feat: auto-apply Snowflake ownership object tags on publish

[Vinay Shende (vinay79n)]

Implements the table-ownership RFC (RFC: Snowflake table ownership via object tags) by tagging tables through the existing publish chokepoint instead of hundreds of hand-written ALTER statements.

• Add _snowflake/object_tags.py:
  • build_table_tags — derives TEAM/DOMAIN/PROJECT from Metaflow context, defaults STATUS=active, validates STATUS/SLA against allowed-value lists. OWNER is resolved by priority: explicit tags={"owner": ...} → owning-team alias ds-<domain>-team when the domain is known → unknown (we don't use current.username, which on argo/deployed runs is a service identity, not a person).
  • build_set_tag_sql / apply_table_tags — build and run the ALTER ... SET TAG, with identifier validation to block SQL injection and warn-don't-fail behavior (a missing tag def or missing APPLY never breaks a successful publish).
• Wire automatic tagging into publish() and publish_pandas() via a new optional tags override. Tags are applied only to production tables (DATA_SCIENCE); non-prod (DATA_SCIENCE_STAGE) runs are left untouched.
• Add metaflow/registry.py: create_ownership_registry_view() + TABLE_OWNERSHIP_REGISTRY view (plain view, adoption-based, no refresh needed; ~2h ACCOUNT_USAGE lag — use INFORMATION_SCHEMA.TAG_REFERENCES / SYSTEM$GET_TAG for real-time checks).
• Export create_ownership_registry_view from the metaflow package.
• Tests: unit tests with 100% coverage of object_tags.py (derivation, owner priority, validation, warn-don't-fail).
• Docs: publish / publish_pandas ownership-tags sections + registry view page.
• Bump version 0.4.2 → 0.5.0.

Admin prerequisites (per the RFC §3): in PATTERN_DB.DATA_SCIENCE, create the 7 TABLE_* tag definitions (with ALLOWED_VALUES on TABLE_STATUS / TABLE_SLA) and grant the publishing role APPLY on each. Until then, tagging is skipped with a warning (publish still succeeds).

ClickUp card - DS-2491

[Copilot]

Pull request overview

Adds first-class Snowflake table-ownership tagging to the Metaflow publish “chokepoints” and introduces an admin helper view that surfaces those tags via a central registry view.

Changes:

• Introduces _snowflake/object_tags.py to derive/validate ownership tags, generate ALTER TABLE … SET TAG SQL, and apply tags (warn-only on tagging failures).
• Wires optional tags overrides into publish() and publish_pandas() and applies tags for production tables.
• Adds create_ownership_registry_view() + docs to create/query the TABLE_OWNERSHIP_REGISTRY view; bumps version to 0.5.0.

Reviewed changes

Copilot reviewed 11 out of 12 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
uv.lock	Bumps locked project version to 0.5.0.
pyproject.toml	Bumps package version to 0.5.0 and updates authors.
src/ds_platform_utils/_snowflake/object_tags.py	New tag derivation/SQL builder/apply helpers for Snowflake object tags.
src/ds_platform_utils/metaflow/write_audit_publish.py	Adds tags parameter and applies tags post-SWAP in production.
src/ds_platform_utils/metaflow/pandas.py	Adds tags parameter and applies tags in production for both direct and S3-stage paths.
src/ds_platform_utils/metaflow/registry.py	New helper to create the ownership registry view in Snowflake.
src/ds_platform_utils/metaflow/init.py	Exports create_ownership_registry_view.
tests/unit_tests/snowflake/test__object_tags.py	New unit tests for tag derivation, SQL construction, and apply behavior.
README.md	Links to the new create_ownership_registry_view docs page.
docs/metaflow/publish.md	Documents the new tags param and ownership-tag behavior.
docs/metaflow/publish_pandas.md	Documents the new tags param and ownership-tag behavior.
docs/metaflow/create_ownership_registry_view.md	New docs for creating/querying the registry view.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.