This project is a FastAPI application for evaluating GenAI outputs with a simple web UI and API endpoints.
VerifyAI/
├── app/ # Main application source code
│ ├── main.py # FastAPI entry point
│ ├── static/ # Static files (frontend)
│ │ └── index.html # Main UI page
│ ├── evaluations/ # Evaluation logic
│ │ └── check_sentiment.py # Sentiment evaluation
│ ├── models/ # Model wrappers/utilities
│ │ └── ollama_chat.py # Example: Ollama chat model
│ └── utils/ # Utility functions
│ └── get_cosine_similarity_score.py
├── tests/ # Test files
│ ├── api/ # API tests
│ │ └── test_users_api.py
│ ├── ui/ # UI tests
│ │ ├── page_objects/
│ │ └── test_login.py
│ ├── data/ # Test data
│ │ ├── global_data.yml
│ │ ├── static_data.yml
│ │ └── dynamic_data.yml
│ └── conftest.py # Pytest fixtures/config
├── requirements.txt # Project dependencies
└── pyproject.toml # Project metadata (optional)
-
Clone the repository:
git clone <repository-url> cd VerifyAI
-
Create and activate a virtual environment (recommended):
python3 -m venv .venv source .venv/bin/activate
-
Install dependencies:
pip install -r requirements.txt
-
Run the application:
uvicorn app.main:app --reload
- Access the web UI at: http://localhost:8000/
- API documentation: http://localhost:8000/docs
To run all tests:
pytest
This will execute all tests in the tests/
directory.
Note:
- The UI is served from
app/static/index.html
. - Update
requirements.txt
if you add new dependencies.