Skip to content

CHANGELOG

All notable changes to the mlflow-assistant package will be documented in this file.

This project adheres to Semantic Versioning and follows changelog conventions inspired by Keep a Changelog.

[v0.1.5] – 2025-09-18

Added

  • Providers This allows users to interact with different LLM services (OpenAI, Ollama and Databricks).
  • Agent Workflow Replaced the mocked answer for an answer generated by a LangGraph agent that connects to MLflow Server via tools.

[v0.1.4] – 2025-07-15

Added

  • Setup This allows users to setup mlflow-assistant.
  • Start This allows users to start the session to interact with mlflow interactively.

Notes

  • Start functionality is currently limited as it is calling a mock function

[v0.1.3] – 2025-07-15

Fixed

  • Fixed navigation structure in mkdocs.yml to correctly point to reference/ directory instead of reference/SUMMARY.md
  • Improved compatibility with literate-nav and section-index plugins for proper code reference navigation

[v0.1.2] – 2025-07-15

Added

  • Comprehensive documentation system with MkDocs Material theme
  • Versioned documentation deployment using mike for GitHub Pages
  • Automated documentation generation from code docstrings using mkdocstrings
  • Code reference navigation with auto-generated API documentation
  • Custom Geist font for modern, clean documentation appearance
  • Integrated documentation deployment in the main CI/CD pipeline

Changed

  • Enhanced CI/CD pipeline to include documentation deployment after successful releases
  • Documentation now automatically deploys on version bumps and PyPI releases

Notes

[v0.1.1] – 2025-07-14

Added

  • Introduced a unified CI/CD pipeline for:
  • Linting with ruff.
  • Type checking with mypy.
  • Test execution with pytest and coverage reporting via pytest-cov.
  • Publishing to PyPI on new Git tags.

Changed

  • Improved test fixture logic for integration tests to correctly check for a live MLflow Tracking Server before execution.

Notes

  • This pipeline simplifies maintenance by consolidating quality checks and publishing into a single workflow, triggered on pushes and releases.

[v0.1.0] – 2025-05-08

Added

  • Initial placeholder release of mlflow-assistant on PyPI.
  • Included a minimal MLflowClient wrapper module.

Notes

  • This version was intentionally published with limited functionality to reserve the mlflow-assistant package name on PyPI.
  • Full development will begin in upcoming versions with a focus on making MLflow easier to use through high-level utilities and assistant-like automation features.