Write integration test that confirms the SPARQL query works against a current graph

[surchs]

#282 is a good example of a PR that looks good but to be completely sure a reviewer would have to

• clone the PR branch
• spin up a local graph
• configure the graph (although that is easier now with [ENH] Add GraphDB setup script recipes#19)
• put data into the local graph
• either spin up the local PR branch directly and connect to the Graph by hand OR
• build a new docker image from the local PR branch, edit the docker-compose.yml and then spin up the local branch API with the graph

That's silly work and a bot should do it. As part of a test. We should be most of the way there already, but the rest is probably still a good bit of work.

The crucial part of the issue entails testing that the query template returns expected results, rather than the content of the query string itself. Specifically, this may requiring construct test queries, and then asserting over responses from the graph.

This would help catch issues such as:

• [BUG] Healthy Control results are incorrect #324
• Subjects missing imaging session excluded from result? #331

[neurobagel-bot]

🚀 Issue was released in v0.4.0 🚀