fix: Primary key declaration

[vrkirillov]

Summary

• Added support for using PRIMARY KEY in MergeTree family table definitions in place of ORDER BY
• Added support lists of columns in PRIMARY KEY declarations

Checklist

Delete items not relevant to your PR:

• Unit and integration tests covering the common scenarios were added
• A human-readable description of the changes was provided to include in CHANGELOG

---

Note

Medium Risk
Changes default DDL for MergeTree tables without order_by (PK-only vs ORDER BY (tuple())), which can affect physical layout and query performance for existing models.

Overview
Adds MergeTree primary_key model config (string or column list) so DDL can emit PRIMARY KEY (...) for table, distributed_table, and seed materializations, with docs and YAML/SQL examples in the changelog.

Refactors shared DDL helpers in table.sql: make_columns, renamed order_by_clause / partition_by_clause, and primary_key before order by / partition by. When only primary_key is set (no order_by), the adapter no longer forces default ORDER BY (tuple()), matching ClickHouse’s ability to rely on the primary key for sorting.

Integration tests assert system.tables primary_key and sorting_key for models with PK-only vs PK + explicit order_by.

^{Reviewed by Cursor Bugbot for commit c8c860d. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{Reviewed by Cursor Bugbot for commit 25ec4af. Configure here.}

[vrkirillov]

fix recommendations in Cursor