convez.ai
PluginDados & AnalyticsAvançado
jettyio

MLC Bakery

jettyio/mlcbakery

Proveniência e linhagem de modelos ML com API REST

Ver no GitHub
InicialPluginScore · de 0 a 100

Serviço Python para gerenciar proveniência e linhagem de modelos ML, construído com FastAPI e SQLAlchemy. Oferece gestão de datasets, rastreamento de entidades, logging de atividades e validação de metadados Croissant através de API RESTful.

Instalação

$npx convez-ai jettyio/mlcbakery

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
7
Forks
2
Acompanhando
2
Saúde
25%
Último commit
há 8 meses
Versão
0.1.3
Issues abertas
0
Contribuidores
6

Casos de uso

  • Rastreamento de proveniência e linhagem em pipelines ML
  • Gestão centralizada de datasets com metadados Croissant
  • Auditoria e logging de atividades em experimentos de ML
  • Documentação automática de relacionamentos entre artefatos de dados

Veredito

Ferramenta robusta para times que precisam rastrear proveniência em pipelines ML complexos. Requer conhecimento técnico em Python e infraestrutura (Docker, PostgreSQL), mas oferece solução madura com validação Croissant integrada.

README

A tradução para o português deste README ainda está na fila.

MLC Bakery

A Python-based service for managing ML model provenance and lineage, built with FastAPI and SQLAlchemy. Support for Croissant metadata validation.

Features

  • Dataset management with collection support
  • Entity tracking
  • Activity logging
  • Provenance relationships tracking
  • RESTful API endpoints

Running with Docker

  1. Set up Environment Variables: Create a .env file in the project root by copying the example:

    cp env.example .env
    
  2. Start docker containers: The bakery relies on a postgres database and Typesense for search. The MCP server makes REST calls to the API server, which then calls the persistence layer.

    docker compose up -d
    
  3. Run Database Migrations: Apply the latest database schema using Alembic. uv run executes commands within the project's managed environment.

    docker compose exec db psql -U postgres -c "create DATABASE mlcbakery;"
    docker compose exec api alembic upgrade head
    

Access the bakery

By default, the API will be available on localhost.

  • Swagger UI: http://bakery.localhost/docs
  • ReDoc: http://bakery.localhost/redoc
  • Streamable MCP HTTP: http://mcp.localhost/mcp (you may need to add this to your /etc/hosts for local development)

Running the Server (Locally)

Prerequisites

  • Python 3.12+
  • uv (Python package manager)

Development steps

  1. Clone the repository:

    git clone git@github.com:jettyio/mlcbakery.git
    cd mlcbakery
    
  2. Install Dependencies: uv uses pyproject.toml to manage dependencies. It will automatically create a virtual environment if one doesn't exist.

    curl -LsSf https://astral.sh/uv/install.sh | sh
    
    pip install poetry uvicorn
    uv run poetry install --no-interaction --no-ansi --no-root --with mcp
    

Start the FastAPI application using uvicorn:

# Make sure your .env file is present for the DATABASE_URL
uv run uvicorn mlcbakery.main:app --reload --host 0.0.0.0 --port 8000

Authentication

The Bakery is setup to authenticate requests with two methods: JWT Tokens and a "Master Admin Token". Both are configured in the ENV variables (.env file). Both JWT tokens and the Master Admin Token should be provided as "Bearer" Authorization header values.

  • ADMIN_AUTH_TOKEN: A fixed value that is the token a user would need to provide to have admin permissions (unrestricted access to all resources).
  • JWT_VERIFICATION_STRATEGY: The URL of a trusted JWT token issuer, such as Clerk. We have a development instance of Clerk running that you can use. You can sign up for an account via flows.jetty.io (alpha), or contact dev@jetty.io for access to Jetty's Cloud.

Running Tests

The tests are configured to run against a PostgreSQL database defined by the DATABASE_URL environment variable. You can use the same database as your development environment or configure a separate test database in your .env file if preferred (adjust connection string as needed).

# Ensure DATABASE_URL is set in your environment or .env file
uv run pytest
Ver o original no GitHub

Bom para

Relacionados