Skip to content

feat(databricks): unify configured client composition and UC data-product import - #242

Merged
ElliotSun merged 43 commits into
mainfrom
feat/databricks-configmanager-importer
Sep 24, 2026
Merged

ElliotSun merged 43 commits into
mainfrom
feat/databricks-configmanager-importer

Conversation

@ElliotSun

Copy link
Copy Markdown
Collaborator

Summary

This keeps the real-workspace findings from the local Databricks smoke while removing environment-specific test conveniences.

  • add typed databricks project/global configuration;
  • resolve SemaPact connection hints explicitly at Databricks composition boundaries;
  • keep the low-level WorkspaceClient factory configuration-neutral;
  • preserve Databricks SDK unified authentication when SemaPact hints are absent;
  • make runtime registry, readiness, Unity import, and live smoke use the same configured-client composition;
  • support catalog.schema Unity import as one ODCS data product containing all discovered assets;
  • use the existing authenticated SDK TableInfo metadata for relationship extraction instead of a second raw HTTP/token path;
  • aggregate Unity relationship metadata across the entire imported data product;
  • declare the directly used Databricks SDK dependency at the provider surface validated by the live run.

Safety / scope

Local-only defaults and the environment-specific live regression script were deliberately removed. The canonical live smoke retains its protected semapact_smoke_* namespace guard and explicit inputs.

The #239/#240/#241 release/deployment workflow is unchanged; this PR changes Databricks configuration/import composition only.

Validation context

The originating branch was exercised successfully against a real Databricks workspace before this cleanup. CI should now validate the cleaned implementation across the supported Python matrix and Databricks extra.

…treamline Unity importer

- Resolve Databricks connection hints (workspace_url/host, token, profile) through ConfigManager fallback hierarchy in create_databricks_workspace_client
- Remove _databricks_env and process-wide environment variable mutations in Unity Catalog importer
- Delegate table discovery and metadata retrieval directly to authenticated WorkspaceClient and pure schema mapping
- Add end-to-end regression script for live Databricks schema Data Products
- Update unit and live smoke tests with proper configuration isolation
@ElliotSun
ElliotSun merged commit cfa8b67 into main Sep 24, 2026
11 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant