vash02/comp-model-mcp-server

Physics Model MCP Server

A modular platform for defining, generating, executing, and tracking physics model experiments with reproducibility.

🚀 Quick Start

1. Clone & install dependencies

   git clone https://github.com/your-org/comp-model-mcp-server.git
   cd comp-model-mcp-server
   pip install -r requirements.txt
   

2. Set environment variables

   export OPENAI_API_KEY="your-openai-api-key"
   export DB_PATH="mcp_server.db"
   export GENERATED_MODELS_DIR="generated_models"
   export SANDBOX_TIMEOUT=30
   

3. Run the server

   uvicorn main_server:app --reload
   

🔁 Reproduce an Experiment

1. Create a Model

Submit a model specification:

curl -X POST http://localhost:8000/models \
  -H "Content-Type: application/json" \
  -d '{
    "model_name": "lorenz",
    "equations": ["dx/dt = sigma*(y - x)", "dy/dt = x*(rho - z) - y", "dz/dt = x*y - beta*z"],
    "parameters": [
      {"name":"sigma","description":"Prandtl number","value":"10.0"},
      {"name":"rho","description":"Rayleigh number","value":"28.0"},
      {"name":"beta","description":"Geometric factor","value":"8.0/3.0"}
    ],
    "initial_conditions":[
      {"name":"x0","description":"initial x","value":"1.0"},
      {"name":"y0","description":"initial y","value":"1.0"},
      {"name":"z0","description":"initial z","value":"1.0"}
    ]
}'

Save the returned model_id.

2. Approve the Model

curl -X POST http://localhost:8000/models/{model_id}/approve

3. Run a Single Experiment

curl -X 'POST' \
  'http://127.0.0.1:8000/models/lorenz_1ac9a383/run' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "parameters": [
    { "name": "sigma", "value": 10.0 },
    { "name": "rho",   "value": 28.0 },
    { "name": "beta",  "value": 2.6666666666666665 }
  ],
  "initial_conditions": [
    { "name": "x0", "value": 1.0 },
    { "name": "y0", "value": 1.0 },
    { "name": "z0", "value": 1.0 }
  ]
}
'

Response includes experiment results.

4. Retrieve Experiment Results

curl http://localhost:8000/models/{model_id}/results

🧬 CLI-Based Experiment Reproduction

1. Fetch script

   curl http://localhost:8000/models/{model_id}/code -o model.py
   

2. Run the script locally

   python model.py --params '{
     "params":{"sigma":10.0,"rho":28.0,"beta":2.6667},
     "initial_conditions":{"x0":1.0,"y0":1.0,"z0":1.0}
   }'
   

   The output will match the experiment results.

🛠️ Cleanup (Optional)

Delete the model and all associated experiments:

curl -X DELETE http://localhost:8000/models/{model_id}

🤖 Embeddings & Analytics

Stored models and experiments include:

• Text embeddings of model specification
• Numeric + text joint embeddings of experiment parameters/results

These can be used for clustering, similarity search, or advanced analytics.

📦 Requirements

• Python 3.10+
• Dependencies in requirements.txt:
  • fastapi, uvicorn, openai, sentence-transformers, torch, numpy, scipy, sympy, etc.

To ensure reproducibility, use the same environment setup and input parameters.