Added fields for MultiTermVectors (#42232)

[ojasgulati]

Fixes #42232

[elasticmachine]

Pinging @elastic/es-core-features

[gwbrown]

@elasticmachine test this please

It looks like this bug happened because the single termvectors API can take the list of fields as a URL parameter, as an alternative to the fields list in the body, and the URL parameter method is how the HLRC implements the single termvectors API, but the multi termvectors API can only take the list of fields in the body. This change, as is, will result in the HLRC putting the fields list in both the URL parameters and the request body for single-termvector requests - I'm not sure if that will cause a problem or not, we may see some failing tests.

We may end up having to change the HLRC implementation of the single termvectors API to not use the URL parameters since they will now show up in the body. If that's the case, we'll have to adjust this test case as well.

@mayya-sharipova I'm not very familiar with the term vectors API, is there a reason we decided to have the HLRC put the fields list as a URL parameter rather than in the body?

[mayya-sharipova]

@ojasgulati @gwbrown Sorry for a late reply. There is no any particular reason to have fields list as a URL parameter, and we can move them into the the body and fix the tests.

[mayya-sharipova]

@ojasgulati Thanks for submitting a PR. We may need to do several more things here that @gwbrown pointed out:

• We don't need URL parameter of params.withFields(tvrequest.getFields()); in here as we now will pass fields in the body
• We will need to fix the test pointed out by @gwbrown to remove all mentions of fields parameter
• Add a test to test that MultiTermVectors indeed work well fields. We can enhance CrudIT::testMultiTermvectors for this.

[ojasgulati]

@mayya-sharipova Thank you for your reply. I will make changes as you described in your comment.

[gwbrown]

@ojasgulati Are you still planning to work on this?

[gwbrown]

Given that this hasn't seen any activity for quite a while, I'm going to close this PR. Feel free to reopen this one or open a new one if you want to continue working on this.