Can we now normalize all JSON+LD that goes into MongoDB?

[justinccdev]

So, now that the JSON is going into MongoDB, we can do some nice stuff like find the total count of entries which are about samples

db.samples.find({ 'mainEntity.@type': 'Sample'}).count()

However, for this to work, the JSON+LD going in needs to be normalized (and we need to transform RDFa and Microdata to this normalized form as much as possible when we handle those).

@innovationchef I think you looked at this before in the older scraper, using the pyld library. I remember it wasn't possible to use where it wasn't completely happy about the jsonld format. But perhaps we could use it in the short term with biosamples at least. What do you think?