Add functionality to add more identifiers to GeneProducts in polish

[GwennyGit]

Feature:
Extend the polish module to add KEGG gene, UniProt as well as RefSeq identifiers.

Implementation:

• For bacterial strains where the genome is in the KEGG database:
  • User needs to provide organism code of strain in KEGG through the config.yaml(E.g. For Staphylococcus haemolyticus JCSC1435 this would be sha.)
    → From this code & the locus tags of the model the KEGG gene identifier can be obtained
  • Access KEGG API to retrieve UniProt identifiers
  • Retrieve ncbiprotein identifier additionally from KEGG & Compare ncbiprotein identifier from KEGG with the one in the model
    → No mapping from KEGG Gene to NCBI Protein identifier possible
  • Add all identifiers to the model
• Add RefSeq identifiers from the Reference Sequence .gff file (Findable at the Genome Assembly site of NCBI, For example implementation see: get_ids_from_gff)
• Maybe, also get the name & locus tag from the .gff file 🤔 → Not possible for lab_strain
• Get UniProt identifiers with RefSeq identifiers
  → Result of u = UniProt(), u.mapping('RefSeq_Protein', 'UniProtKB', refseq_id) yields no feasible results. Comparing this result with the result from the KEGG mapping for a protein of Staphylococcus haemolyticus JCSC1435 showed that the resulting identifier is not the same.

[famosab]

We already stumbled upon this in another issue (relates to #52), I think that working with the stuff that is implemented in genecomp might already cover many parts that we have here.

[famosab]

Overview on Identifiers of interest:

• Genbank: Locus tag & NCBI Protein ID
• RefSeq: Locus tag, Old locus tag (== Genbank locus tag, not always present) & RefSeq ID
• Genbank GFF == GFF of e.g. chromosome accession website -> this will hold all info on chromosome and plasmids

[GwennyGit]

If cv_ncbiprotein is adjusted to retrieve identifiers from multiple databases or another function is created for that the identifiers could be stored in a dictionary or table mapping to the database and add_curie_set could also be used here.

[GwennyGit]

As I had already written a script to enhance the GeneProduct annotations here I decided not to create a table for the annotation enhancement but integrated my script into refineGEMs. The adjustments were tested on two VMH models created with AGORA2 and three models which were previously polished with refineGEMs' former polish function. The results look good.

[GwennyGit]

The changes described here are all merged into dev-2 which is the branch for refineGEMs 2.0.0. Thus, this issue can be closed.