[Bugfix] Druid run_query dimensions part 3 + Unit tests

[Mogball]

#3920 (comment)

Hopefully this'll be the last one. Added some unit tests to try to save some face 🥇

[mistercrunch]

this is a python gotcha:
http://docs.python-guide.org/en/latest/writing/gotchas/#mutable-default-arguments,
just set it to None here and go with extras = extras or {} in the function body

[Mogball]

👍

[ghulands]

Just seeing if you tested this with a time series chart too? That too was still broken after #3920.

[Mogball]

Yes I checked it with that. The test case is the one where timeseries_limit is specified

[ghulands]

I just tested this and it fails for the time series charts. I'm using a column in the group by having the same setup as I reported in #3943. Using a data table view works fine.

Traceback (most recent call last):
  File "/app/.heroku/python/lib/python2.7/site-packages/superset/viz.py", line 275, in get_payload
    df = self.get_df()
  File "/app/.heroku/python/lib/python2.7/site-packages/superset/viz.py", line 97, in get_df
    self.results = self.datasource.query(query_obj)
  File "/app/.heroku/python/lib/python2.7/site-packages/superset/connectors/druid/models.py", line 1078, in query
    client=client, query_obj=query_obj, phase=2)
  File "/app/.heroku/python/lib/python2.7/site-packages/superset/connectors/druid/models.py", line 865, in get_query_str
    return self.run_query(client=client, phase=phase, **query_obj)
  File "/app/.heroku/python/lib/python2.7/site-packages/superset/connectors/druid/models.py", line 1056, in run_query
    filters,
  File "/app/.heroku/python/lib/python2.7/site-packages/superset/connectors/druid/models.py", line 874, in _add_filter_from_pre_query_data
    f = Dimension(dim) == row[dim]
  File "/app/.heroku/python/lib/python2.7/site-packages/pandas/core/series.py", line 642, in __getitem__
    return self._get_with(key)
  File "/app/.heroku/python/lib/python2.7/site-packages/pandas/core/series.py", line 683, in _get_with
    return self.loc[key]
  File "/app/.heroku/python/lib/python2.7/site-packages/pandas/core/indexing.py", line 1328, in __getitem__
    return self._getitem_axis(key, axis=0)
  File "/app/.heroku/python/lib/python2.7/site-packages/pandas/core/indexing.py", line 1541, in _getitem_axis
    return self._getitem_iterable(key, axis=axis)
  File "/app/.heroku/python/lib/python2.7/site-packages/pandas/core/indexing.py", line 1081, in _getitem_iterable
    self._has_valid_type(key, axis)
  File "/app/.heroku/python/lib/python2.7/site-packages/pandas/core/indexing.py", line 1418, in _has_valid_type
    (key, self.obj._get_axis_name(axis)))
KeyError: "None of [[u'outputName', u'outputType', u'type', u'extractionFn', u'dimension']] are in the [index]"

[Mogball]

Looks like it's an issue with _add_filter_from_pre_query_data, which doesn't properly handle dimensions specs (and is called for timeseries). I think that's a bit outside the scope of this fix. And I'm not really familiar with PyDruid.