Alan-Jowett / CoPilot-For-Consensus / 20684158781

78%

push

github

Add Cosmos DB support with 'cosmos' alias and managed identity authentication (minimal implementation) (#707)

* Initial plan

* Add Cosmos DB support with 'cosmos' alias and COSMOS_DB_ENDPOINT env var

- Updated create_document_store() to support 'cosmos' as an alias for 'azurecosmos'
- Added support for COSMOS_DB_ENDPOINT environment variable (Azure infra uses this)
- Updated ingestion config schema to include Cosmos DB parameters
- Updated ingestion service to use helper function for building document store kwargs
- Added unit tests for cosmos alias and COSMOS_DB_ENDPOINT environment variable

Co-authored-by: Alan-Jowett <20480683+Alan-Jowett@users.noreply.github.com>

* Add managed identity support for Cosmos DB

- Updated AzureCosmosDocumentStore to use DefaultAzureCredential when key is not provided
- Added azure-identity to dependencies in setup.py
- Added test for managed identity authentication
- All Azure Cosmos DB tests pass successfully

Co-authored-by: Alan-Jowett <20480683+Alan-Jowett@users.noreply.github.com>

* Fix linting issues (remove trailing whitespace)

Co-authored-by: Alan-Jowett <20480683+Alan-Jowett@users.noreply.github.com>

* Address PR review comments - fix env vars, logging, validation, and tests

- Fixed COSMOS_DB_KEY env var reading (was only checking COSMOS_KEY)
- Added validation for required Cosmos DB endpoint parameter
- Filtered sensitive credentials (key, password) from logs
- Updated tests to properly verify environment variable reading
- Added test for backward compatibility with legacy COSMOS_KEY env var
- Clarified docstring for managed identity usage

Co-authored-by: Alan-Jowett <20480683+Alan-Jowett@users.noreply.github.com>

* Simplify to minimal changes - remove ingestion service complexity

Reduced to only essential changes to avoid CodeQL warnings:
- Kept: cosmos alias in factory, COSMOS_DB_ENDPOINT/KEY support, managed identity
- Removed: ingestion service helper function, validation logic, log filtering... (continued)

10 of 15 new or added lines in 2 files covered. (66.67%)

6012 of 7876 relevant lines covered (76.33%)

0.85 hits per line