Missing annotations when converting from SBML3 to SBtab

[smoretti]

Hi

I found several issues when converting from SBML3 to SBtab (or ObjTables) with sbtab_sbml2sbtab.py:

• When several annotations are provided for the same entity (e.g. CHEBI) only the last one is reported in SBtab.

  <bqbiol:isRelatedTo>
  <rdf:Bag>
       <rdf:li rdf:resource="https://identifiers.org/CHEBI:29234"/>
       <rdf:li rdf:resource="https://identifiers.org/CHEBI:5584"/>
  </rdf:Bag>
  </bqbiol:isRelatedTo>

  returns only

  !Identifiers:obo.chebi
  CHEBI:5584

• Long and/or with / identifiers.org annotations look to break the annotation parsing https://identifiers.org/envipath:32de3cf4-e3e6-4168-956e-32fa5ddb0ce1/compound/41e4c903-407f-49f7-bf6b-0a94d39fa3a7 returns

  !Identifiers:envipath:32de3cf4-e3e6-4168-956e-32fa5ddb0ce1/compound
  41e4c903-407f-49f7-bf6b-0a94d39fa3a7

  instead of (expect)

  !Identifiers:envipath
  32de3cf4-e3e6-4168-956e-32fa5ddb0ce1/compound/41e4c903-407f-49f7-bf6b-0a94d39fa3a7

• Not sure the is qualifier is parsed in SBML annotations.

  <bqbiol:is>
  <rdf:Bag>
      <rdf:li rdf:resource="https://identifiers.org/CHEBI:28938"/>
  </rdf:Bag>
  </bqbiol:is>

I joined an SBML example file. bigg_e_coli_core.sbml3.xml.zip

I tested on Linux CentOS 7.4, with Python 3.5.2.

[eladnoor]

@tlubitz - could you take a look at this? These identifiers are important for equilibrator in order to map the compounds in metanetx correctly.

[tlubitz]

@eladnoor I do have it on my list! Alas, my working time right now is reduced to an absolute minimum, due to lockdown