[ML] Add embedded map to geo_point fields for Data Visualizer

[qn895]

Summary

This PR is part of #86387. It adds embedded maps to the geo_point fields for the data visualizer (both the file based and index based). Changes include:

• A new <MlEmbeddComponent/> that can be consumed in other places inside Ml
• Maps plugin as a required plugin
• Updated functional tests with a new file for checking geo_point fields
• Update the kibanaField formatting for the preview distribution for Number rows

Future Todos

• Support for geo_shape fields

Checklist

Delete any items that are not applicable to this PR.

• Any text added follows EUI's writing guidelines, uses sentence case text and includes i18n support
• Unit or functional tests were updated or added to match the most common scenarios

[elasticmachine]

Pinging @elastic/ml-ui (:ml)

[pheyos]

Really cool to have the maps in the data visualizer row expansions, great addition!

Functional tests in the file data viz LGTM, just one small suggestion.

I've noticed that the expanded row content is different in file data viz compared to the index data viz after import:

File data viz before import

This shows document stats, top values and the map.

Index data viz after import

This shows document stats, examples and the map.
Is this difference intended?

And related to that: The service method assertGeoPointFieldContents will currently only work for the file data viz (because of that difference). And it would be good to also have functional tests for maps in the index data viz, so we should either make both row contents the same or have separate service methods to validate.

[peteharverson]

I don't think this is related to your PR, but for this filebeat-apache-2019.01.30 data set, no examples are displayed for the traefik.access.geoip.location field.

I remember seeing something like this in the card based version, which was related to the query we were using to get the examples - https://github.com/elastic/kibana/blob/7.9/x-pack/plugins/ml/public/application/datavisualizer/index_based/components/field_data_card/content_types/geo_point_content.tsx

If the query to get the examples returns no docs, we should probably show some sort of message in that cell saying something like No examples were obtained for this field

[qn895]

I added this message here f537b8c

[peteharverson]

Showing the No examples were obtained for this field is good here, but I think we can actually improve the search used to obtain the examples in ml/server/models/data_visualizer/data_visualizer.ts getFieldExamples too. I think we can edit that search to use the fields parameter in place of the current _source parameter. I gave this a quick try and examples now show up for the traefik.access.geoip.location field above:

[qn895]

Updated here d44517f

The example lists are now using fields API, which will look different since the API auto detects and auto formats.

Previous format:

[peteharverson]

@pheyos regarding #88880 (review), I think we have to show examples for the index based view as geo fields don't support aggregations. Showing the examples list for the file based view would work for me, if we want to use the same component for both. I guess the radius of the circles on the map gives an indication of the relative frequencies (which is what the top values list provides too).

[pheyos]

I agree @peteharverson, showing the examples list in both views sounds good.

[pheyos]

Good to see that file data viz and index data viz are showing the same components for the geo row expansion now 🎉
One suggestion and one comment about the failing tests:

[pheyos]

Functional tests LGTM 🎉
Just one question around variable naming.

[darnautov]

did you consider this instead of any?

Suggested change

	const doc: any = get(hits[i].fields, field);
	const doc: object[] | undefined = get(hits[i].fields, field);

[peteharverson]

Tested latest edits. LGTM. A great addition!

[darnautov]

LGTM

[kibanamachine]

💚 Build Succeeded

• continuous-integration/kibana-ci/pull-request
• Commit: dca4051

Metrics [docs]

Module Count

Fewer modules leads to a faster build time

id	before	after	diff
ml	1723	1728	+5

Async chunks

Total size of all lazy-loaded chunks that will be downloaded as the user navigates the app

id	before	after	diff
maps	2.6MB	2.6MB	+115.0B
ml	6.5MB	6.5MB	+9.4KB
total			+9.5KB

Page load bundle

Size of the bundles that are downloaded on every page load. Target size is below 100kb

id	before	after	diff
ml	56.6KB	57.2KB	+537.0B

History

• 💚 Build #101219 succeeded 647a5d2
• 💚 Build #100764 succeeded 7e81cd3
• 💔 Build #100458 failed 62acfc4
• 💔 Build #100215 failed b925e59
• 💔 Build #100110 failed 11eaf09

To update your PR or re-run it, just comment with:
@elasticmachine merge upstream