Skip to main content

Install from Source

This guide will walk you through the process of installing Evalap from source code. Installing from source is recommended for developers who want to contribute to the project or need the latest features that may not be available in the released versions.

Prerequisites​

Before you begin, ensure you have the following installed on your system:

  • Python 3.10 or higher
  • pip (Python package installer)
  • Git

Clone the Repository​

git clone https://github.com/etalab-ia/evalap.git
cd evalap
python -m venv venv
source venv/bin/activate  # On Windows, use: venv\Scripts\activate

Install Dependencies​

pip install .

Configure the Application​

To protect the API sensitive request such as deleting experiment or dataset you can set an admin token

export ADMIN_TOKEN="Your evalap admin token"

You can access LLM models from major providers by setting up your API keys if you have accounts with:

export OPENAI_API_KEY="Your secret key"
export ANTHROPIC_API_KEY="Your secret key"
export MISTRAL_API_KEY="Your secret key"
export ALBERT_API_KEY="Your secret key"

You can also define environment variables in a .env file at the root of the project.

All project global settings and environment variables are handled in evalap/api/config.py.

Database Initialization​

  1. Launch the development services:
docker compose -f compose.dev.yml up
  1. Create the first migration script:
alembic -c evalap/api/alembic.ini revision --autogenerate -m "Table Initialization"
  1. Initialize/Update the database schema:
alembic -c evalap/api/alembic.ini upgrade head

Run the Application​

# Step 1: Run the API server
uvicorn evalap.api.main:app --reload --host 0.0.0.0 --port 8000

# Step 2: In a separate terminal, activate your virtual environment if needed, then run the runner
PYTHONPATH="." python -m evalap.runners

Verify Installation​

To verify that Evalap is running correctly, open your web browser and navigate to:

http://localhost:8000/redoc

You should see the API documentation page. You can also use http://localhost:8000/docs if you prefer the swagger version.

Logging Configuration​

You can adjust the logging level for more detailed output:

# Run with debug logging enabled
LOG_LEVEL="DEBUG" PYTHONPATH="." python -m evalap.runners

Troubleshooting​

If you encounter issues starting the application:

  1. Ensure all dependencies are correctly installed
  2. Verify that the database is running (check Docker containers)
  3. Check that environment variables are properly set
  4. Look for error messages in the terminal output

The API should now be running at http://localhost:8000.

Run the Streamlit Frontend (Optional)​

streamlit run evalap/ui/demo_streamlit/app.py --server.runOnSave true

Next Steps​

Now that you have Evalap installed, you can: