Model Registry using PostgreSQL for ML Experiments

This project provides a model registry for machine learning experiments using FastAPI and PostgreSQL. It allows you to manage projects, models, and experiments efficiently.

Features

• Projects: Create and manage projects.
• Models: Add and retrieve models associated with projects.
• Experiments: Track experiments, including parameters, datasets, evaluations, and artifact file paths.

Endpoints

Projects

• GET /projects/all: Retrieve all projects.
• GET /projects/id/{project_id}: Retrieve a project by its ID.
• GET /projects/name/{project_name}: Retrieve a project by its name.
• POST /projects: Add a new project.

Models

• GET /models/all: Retrieve all models.
• GET /models/id/{model_id}: Retrieve a model by its ID.
• GET /models/project-id/{project_id}: Retrieve models by project ID.
• GET /models/name/{model_name}: Retrieve models by name.
• POST /models: Add a new model.

Experiments

• GET /experiments/all: Retrieve all experiments.
• GET /experiments/id/{experiment_id}: Retrieve an experiment by its ID.
• GET /experiments/model-version-id/{model_version_id}: Retrieve experiments by model version ID.
• GET /experiments/model-id/{model_id}: Retrieve experiments by model ID.
• GET /experiments/project-id/{project_id}: Retrieve experiments by project ID.
• POST /experiments: Add a new experiment.
• POST /experiments/evaluations/{experiment_id}: Update experiment evaluations.
• POST /experiments/artifact-file-paths/{experiment_id}: Update experiment artifact file paths.

Setup

Environment Variables

1. Create a .env file with the following content from the .env.example

2. Build the containers:

make build

3. Start the containers:

make up

3. [Optionally] push the containers to your registry:

make push

4. Shut down the containers

make down