fix(eap): fix bug with data_present

[davidtsuk]

Fixes https://github.com/getsentry/eap-planning/issues/144

Additional Context

When we perform aggregations over attributes, the function is converted into function_nameIf (e.g. countIf) which returns a default value if no rows match the condition. This means that there is no way to distinguish between an aggregate result that is 0 because it's for example a sum of values that add up to 0, and a result that is 0 because no values matched the given condition. To deal with this, we compute the number of events being aggregated even when we aren't extrapolating so we can determine if data was present or not.

[codecov]

❌ 1 Tests Failed:

Tests completed	Failed	Passed	Skipped
2741	1	2740	6

View the top 1 failed tests by shortest run time

tests.web.rpc.v1.test_endpoint_time_series.test_endpoint_time_series.TestTimeSeriesApi::test_with_no_data_present

Stack Traces | 0.277s run time

Traceback (most recent call last):
  File ".../v1/test_endpoint_time_series/test_endpoint_time_series.py", line 417, in test_with_no_data_present
    assert sorted(response.result_timeseries, key=lambda x: x.label) == [
AssertionError: assert [label: "avg"\nbuckets {\n  seconds: 1736236800\n}\nbuckets {\n  seconds: 1736237100\n}\nbuckets {\n  seconds: 1736237400\n}\nbuckets {\n  seconds: 1736237700\n}\nbuckets {\n  seconds: 1736238000\n}\nbuckets {\n  seconds: 1736238300\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\n,\n label: "sum"\nbuckets {\n  seconds: 1736236800\n}\nbuckets {\n  seconds: 1736237100\n}\nbuckets {\n  seconds: 1736237400\n}\nbuckets {\n  seconds: 1736237700\n}\nbuckets {\n  seconds: 1736238000\n}\nbuckets {\n  seconds: 1736238300\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\n] == [label: "avg"\nbuckets {\n  seconds: 1736236800\n}\nbuckets {\n  seconds: 1736237100\n}\nbuckets {\n  seconds: 1736237400\n}\nbuckets {\n  seconds: 1736237700\n}\nbuckets {\n  seconds: 1736238000\n}\nbuckets {\n  seconds: 1736238300\n}\ndata_points {\n  data: 1\n  data_present: true\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\n,\n label: "sum"\nbuckets {\n  seconds: 1736236800\n}\nbuckets {\n  seconds: 1736237100\n}\nbuckets {\n  seconds: 1736237400\n}\nbuckets {\n  seconds: 1736237700\n}\nbuckets {\n  seconds: 1736238000\n}\nbuckets {\n  seconds: 1736238300\n}\ndata_points {\n  data: 1\n  data_present: true\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\n]
  At index 0 diff: label: "avg"\nbuckets {\n  seconds: 1736236800\n}\nbuckets {\n  seconds: 1736237100\n}\nbuckets {\n  seconds: 1736237400\n}\nbuckets {\n  seconds: 1736237700\n}\nbuckets {\n  seconds: 1736238000\n}\nbuckets {\n  seconds: 1736238300\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\n != label: "avg"\nbuckets {\n  seconds: 1736236800\n}\nbuckets {\n  seconds: 1736237100\n}\nbuckets {\n  seconds: 1736237400\n}\nbuckets {\n  seconds: 1736237700\n}\nbuckets {\n  seconds: 1736238000\n}\nbuckets {\n  seconds: 1736238300\n}\ndata_points {\n  data: 1\n  data_present: true\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\ndata_points {\n}\n
  Full diff:
    [
     label: "avg"
    buckets {
      seconds: 1736236800
    }
    buckets {
      seconds: 1736237100
    }
    buckets {
      seconds: 1736237400
    }
    buckets {
      seconds: 1736237700
    }
    buckets {
      seconds: 1736238000
    }
    buckets {
      seconds: 1736238300
    }
    data_points {
  -   data: 1
  -   data_present: true
    }
    data_points {
    }
    data_points {
    }
    data_points {
    }
    data_points {
    }
    data_points {
    }
    ,
     label: "sum"
    buckets {
      seconds: 1736236800
    }
    buckets {
      seconds: 1736237100
    }
    buckets {
      seconds: 1736237400
    }
    buckets {
      seconds: 1736237700
    }
    buckets {
      seconds: 1736238000
    }
    buckets {
      seconds: 1736238300
    }
    data_points {
  -   data: 1
  -   data_present: true
    }
    data_points {
    }
    data_points {
    }
    data_points {
    }
    data_points {
    }
    data_points {
    }
    ,
    ]

To view more test analytics, go to the Test Analytics Dashboard
📢 Thoughts on this report? Let us know!

[onkar]

LGTM

[onkar]

nit: should this be renamed for is_data_present since it returns an answer to this question?

[onkar]

I am wondering if this return can be collapsed with the return on L130 for better readability. All we are checking is if confidence_interval is None and if so, we pass None.

[davidtsuk]

yeah you are right, it can be collapsed, nice catch