Duplicate 'root' MF annotations in the GPAD should be filtered

[ukemi]

I thought we had fixed this before, but I am seeing it in the 'live' files. In this model, http://noctua.geneontology.org/editor/graph/gomodel:MGI_MGI_1918926? three root annotations were imported from the MGI import.
When I look at the GPAD output on snapshot, I see four annotations. The MF root annotation is duplicated, presumably because it is being generated from the asserted MF annotation and it is being generated from the 'root' BP annotation. I thought @balhoff had made a fix to correct these duplicates from being generated.

MGI MGI:1918926 involved_in GO:0008150 GO_REF:0000015 ECO:0000307 20210426 MGI noctua-model-id=gomodel:MGI_MGI_1918926|contributor=https://orcid.org/0000-0003-2689-5511|model-state=production
MGI MGI:1918926 is_active_in GO:0005575 GO_REF:0000015 ECO:0000307 20210426 MGI noctua-model-id=gomodel:MGI_MGI_1918926|contributor=https://orcid.org/0000-0003-2689-5511|model-state=production
MGI MGI:1918926 enables GO:0003674 GO_REF:0000015 ECO:0000307 20210426 MGI part_of(GO:0008150) noctua-model-id=gomodel:MGI_MGI_1918926|contributor=https://orcid.org/0000-0003-2689-5511|model-state=production
MGI MGI:1918926 enables GO:0003674 GO_REF:0000015 ECO:0000307 20210426 MGI noctua-model-id=gomodel:MGI_MGI_1918926|contributor=https://orcid.org/0000-0003-2689-5511|model-state=production

[balhoff]

@ukemi let's discuss these sometime soon. I feel like the standalone MF annotation is just redundant in the model and should be deleted, or else we should use a different evidence code for these placeholder MFs.