load_obo parser needs some specific parsing on definition field

[newgene]

Here is an example from NCIT API:

https://biothings.ncats.io/ncit/node/NCIT:C47977?fields=def

{
    "_id": "NCIT:C47977",
    "_version": 1,
    "def": "\"The hydrochloride salt form of clindamycin, a semi-synthetic, chlorinated broad spectrum antibiotic produced by chemical modification of lincomycin. Clindamycin hydrochloride is used as a solid in capsules.\" []"
}

Minimally, we should trim off the trailing [] and the extra double quotes. We resolved the similar issue in MyChem.info before, see biothings/mychem.info#107.

This is likely not a ncit-specific issue, we might want to fix it at our generic obo parser at hub.dataload.data_parsers.load_obo.

In annotator.py, we currently do an extra clean up (48295bd), which can be removed after the source values are fixed.