Import fix_evidence too

[hammer]

https://github.com/opentargets/data_pipeline/blob/master/mrtarget/common/EvidenceString.py#L153

Going to be harder because it relies on targets, diseases, and eco.

[hammer]

Did eco_scores, still need to do:

• self.fix_target_id: --> normalise_target_id --> get_uniprot2ensembl, get_reference_ensembl_id. Hot mess I don't wanna handle right now.
• self.fix_disease_id: --> get_ontology_code_from_url
• get_ontology_code_from_url

[hammer]

Turns out get_ontology_code_from_url is easy to import from https://github.com/opentargets/data_pipeline/blob/master/mrtarget/modules/EFO.py#L21, which makes fix_disease_id easy to import too.

[hammer]

• get_uniprot2ensembl defined in https://github.com/opentargets/data_pipeline/blob/master/mrtarget/common/LookupTables.py, which ultimately queries the gene index. This seems to come from the dbxref_extended field in the data sourced from uniprot-uri. It looks like https://www.uniprot.org/docs/dbxref.txt documents how that's populated. Maybe best way to do this mapping is with https://www.uniprot.org/help/download_mapping? Could use the XML or the flat file.
• get_reference_ensembl_id calls out to EvidenceManager._map_to_reference_ensembl_gene

[hammer]

Where does https://github.com/opentargets/data_pipeline/blob/master/mrtarget/resources/genes_with_non_reference_ensembl_ids.tsv come from?! Called in https://github.com/opentargets/data_pipeline/blob/master/mrtarget/common/LookupHelpers.py.