Add stream fields to each dataset

[ruflin]

During the generation of the datasets, a stream.yml file is created with the stream fields inside. These are required for each dataset.

[ruflin]

This PR requires #295 to be merged first.

@neptunian @skh This adds the stream fields to each dataset. With this, EPM does not have to do anything special. I checked and it looks like template generation and index pattern generation works with the fields. On the index pattern side: Are constant_keyword fields just treated as keyword or does KB have support for it already?

[neptunian]

@ruflin for index patterns there is no support for the constant_keyword field. if the type is unknown, it should not be set, so the field would have no type set. For mappings, if it doesn't recognize the type, it just passes it through. I created an issue to support it here elastic/kibana#61112 . keyword is converted to a string for index patterns, I imagine we will do the same for constant_keyword?

[ruflin]

@neptunian Is it ok to merge this PR as I think it does not break anything at the moment?

Moving the other conversation to elastic/kibana#61112

[neptunian]

yes, it won't break anything.