Add raw_data query support to backend

[fridgepoet]

What this PR does / why we need it:
This recreates the raw_data query processing that exists in the frontend in the backend. It redirects any queries with only raw_data to the backend.
This is based on equivalent work in Elasticsearch (grafana/grafana#63208 and grafana/grafana#59741) and on the existing query logic in OpenSearch's frontend.

1. For each hit in the hits response, we create an object:
   • with entries for each object in _source -- these are flattened to a maximum depth of 10
   • with entries _id, _type, _index
   • with an entry for timestamp from fields if possible and falling back to the timestamp _source if possible
2. Each entry becomes a data frame field with the values associated with each hit
   • To detect the data frame type, we look at first non-null value and set type based on it.
3. The fields are then sorted alphabetically, except for the time field which is specially placed at the beginning.

For testing, one way is to use the sample data and create a query > Metric > Raw Data.
Compare a version without these changes, e.g. v2.6.2.
In our squad, check the data source set up here https://clouddatasources.grafana.net/goto/BTjFTmuVR?orgId=1 under the Lucene Raw Data section.
This should be running the frontend logic.

Build locally this PR and make the same query.
The same data results should be observed. There is a difference because the data frame types are now interpreted (previously values were all strings), so some default format options round the numbers.

Here are some examples of a visual comparison:
Black is the frontend flow

White is the locally-built backend flow

Which issue(s) this PR fixes:

Fixes #196

Special notes for your reviewer:
When possible, tests were pulled from the Elasticsearch PR and adapted for our code.
We also included this change grafana/grafana#67767 as we ran into the issue. We'll likely include it for other types of queries as well.

If there are too many changes, let me know if I should try to break it up. I'm also happy to go over it together.

[iwysiu]

A couple questions/nits, but nice work!

[idastambuk]

We'll likely include it for other types of queries as well.

We should probably add this in the description of either the epic ticket or the tasks in it

[fridgepoet]

Thanks @iwysiu and @idastambuk for having a look! I think I've addressed the feedback so far. Let me know if you have any other suggestions.

[iwysiu]

Nice work! One last little comment nit that was just me being unclear.

Was there a reason you needed to force push? If you can, once people have started reviewing a pr you should avoid force pushing, it messes up the links for comments. (If you need to fix merge conflicts you can do a rebase force push after its been approved)

[fridgepoet]

Thanks for having a look @iwysiu and good point, sorry about that! I will avoid that in the future.