chatsorter 0.1.0

Memory processing system for chat applications with fact extraction and summarization

ChatSorter

Memory processing system for chat applications with automatic fact extraction, summarization, and intelligent memory retrieval.

Installation

From PyPI

pip install chatsorter

From Source (Development)

git clone https://github.com/yourusername/chatsorter.git
cd chatsorter
pip install -e .

Quick Start

Python API

from chatsorter import ChatSorter

# Initialize with your API key
cs = ChatSorter(api_key="cs_demo_xxxx")

# Remember a message
cs.remember(chat_id="user123", message="Hi, I'm John from Boston working at TechCorp")

# Recall memories
memories = cs.recall(chat_id="user123", query="where does John work?")
print(memories['memories'])  # Related memories from conversation history

Command Line

# Remember a message
chatsorter --api-key cs_demo_xxxx "Hi, I'm John from Boston" --chat-id user123

# Recall memories
chatsorter --api-key cs_demo_xxxx --recall "where does John work" --chat-id user123

# Check API health
chatsorter --api-key cs_demo_xxxx --health

Command Line Interface

# Process a message
chatsorter "Hello, my name is Alice" --chat-id alice_chat

# Search memories
chatsorter --search "what is my name" --chat-id alice_chat

# Show all facts
chatsorter --facts --chat-id alice_chat

# Clear chat
chatsorter --clear --chat-id alice_chat

Features

• Automatic Fact Extraction: Extracts names, locations, jobs, preferences, and more
• Memory Summarization: Creates narrative summaries of conversations
• Intelligent Search: Vector-based semantic search through conversation history
• Importance Scoring: Rates message importance for better memory management
• Multi-Chat Support: Handle multiple conversations simultaneously
• Background Processing: Asynchronous extraction and summarization

API Reference

ChatSorter Class

__init__(data_dir="data", api_key=None, company_id=None, user_id="default_user")

Initialize ChatSorter instance.

process(message, chat_id, retrieve_memories=True, clear_chat=False)

Process a message and return extracted information.

Returns:

{
    "facts": {...},        # Extracted facts
    "importance": 0.0,     # Importance score
    "memories": [...],     # Related memories
    "response": "...",     # Processing status
    "analysis": {...}      # Additional metadata
}

search(query, chat_id, max_results=5)

Search for memories related to a query.

get_facts(chat_id)

Get all extracted facts for a conversation.

clear_chat(chat_id)

Clear all memories and facts for a conversation.

Configuration

Set these environment variables:

• OPENAI_API_KEY: Your OpenAI API key (for LLM-powered features)
• OLLAMA_BASE_URL: Ollama server URL (optional, for local LLM)

Development

Local Development

pip install -r requirements.txt
python api.py  # Start API server

Testing

python example.py  # Run example script
python quicktest.py  # Run diagnostic tests

Building Package

python -m build
pip install dist/chatsorter-0.1.0.tar.gz

License

MIT License