Identify test cases for NEO (and autocomplete) builds

[kltm]

Generally, we'd like to have some more objective measure of whether a build of NEO (and the autocomplete/total entity space) has what we think it has. This applies to both the ontology builds and the load into Solr.

As a starting point, we'd like to:

• get a set of identifiers for things we expect to see in the build

Ideally, this would slot into having some script to execute them against a product (owl? solr index?) in a pipeline so that failure can prevent publication

While not strictly NEO, we can start there and get a lot of work done. We can start with identifiers listed in #51 #52 #53.

Tagging @vanaukenk @goodb for feedback.

[kltm]

PR:000050230
PR:000050222
PR:000050061
RGD:1309181
ZFIN:ZDB-LINCRNAG-190911-1
ZFIN:ZDB-GENE-190924-2
ZFIN:ZDB-GENE-200114-3
CHEBI:23367