Change return type from features.query.

[tkaemming]

This more closely matches the upcoming API response, and reflects how people will likely actually use this function.

Before:

>>> features.query(Group.objects.get(id=550))
{'exception:message:character-shingles': [(550, 1.0)],
 'exception:stacktrace:application-chunks': [(550, 1.0)],
 'exception:stacktrace:pairs': [(550, 1.0),
  (354, 0.225),
  (355, 0.225),
  (356, 0.225),
  …

After:

>>> features.query(Group.objects.get(id=550))                                                                                                             
[(550,
  {'exception:message:character-shingles': 1.0,
   'exception:stacktrace:application-chunks': 1.0,
   'exception:stacktrace:pairs': 1.0,
   'message:message:character-shingles': 1.0}),
 (385, {'exception:stacktrace:pairs': 0.225}),
 (394, {'exception:stacktrace:pairs': 0.225}),
 …

[tkaemming]

This is pretty naive, but should work well enough until we decide what kind of default weighting makes sense.

[getsentry-bot]

	1 Warning
⚠️	Changes require @getsentry/security sign-off

Security concerns found

• src/sentry/similarity.py

Generated by 🚫 danger

[JTCunning]

I like it.