Validation of SBML-comp models

[Daniel Weindl]

Validation of an SBML-comp model (SBML semantic test suite case 01161) does not give any validation errors/warnings initially, but after applying CompFlatteningConverter it does. The warnings seem unrelated to flattening the model. Why aren't they shown before? Are those checks skipped if comp is used?

To reproduce via python-libsbml:

import libsbml
import requests

url = "https://raw.githubusercontent.com/sbmlteam/sbml-test-suite/release/cases/semantic/01161/01161-sbml-l3v2.xml"
r = requests.get(url, allow_redirects=True)

sbml_reader = libsbml.SBMLReader()
sbml_doc = sbml_reader.readSBMLFromString(r.content.decode("utf-8"))
assert sbml_doc.validateSBML() == 0
for i_error in range(sbml_doc.getNumErrors()):
    error = sbml_doc.getError(i_error)
    print(f'{error.getCategoryAsString()} ({error.getSeverityAsString()}): {error.getMessage()}')

conversion_properties = libsbml.ConversionProperties()
conversion_properties.addOption("flatten comp", True)
conversion_properties.addOption("leave_ports", False)
conversion_properties.addOption("abortIfUnflattenable", "none")
assert sbml_doc.convert(conversion_properties) == libsbml.LIBSBML_OPERATION_SUCCESS
str_model = libsbml.SBMLWriter().writeSBMLToString(sbml_doc)

print(str_model)

n_errors = sbml_doc.validateSBML()
for i_error in range(sbml_doc.getNumErrors()):
    error = sbml_doc.getError(i_error)
    print(f'{error.getCategoryAsString()} ({error.getSeverityAsString()}): {error.getMessage()}')

# FAILS
assert n_errors == 0, n_errors

[Sarah Keating]

This is a bug thanks for the report.