Integration with Elasticsearch Transform

First, it would be helpful to stream aggregated data down via composite aggs. This would allow users to aggregate, filter, and then push their data into a pandas dataframe.

Then that data could be pre-aggregated, put in a dataframe, train a model, and then that model pushed to Elasticsearch.

The issue is a way to "deploy" the feature aggregation.

Eland could solve this by deploying a [transform](https://www.elastic.co/guide/en/elasticsearch/reference/current/transform-apis.html). This allows for data to be continually transformed via aggregations. Then that data can be pushed through a Pipeline which references the trained model.

No repository files, tests, or entry points are named. Start by reviewing Eland's existing Elasticsearch and DataFrame integrations, then read the linked Elasticsearch Transform API and composite aggregation documentation; done means agreeing on a scoped way to deploy feature aggregation through a transform and model pipeline.

Written by the indexing model from the issue text.