Rewrite SQL node early in the planning phase to support parameterization of MSQ DML queries

[abhishekrb19]

It's useful to supply dynamic parameters to automate periodic jobs. However, previously only MSQ SELECT queries could be parameterized. For example, consider the following MSQ REPLACE query with valid dynamic parameters:

REPLACE INTO foo
OVERWRITE WHERE __time >= ? AND __time < ?
SELECT TIME_PARSE(ts) AS __time, c1
FROM (VALUES('2023-01-01', 'day1_1'), ('2023-01-01', 'day2')) AS t(ts, c1)
WHERE c1 = 'day2'
PARTITIONED BY DAY

[
  {
    "type": "TIMESTAMP",
    "value": "2023-01-01"
  },
  {
    "type": "TIMESTAMP",
    "value": "2023-01-05"
  },
]

The expectation is that the parameters will bind to the ? placeholders in the query in place of the literal, but the query fails with the following exception:

Invalid OVERWRITE WHERE clause [`__time` >= ?]: Cannot get a timestamp from sql expression [?]

This is because there is a bunch of validation in the IngestHandler code that happens during the planning stage, which runs on the node where the rewrite hasn’t happened yet. This change moves the call to rewriteParameters() earlier in the planning stage so any validations will run on the rewritten root node.

This PR has:

• been self-reviewed.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
• been tested in a test Druid cluster.

[clintropolis]

👍