Add Lakeview dashboard CRUD + Publish tools to UC server

[priyal-c]

Summary

• Adds 8 Lakeview (AI/BI Dashboard) tools to the Unity Catalog MCP server: list, get, create, update, delete, publish, unpublish, get_published
• Uses WorkspaceClient().lakeview SDK methods, follows existing BaseTool / factory pattern (same as genie.py, vector_search.py)
• Gated by enable_lakeview CLI setting (default: True), opt-out with --enable-lakeview false
• Includes 18 unit tests covering tool listing, execution, and integration via mocked WorkspaceClient

Motivation

Customer (Mural) asked whether the managed Databricks SQL MCP server supports dashboard creation. It doesn't — only execute_sql, execute_sql_read_only, and poll_sql_result are exposed, and the managed server is closed-source. This gives the community a way to manage AI/BI dashboards via MCP using the open-source UC server.

Tools Added

Tool	Description
lakeview_list_dashboards	List dashboards with pagination
lakeview_get_dashboard	Get metadata + serialized definition
lakeview_create_dashboard	Create a new dashboard
lakeview_update_dashboard	Update display name, warehouse, or definition
lakeview_delete_dashboard	Move to trash (recoverable)
lakeview_publish_dashboard	Publish a dashboard version
lakeview_unpublish_dashboard	Remove published version
lakeview_get_published_dashboard	Get published version details

Test plan

• make lint passes (black --check + ruff check)
• uv run pytest tests/ — 42 tests pass (18 new + 24 existing)
• Integration test against a live workspace (manual)
• Verify tools/list MCP call includes lakeview tools via stdio

This pull request and its description were written by Isaac.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}