Flatten schema explorer tree in Flux builder, update searching

[alexpaxton]

• The tree should look like this:

  Bucket
    MEASUREMENTS
      Measurement Name (has Make Filter button)
        Field Name (Click to add filter)
    TAGS
      Tag Key
          Tag Value (Click to add filter)
    FIELDS
      Field name (Click to add filter)
  

• There should be a search box under the MEASUREMENTS and FIELDS headings, and under each tag key

• When searching under MEASUREMENTS or FIELDS, all matching children and the path in the tree that leads to them should be displayed

• All data for searching under MEASUREMENTS or FIELDS should exist prior to being able to search (synchronous search)

• When searching under a tag key, an asynchronous regex-based search with a limit should be used (debounced)

• When data is loading, a loading state should be displayed

• If data fails to load, an error state should be displayed

[chnn]

This story requires loading most of the data for the tree at once. That can be done with the following three queries:

RANGE_START = -1h

// Fetch all the measurements
from(bucket: "telegraf")
  |> range(start: RANGE_START)
  |> group(by: ["_measurement"])
  |> distinct(column: "_measurement")
  |> group(none: true)
  |> map(fn: (r) => ({_measurement: r._measurement}))
  |> yield(name: "measurements")
  
// Fetch all the tag keys
from(bucket: "telegraf")
  |> range(start: RANGE_START)
  |> group(none: true)
  |> keys(except:["_time", "_value", "_start", "_stop"])
  |> map(fn: (r) => ({tag: r._value}))
  |> yield(name: "tags")

// Fetch all the fields and their associated measurement
from(bucket: "telegraf")
  |> range(start: RANGE_START)
  |> group(by: ["_field", "_measurement"])
  |> distinct(column: "_field")
  |> group(none: true)
  |> map(fn: (r) => ({_measurement: r._measurement, _field: r._field}))
  |> yield(name: "fieldsByMeasurement")

Since Flux does not seem to support multiple yields yet, these should will have to be issued as separate queries for now.

Note that the flat list of fields (what is displayed in the FIELDS tree) can be extracted from the third query. However, the flat list of measurements has to be issued as a separate query (the first one), since a measurement does not necessarily have any fields associated with it.

These queries will need to be parsed at the bucket level into a data structure that looks something like the following:

{
  "measurements": {
    "measurement_1": [
      "field_1",
      "field_2",
    ],
    "measurement_2": [
      "field_3",
      "field_4",
    ],
  },
  "tags": [
    "tag_key_1",
    "tag_key_2",
    "tag_key_3",
  ],
  "fields": [
    "field_1",
    "field_2",
    "field_3",
    "field_4",
  ]
}