More options and enhancements for gap-filling

[cb-Hades]

Add more options to enable fine-grained gap-filling.

Ideas:

• Better handling of compartments / case, when database does not have information about the compartment of a metabolite / reaction -> maybe sth for a classifier/ML tool?

• Enable the parameter reac_str for build_reaction_bigg to allow building a reaction from a string only without an ID (mainly for manual curation or similar

• Option to add either all new content from gap analysis to the model or restricted by namespace

• a type of template gap filling that utilises a genome and a template model to extract information of homologous genes and reactions to improve the current model

• A ManualGapFiller that is initialised with tables containing information on missing genes and reactions

  • Tables should be similar to those used in all GapFiller classes.
    -> Maybe add getter & setter for missing_reactions & missing_genes?

• Extension of the KEGGapfiller to also accept and use a different organism from KEGG as reference 🤔

• Add a check similar to verifyGapfilledReactions to GapFiller.fill_model

• Implement an option for the GeneGapFiller to use ML to predict EC numbers
  -> Already looked at tools:

  Tool	Usage	GitHub
  DeepECTransformer	Not installable, no license	https://github.com/kaistsystemsbiology/DeepProZyme
  ECRECer Web service	Does not work/return	https://github.com/kingstdio/ECRECer
  CLEAN Web service	Does not work out of the box	https://github.com/tttianhao/CLEAN

• Implement charge addition via ChEBI and/or BiGG -> See idea in dev.old_funcs.add_charges_chemical_formulae_to_metabs

---

Notes:

• when starting any of the ideas above, fell free to open a sub-issue or similar
• adding GPRs only works using the OR connection (currently)