[ML] AIOps Log Rate Analysis: Support for text fields.

[walterra]

With the recent work done for the data drift view, we now have utilities in place to calculate p-values for any kind of data to be compared. We can make use of that to add support for identifying significant log categories. Within log rate analysis, we could categorize logs for the baseline and deviation time range and then use chi2test to identify significant changes. The results could then be part of the analysis results table. The trickier part will be to make this part of the grouping as the log categories cannot be used within the frequent_item_set aggregation. So this would have to be something on top, like augmenting the keyword based groups with matching log categories.

Tasks

Beta Give feedback

1. Move chi2test to package [ML] Move chi2test to package #167237
2. Move log pattern analysis queries to common [ML] AIOps: Move log pattern analysis fetch code to common to be available via server #167465
3. Support text fields in log rate analysis [ML] AIOps: Supports text fields in log rate analysis #165124
4. Support text fields for log rate analysis grouping. 3af27fd71e9662d9e066fa83edbdb3323b8d7d94
5. Disable link to log pattern analysis for table rows with a log pattern already as the value. 16dd091
6. Fix document count chart for log patterns. 8dc72584eca
7. Fix loading percentage for combined log pattern and keyword histograms. 103384bcabd5b7b33842baf0d62f72b8e83ffaf1
8. Fix getting counts for baseline and deviation for categories. d7ecabb547f6660b16c33837af0fecd61210f3bb
9. Fix histograms for groups with text fields.

[elasticmachine]

Pinging @elastic/ml-ui (:ml)

[walterra]

Moved follow up tasks to the meta issue here #168459. Closing as this covers the work done for 8.11.