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 asdlml), 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, andvalidate_asset/validate_assets.Command-line interface (
dlml):dlml explorerlaunches the web app;dlml listanddlml infoinspect 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/dlmllayout with the data undersrc/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’sreleases/latestcontinues 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.