Split QueryExecutor::query into separate methods for SQL and InfluxQL

[hiltontj]

Problem statement

As part of #25751, when executing InfluxQL queries for the /query API, we need the ability to capture information about the incoming InfluxQL query and use that to augment how the HTTP API response is structured.

The only case for this that I am currently aware of is when using a GROUP BY clause.

Proposed solution

Split the QueryExecutor::query method into two methods: QueryExecutor::query_sql and QueryExecutor::query_influxql.

The query_sql method will keep the current API, less the query_kind parameter, and can function as is.

The query_influxql method will change in the following way:

• accept a parsed Statement (it may also accept the raw query string if that is needed for query logging or otherwise)
• use the InfluxQLQueryPlanner::statement_to_plan method to create the logical plan

Otherwise, its API will still stay the same such that it functions as is with the /api/v3/query_influxql API

This will allow us to parse the InfluxQL query in the API handler, which we are already doing (see here), capture some state about the query, e.g., the GROUP BY clause, and use that to handle the record batch stream on its way out of the query execution. It also keeps us from parsing the query string twice (see here).

Alternatives considered

N/A

Additional context

See related issue: #25751