Support query filters matching list values in generic and mapped tables

[pavelhoral]

PR adds support for query filters matching list values as defined in the QueryFilter contract (#167) and adds support for JSON based queries on mapped tables in PostgreSQL database.

• Generic tables index each array item under two distinct property table rows. One for the exact JSON pointer (e.g. /foo/0) and the other for parent property pointer (e.g. /foo). This forced switching generic table queries from using JOINs to using subqueries with EXISTS (...) clause.
• I did test for performance impact on the JOIN->EXISTS switch using PostgreSQL and both approaches seem to yield comparable results. I hope other DB engines will behave in a similar way.
• Overall architecture of query rendering for mapped tables was significantly tuned (and kind of simplified). Now each column gets its own query filter visitor and there are separate methods and classes for working with JSON based columns and for other simple value columns.
• Right now, the only database that got support for working with JSON columns in mapped tables is PostgreSQL. Nothing prevents other databases from getting the support as well. The basic JsonFieldFilterVisitor is using standard SQL/JSON functions which is AFAIK being supported by Oracle and DB2 (from the list of databases supported by Wren:IDM).

Data migration for preexisting installations:

• Searching JSON arrays is possible only after objects stored in generic table format are updated (forcing cleanup and generation of values in the property table). There is currently no automatic migration. I am not sure it is worth the effort... we just might provide guidance for manual migration in release notes (e.g. what SQL queries can be used to unwrap existing arrays data).

This PR replaces #268.

[sonarqubecloud]

Quality Gate failed

Failed conditions
64.6% Coverage on New Code (required ≥ 80%)

See analysis details on SonarQube Cloud