
memory = MemoryGraphMemory(
api_key=api_key,
session_id="example-session-2",
return_messages=True, # Return as Message objects
)
llm = ChatOpenAI(temperature=0.7, model="gpt-3.5-turbo")
chain = ConversationChain(llm=llm, memory=memory, verbose=True)
# Have a conversation
print("\n1. First message:")
response = chain.run("What are the benefits of using graph databases?")
print(f"Response: {response}\n")
print("2. Ask about conversation history:")
response = chain.run("What did I ask about?")
print(f"Response: {response}\n")
def persistent_session_example():
"""Demonstrate memory persistence across different script runs."""
print("=" * 60)
print("Example 3: Persistent Sessions")
print("=" * 60)
api_key = os.getenv("MEMORYGRAPH_API_KEY")
if not api_key:
print("Error: MEMORYGRAPH_API_KEY environment variable not set")
return
session_id = "persistent-session-demo"
# Create memory - this will load any existing conversation history
memory = MemoryGraphMemory(
api_key=api_key,
session_id=session_id,
return_messages=False,
)
llm = ChatOpenAI(temperature=0.7, model="gpt-3.5-turbo")
chain = ConversationChain(llm=llm, memory=memory, verbose=True)
print(f"\nUsing session_id: {session_id}")
print("This session persists across runs. Try running this script multiple times!")
# Add new message to the conversation
print("\n1. New message:")
response = chain.run("Remember: My favorite programming language is Python.")
print(f"Response: {response}\n")
print("2. Test memory:")
response = chain.run("What's my favorite programming language?")
print(f"Response: {response}\n")
print(
f"\nNote: To start a fresh conversation, use a different session_id "
f"or change the value of 'persistent-session-demo'"
)
def custom_tags_example():
"""Example showing how to use custom context and tags."""
print("=" * 60)
print("Example 4: Advanced - Direct Memory Access")
print("=" * 60)
api_key = os.getenv("MEMORYGRAPH_API_KEY")
if not api_key:
print("Error: MEMORYGRAPH_API_KEY environment variable not set")
return
# Create memory
memory = MemoryGraphMemory(
api_key=api_key,
session_id="advanced-example",
)
# You can also access the underlying MemoryGraph client directly
# for more advanced operations
print("\nAccessing conversation memories directly:")
memories = memory.client.search_memories(
tags=["session:advanced-example"],
memory_types=["conversation"],
limit=10,
)
if memories:
print(f"\nFound {len(memories)} conversation memories:")
for mem in memories:
role = mem.context.get("role") if mem.context else "unknown"
print(f" - [{role}] {mem.title}")
else:
print("\nNo conversation memories found yet. Run the chain first!")
# You can also create relationships between memories
# or search across sessions
print("\nYou have full access to MemoryGraph features!")
def main():
"""Run all examples."""
print("\n" + "=" * 60)
print("MemoryGraph + LangChain Integration Examples")
print("=" * 60 + "\n")
# Check for required environment variables
if not os.getenv("MEMORYGRAPH_API_KEY"):
print("ERROR: Please set MEMORYGRAPH_API_KEY environment variable")
print("Example: export MEMORYGRAPH_API_KEY='mgraph_...'")
return
if not os.getenv("OPENAI_API_KEY"):
print("ERROR: Please set OPENAI_API_KEY environment variable")
print("Example: export OPENAI_API_KEY='sk-...'")
return
# Run examples
try:
basic_conversation_example()
print("\n" + "-" * 60 + "\n")
message_format_example()
print("\n" + "-" * 60 + "\n")
persistent_session_example()
print("\n" + "-" * 60 + "\n")
custom_tags_example()
print("\n" + "=" * 60)
print("All examples completed!")
print("=" * 60 + "\n")
except Exception as e:
print(f"\nError running examples: {e}")
import traceback
traceback.print_exc()
if __name__ == "__main__":
main()
