Version 3.0 (July 13, 2026)

This release turns the library into an installable Python package and adds a programmatic API, a command-line interface, and the DLML Explorer web app. The model data itself is unchanged and remains backward compatible.

Added

  • Installable package: the library is published to PyPI as simcenter-dlml (imported as dlml), with the model data bundled inside the wheel — no runtime download.

  • Python API (import dlml) for discovering datasets and reading model parameters, metadata, and input schemas: list_datasets, available_collections, get_fragility / get_consequence_repair / get_loss_repair / get_parameters, get_metadata, get_schema, get_file, and validate_asset / validate_assets.

  • Command-line interface (dlml): dlml explorer launches the web app; dlml list and dlml info inspect the packaged datasets.

  • DLML Explorer web app (the simcenter-dlml[explorer] extra) for searching, visualizing, and assembling project-specific model collections.

  • Test suite, continuous integration, and PyPI trusted publishing.

Changed

  • Distribution: the project is now distributed as a pip-installable package in addition to being bundled with Pelicun; the source moved to a src/dlml layout with the data under src/dlml/data.

  • Self-contained: the core library, its API, and the documentation build no longer require Pelicun. The package now bundles the CSV-parsing helper (convert_to_MultiIndex) and the controlled vocabularies it validates against (the recognized demand types and probability-distribution families), so the core install depends only on pandas, numpy, and jsonschema.

Distribution

  • Version 3.0 is distributed through PyPI as simcenter-dlml. The GitHub repository’s releases/latest continues to point to v2.1.0, so existing Pelicun (≤3.9) installations — which download the model data from the latest GitHub release — keep working unchanged. Subsequent v3.x updates are published to PyPI.