feat: complete ORM query builder and model-driven workflows

[KKould]

What problem does this PR solve?

KiteSQL already had the first piece of ORM support, but it still could not cover most of KiteSQL's query and maintenance capabilities through a consistent high-level API.

This PR completes that layer so model-driven workflows can stay inside the ORM for projections, joins, subqueries, set queries, mutations, introspection, and related documentation/tests, instead of falling back to raw SQL or exposed AST details.

Issue link:

What is changed and how it works?

ORM query builder

• Rebuild ORM DQL/DML/DDL construction on top of sqlparser AST while keeping AST details hidden from end users
• Expand expression support with function calls, arithmetic, aggregates, CASE, aliased values, DISTINCT, and richer subquery predicates
• Add group_by, having, tuple/value projections, and DTO projections via #[derive(Projection)]
• Add source aliasing, typed joins, USING, cross joins, and join-aware struct projections
• Unify composable query sources under QueryOperand and support UNION / EXCEPT with result-level ordering, limits, count, exists, get, and explain

ORM mutations and maintenance

• Unify query-driven mutations under from::<M>()...update()/delete()
• Add INSERT ... SELECT support and ORM-side insert/overwrite builders at the end of query chains
• Add convenience helpers such as insert_many, truncate, create_view, create_or_replace_view, drop_view, show_tables, show_views, describe, and explain
• Keep the ORM surface focused on high-level model/query workflows rather than explicit AST export APIs

Fixes, docs, and tests

• Fix correlated IN subqueries, set-query binding/top-level ordering, and EXCEPT DISTINCT / EXCEPT ALL semantics
• Expand ORM macro tests and add SLT coverage for subqueries, unions/except, and insert-from-query cases
• Rework ORM rustdoc and README so the final API surface is documented around the current builder design

Code changes

• Has Rust code change
• Has CI related scripts change

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Note for reviewer

This PR is large because it carries the whole ORM completion pass on top of the initial model-driven ORM groundwork. The main thread is: finish the ORM around the existing KiteSQL capability surface, keep AST details internal, and add enough tests/docs to make the new builder flow stable and understandable.