Generate more idiomatic Flux in query builder

[chnn]

Closes #13775

Updates the Flux code generated by the query builder to use the aggregateWindow helper when possible. Queries previously generated like so:

from(bucket: "telegraf")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r._measurement == "cpu" or r._measurement == "disk")
  |> filter(fn: (r) => r._field == "used_percent")
  |> window(period: v.windowPeriod)
  |> mean()
  |> group(columns: ["_value", "_time", "_start", "_stop"], mode: "except")
  |> yield(name: "mean")

Will now be generated as:

from(bucket: "telegraf")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r._measurement == "cpu" or r._measurement == "disk")
  |> filter(fn: (r) => r._field == "used_percent")
  |> aggregateWindow(every: v.windowPeriod, fn: mean)
  |> yield(name: "mean")

[ebb-tide]

looks awesome!
any idea how we could keep track of the number type issue nathaniel highlighted? It's certainly going to come up again.

[chnn]

any idea how we could keep track of the number type issue nathaniel highlighted? It's certainly going to come up again.

Yeah not sure... maybe it makes sense to wait until the issue is reported by a user, so we have a more acute description of the problem. Or we could create an issue—my worry is that the issue would be too broad of a problem statement and unactionable.

[nathanielc]

any idea how we could keep track of the number type issue nathaniel highlighted? It's certainly going to come up again.

Eventually I expect this to be something that a Flux language server could expose. Not sure exactly the details of how that will all work, but this kind of thing is a type error which in many cases should be discoverable statically.