COMBINE archives¶

A model is rarely the whole story: a study consists of one or more models, the simulation experiments which were run on them, the data and the figures. The COMBINE archive (OMEX) packages all of it into one file with a manifest.xml which says what every entry is.

sbmlutils uses pymetadata for archives; it is a dependency, so nothing extra has to be installed.

Creating an archive¶

from pathlib import Path

from pymetadata.omex import EntryFormat, ManifestEntry, Omex

from sbmlutils.factory import create_model

# create the models
sbml_path = Path("model.xml")
create_model(model=model, filepath=sbml_path)

# package them
omex = Omex()
omex.add_entry(
    entry_path=sbml_path,
    entry=ManifestEntry(
        location="./models/model.xml",
        format=EntryFormat.SBML_L3V1,
        master=True,
    ),
)
omex.to_omex(Path("study.omex"))

location is the path of the entry inside the archive, format is the identifiers.org URI of the format, and master marks the entry a tool should start with.

Reading an archive¶

from pymetadata.omex import Omex

with Omex.from_omex(Path("study.omex")) as omex:
    print(omex.manifest["./models/model.xml"].format)

    for entry in omex.entries_by_format("sbml"):
        print(entry.location, omex.get_path(entry.location))

The context manager removes the temporary directory the archive was extracted into.

Omex.from_url reads an archive directly from a URL, which is how the models of BioModels are fetched, see Reading and writing.

Reports for an archive¶

SBMLDocumentInfo describes a single model. For an archive, iterate the SBML entries and describe each of them, which is what the report does:

from pymetadata.omex import Omex

from sbmlutils.report.sbmlinfo import SBMLDocumentInfo

with Omex.from_omex(Path("study.omex")) as omex:
    for entry in omex.entries_by_format("sbml"):
        info = SBMLDocumentInfo.from_sbml(omex.get_path(entry.location))
        print(entry.location, len(info.to_json()))

Example¶

examples/combine_archive/omex_models.py creates two models, flattens the hierarchical one and packages all three into an archive:

python -m examples.combine_archive.omex_models