Add test coverage for frame preprocessing in test_video_describer.py

[chigwell]

User Story
As a software developer,
I want to add test coverage verifying that video frames are properly preprocessed for model input in test_video_describer.py
so that we prevent silent failures caused by malformed model inputs.

Background
The current test mocks in test_video_describer.py (specifically test_get_video_descriptions and test_describe_frames) validate general method calls but don't verify critical preprocessing steps:

1. The VideoDescriber.describe_frames() converts BGR frames to RGB using cv2.cvtColor before model processing
2. The Blip model expects RGB-format PIL.Image inputs, but tests mock BlipProcessor without validating input transformations
3. Unvalidated frame conversion risks subtle model performance degradation if preprocessing changes accidentally

Acceptance Criteria

• Modify tests/test_video_describer.py to validate frame preprocessing:
  • Add assertions verifying mock BlipProcessor receives RGB-format PIL.Image inputs in test_describe_frames
  • Update @patch decorators to capture actual describe_frames arguments instead of generic mocks
• Add concrete test case:
  • Generate test frames with known BGR values using numpy
  • Verify describe_frames converts them to RGB before model processing
  • Confirm tests fail if cv2.cvtColor(frame, cv2.COLOR_BGR2RGB) is removed from production code
• Ensure test validation:
  • Check color channel order in processed images matches model expectations
  • Verify test coverage for both URL-sourced and local-file video processing paths
  • Maintain test isolation (no external API calls or file I/O)