TORCHLITE API

Backend API for the TORCHLITE application.

Pre-requisites

1. Python 3.11 or higher.
2. MongoDB (an alternative is to connect to an existing deployed MongoDB instance containing relevant data)
3. Redis (an alternative is to connect to an existing deployed Redis instance)
4. Docker is installed and running (recommended).

Install Dependencies

1. Create a Python virtual environment

   python3 -m venv .venv
   source .venv/bin/activate

2. Install dependencies

   pip install -r requirements.txt
   poetry lock --no-update
   poetry install

Environment Variables

Create a .env file in the root directory and add the following environment variables, and replace the placeholders with the actual values:

KEYCLOAK_REALM=<realm>
TORCHLITE_CLIENT_ID=<keycloak_client_id>
TORCHLITE_CLIENT_SECRET=<ckeycloak_lient_secret>
FEATURED_WORKSETS_URL=<url>
MONGODB_URL=<mongodb_url>
REDIS_HOST=<redis_host>
ENABLE_CACHE=<True/False>
CACHE_EXPIRE=<cache_expiration_time_in_seconds>
REGISTRY_API_URL=<registry_api_url>
FEATURED_WORKSET_USER=<featured_workset_user>

Running the API using Docker Container (Recommended)

1. Build the Docker image

   docker buildx build --no-cache -t torchlite-backend .

2. Run the TORCHLITE Backend using EITHER of the steps below.

   a. Using Docker Run command. This assumes that MongoDB and Redis are running on the host machine.

   docker run --env-file .env -p 8000:8000 torchlite-backend

   b. Using Docker Compose (further testing and updating the instructions may be needed)

   docker-compose up

Running the API using Poetry (Alternative)

1. Run the application

   poetry run uvicorn htrc.torchlite.app:api --reload --log-config log_conf.yaml

API Documentation

Open the API documentation in your browser at http://127.0.0.1:8000/docs.