Roadmap#

This page outlines the main areas of ongoing and future improvement for pgmpy. It is intended as a contributor-facing roadmap rather than a strict release schedule.

The roadmap is best read as a set of active themes:

  • where the library is being improved,

  • where contributions are most useful,

  • and which parts of the user experience we want to keep strengthening.

Documentation And Learning Resources#

Continue improving the documentation so it is easier to move between:

  • task-oriented guides,

  • runnable quickstart examples,

  • and detailed API reference pages.

This includes clearer workflows, better cross-linking, and stronger example coverage for the core pgmpy feature areas.

Unified APIs Across Workflows#

Keep making the library feel more consistent across:

  • causal discovery,

  • parameter estimation,

  • inference,

  • causal identification and estimation,

  • simulation,

  • and evaluation.

This includes clearer high-level entry points, consistent naming, and better alignment between narrative guides and the public APIs.

Examples, Benchmarks, And Datasets#

Expand the ecosystem of built-in examples, benchmark datasets, and example models so users can evaluate workflows quickly without building everything from scratch.

This includes:

  • better example coverage for major tasks,

  • easier discovery of built-in assets,

  • and stronger benchmark-oriented documentation.

Modeling And Algorithm Coverage#

Continue expanding and refining pgmpy’s support for probabilistic and causal modeling workflows.

This includes ongoing work around:

  • structure learning and causal discovery,

  • parameter estimation,

  • exact and approximate inference,

  • causal reasoning,

  • and support for different model families.

Developer Experience#

Improve the contributor experience by keeping the codebase easier to extend and maintain.

Important themes here include:

  • better extension templates,

  • clearer contribution pathways,

  • stronger automated validation for docs and examples,

  • and consistent public module organization.

How To Contribute Against The Roadmap#

If you want to work on one of these areas:

  1. Start with the Contributing Guide.

  2. Check open issues and pull requests on GitHub.

  3. Open an issue or discussion if you want to propose a roadmap addition or a larger design change.

Useful links: