Add worker to query classifier data to fit cognoml model

[jessept]

@awm33 I created this so we can track it going forward.

A given classifier task has a list of entrezids and disease types. The worker code will query for any samples that match the list of disease types and join that to the mutations table. The result will not be an a [sample_id,mutation_status] form, so the worker needs to transform it into that form and pass it to the cognoml code.

[dhimmel]

The worker code will query for any samples that match the list of disease types and join that to the mutations table.

I'm not sure the worker code has to do this at all. The frontend will need access to this information for their realtime displays. Therefore, we can potentially outsource/consolidate this computation to the frontend?

[awm33]

@dhimmel

The frontend will need access to this information for their realtime displays.

What realtime displays?

Therefore, we can potentially outsource/consolidate this computation to the frontend?

It's just reformatting some JSON results into [sample_id,mutation_status] , I don't think it will be computationally expensive.

If I can just pass something like [sample_id,mutation_status] or a dict, or whatever singular format is chosen, that's all that's needed. Oh, and the congnoml code should know to skip pulling the mutation statues remotely, since it has been passed in by the worker code.

[dhimmel]

If I can just pass something like [sample_id,mutation_status] or a dict, or whatever singular format

@awm33 --- so your main priority is to pass both sample_id and mutation_status status info in a single parameter? Your problem with the current design is that we force you to split this information into two parameters?

If I'm understanding correctly, I'd then favor parameter called sample_to_mutation_status that consumes a dictionary. The dictionary is nice because it makes clear that the ordering is not important.

[awm33]

@dhimmel Ah, it looks like I can still pass it. An earlier version of the refactor PR expected a dataframe.

Then I think we just need to make get_mutations_df run optionally and we should be good.