contextmemory 0.1.3

Long-term memory system for AI conversations

ContextMemory

Long-term memory system for AI conversations.

ContextMemory extracts, stores, and retrieves important facts from conversations, enabling AI Agents to remember user preferences, context, and history across sessions.

Installation

pip install contextmemory

Quick Start

1. Configure

Option A: Programmatic (Recommended)

from contextmemory import configure

configure(
    openai_api_key="sk-...",  # Required
    database_url="postgresql://...",  # Optional - defaults to SQLite
)

Option B: Environment Variables

export OPENAI_API_KEY="sk-..."
export DATABASE_URL="postgresql://..."  # Optional

2. Initialize Database

from contextmemory import create_table

create_table()  # Creates all required tables

3. Use Memory System

from contextmemory import Memory, SessionLocal

# Create session and memory instance
db = SessionLocal()
memory = Memory(db)

# Add memories from a conversation
messages = [
    {"role": "user", "content": "Hi, I'm John and I love Python programming"},
    {"role": "assistant", "content": "Nice to meet you John! Python is great."},
]
memory.add(messages=messages, conversation_id=1)

# Search memories
results = memory.search(
    query="What programming language does the user like?",
    conversation_id=1,
    limit=5
)
print(results)
# {'query': '...', 'results': [{'memory_id': 1, 'memory': 'User loves Python programming', 'score': 0.89}]}

# Update a memory
memory.update(memory_id=1, text="User is an expert Python developer")

# Delete a memory
memory.delete(memory_id=1)

Features

• Automatic Memory Extraction: Uses LLM to extract important facts from conversations
• Semantic Search: Find relevant memories using embedding-based similarity
• Memory Deduplication: Automatically updates or removes duplicate memories
• Flexible Storage: SQLite (default) or PostgreSQL
• Easy Configuration: Programmatic or environment variable setup

Configuration Options

Parameter	Required	Default	Description
openai_api_key	Yes	-	Your OpenAI API key
database_url	No	~/.contextmemory/memory.db	Database connection URL
debug	No	False	Enable debug logging

Database Support

• SQLite (default): No additional setup required
• PostgreSQL: Install with pip install contextmemory[postgres]

API Reference

configure(openai_api_key, database_url=None, debug=False)

Initialize the library configuration.

Memory(db: Session)

Main memory interface class.

Methods:

• add(messages: List[dict], conversation_id: int) - Extract and store memories
• search(query: str, conversation_id: int, limit: int) - Search memories
• update(memory_id: int, text: str) - Update a memory
• delete(memory_id: int) - Delete a memory

create_table()

Create all required database tables.

SessionLocal()

Get a new database session.

License

MIT License - see LICENSE file.

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

Links

• GitHub Repository
• Issue Tracker