refactor how filters are handled

[mcannon068nw]

I'm not sure filters are currently working anymore. Additionally there is some ambiguity on proper way to implement this. May want to revisit how the query string is built as a related issue.

[jsstevenson]

#29 cleans up some of the queries/params, but I do think that some (not all) of the filters have a separate problem (copied from another issue)

I think the problem is with our resolvers, not with the queries or anything we're doing here, but I believe that filters like sourceDbName on interactions are fetching all interactions for genes that have an interaction provided by that source -- rather than fetching just those interactions provided by that source. Ditto pmid and interactionTypes.

I would lean towards removing them for now for this reason -- because they aren't super helpful like this.

[jsstevenson]

Depends on dgidb/dgidb-v5#490