This repository contains a GitHub workflow that automatically processes new issues using pre-encoded memvid memory and ChatGPT to provide intelligent responses.

1. Memory Creation (done once locally): Run create_memory.py to encode knowledge base into memvid format
2. Issue Processing (automated): When a new issue is opened, the workflow:
   • Reads the issue title and body
   • Loads the pre-created memvid memory files
   • Uses quick_chat to retrieve relevant context from the memory (context-only mode)
   • Calls OpenAI GPT with both issue content and memvid context
   • Posts an automated comment to the issue with the response

Note: The system uses memvid in "context-only" mode for memory retrieval and OpenAI for response generation. You may see harmless warnings about "Google provider not available" - these can be safely ignored.

Before committing to the repository, you need to create the memory files:

# Install memory creation dependencies
pip install -r requirements-memory.txt

# Create the memory files
python create_memory.py

This will create memory.mp4 and memory_index.json files that should be committed to the repository.

Set up the following secret in your GitHub repository:

• OPENAI_API_KEY: Your OpenAI API key for ChatGPT access

The GITHUB_TOKEN is automatically provided by GitHub Actions.

The workflow requires the following permissions:

• issues: write - to post comments on issues
• contents: read - to read repository contents

These are configured in the workflow file.

To update the knowledge base, modify the chunks array in create_memory.py, then re-run the script and commit the updated memory files.

The ChatGPT response format can be customized by modifying the prompt in the get_chatgpt_response() function in memvid_agent.py.

• create_memory.py - Script to create memvid memory files (run locally)
• memvid_agent.py - Main agent that processes issues (runs in GitHub Actions)
• .github/workflows/memvid_agent_workflow.yml - GitHub Actions workflow
• requirements.txt - Python dependencies for the workflow agent
• requirements-memory.txt - Python dependencies for memory creation
• memory.mp4 - Pre-encoded video memory (created by create_memory.py)
• memory_index.json - Memory index file (created by create_memory.py)
• README.md - This documentation

Before deploying to GitHub, you can test the agent locally:

# Run setup script
./setup_test.sh

# Activate virtual environment 
source venv/bin/activate

# Load environment variables from .env file
source .env

1. Workflow Simulation - Test the complete GitHub issue workflow:

   python simulate_workflow.py

2. Interactive Testing - Test with custom queries:

   python interactive_test.py

The testing uses your actual OpenAI API key from the .env file but simulates GitHub posting (no real GitHub API calls are made).

1. Initial Setup:

   pip install -r requirements-memory.txt
   python create_memory.py
   git add memory.mp4 memory_index.json
   git commit -m "Add pre-encoded memory files"
   git push

2. Usage: Once set up, the workflow automatically runs when new issues are opened

3. Updating Knowledge Base:

   # Edit create_memory.py with new knowledge
   python create_memory.py
   git add memory.mp4 memory_index.json
   git commit -m "Update knowledge base"
   git push