MCP_hack

This project is a collection of tools and experiments for molecular structure prediction and visualization using various AI and computational methods.

Project Structure

The project consists of four main components:

1. Chai1 Test (chai1_test/)

A Python module for molecular structure prediction using the Chai model. This component:

• Processes molecular structure data
• Performs predictions using pre-trained models
• Generates CIF files for structural visualization
• Uses configuration files for model parameters and inference settings

2. Modal Test (modal_test/)

A test implementation using Modal.com for cloud computation:

• Provides serverless computation capabilities
• Includes utility functions for numerical operations
• Demonstrates Modal.com integration patterns

Run it with

modal run square_root.py

3. Molecular Visualization Notebooks (notebooks_molviewspec/)

Jupyter notebooks for molecular structure visualization:

• Contains example notebooks for structure visualization
• Includes sample data (1CBS protein structure)
• Demonstrates various visualization techniques and annotations
• Uses common molecular structure file formats (CIF, BCIF)

4. Streamlit Interface (streamlit_test/)

A web interface built with Streamlit for:

• Interactive molecular structure visualization
• File upload and processing
• Results visualization and analysis

5. Gradio App

A web interface built with Gradio for:

• Interactive molecular structure visualization
• File upload and processing
• Results visualization and analysis

Run it with

gradio app_molecule3d.py

Setup and Installation

Each component has its own pyproject.toml file for dependency management. To set up any component:

1. Navigate to the component directory
2. If necessary install uv with instructions here: https://docs.astral.sh/uv/getting-started/installation/

curl -LsSf https://astral.sh/uv/install.sh | sh

3. Create a virtual environment: You can create one by folder for instance.

   uv venv 
   source .venv/bin/activate
   uv pip install gradio[mcp] # Works only on a bash shell and not on zsh

4. Create a token to access the Modal API

python3 -m modal setup

See description here: https://modal.com/apps/charlotte-chaps/main This command will open a Modal page to create a API token.

5. Note for a new component creation Start by running the following command to create your pyproject.tomlfile:

uv init

Usage

Please refer to the README files in each component directory for specific usage instructions and examples.

File Structure

.
├── chai1_test/          # Molecular structure prediction
├── modal_test/          # Cloud computation integration
├── notebooks_molviewspec/ # Visualization notebooks
└── streamlit_test/      # Web interface

Contributing

When contributing to this repository, please first discuss the change you wish to make via issue, email, or any other method with the owners of this repository.

License

[Your chosen license]