Add jq to dev dependency

[joverlee521]

Context

Prompted by #1384 (comment)

Improve testing experience by using jq to parse JSON files in functional tests, e.g.

augur/tests/functional/refine/cram/year-bounds.t

Lines 41 to 44 in d2ef346

	TODO: Using jq woud be cleaner, but requires an extra dev dependency.
	$ python3 -c 'import json, sys; print(json.load(sys.stdin)["nodes"]["PAN/CDC_259359_V1_V3/2015"]["date"])' < branch_lengths.json
	2020-12-31

[jameshadfield]

It'd be great to have jq available in our environments. For me it's use is severely limited by its inability to parse large trees. As a tool specifically for "testing experience" this might not matter, but if we are to start including it in our environments I think it'd be the right time to fix that bug (by contibuting a patch to expose parsing depth as an argument, or maintaining a conda recipe with an incresed depth, or some other way).

[victorlin]

Do you mean in our Nextstrain runtimes or dev environments? It's already available in the managed runtimes (docker, conda).

For the Augur dev environment, I don't see inability to parse large trees as a blocker. I think there are few (if any) tests with trees large enough to break jq, and for those cases, whatever's being used right now can stay as-is.