Add transformers for videos , bounding boxes

[vishakha041]

On annotations: annotation_source=coco (even better, this should be whatever model was used to extract the anntoation, e.g. SAM), annotation_mode=auto (could be manual or none)

[ad-claw000]

The issue requests new transformers specifically for handling videos and bounding boxes (annotations), with bounding boxes receiving properties like annotation_source and annotation_mode.

Here are a few possible approaches for the implementation:

Approach 1: Create dedicated VideoProperties and BoundingBoxProperties transformers

We can add new files aperturedb/transformers/video_properties.py and aperturedb/transformers/bounding_box_properties.py, which inherit from Transformer. Transformer.__init__ would be updated to track _add_video_index and _add_bounding_box_index.

• Pros: Follows the established modular pattern of ImageProperties; allows embedding specific logic (e.g., extracting video duration or box areas later).
• Cons: Requires users to chain multiple specific transformers together in their data pipelines.

Approach 2: Expand CommonProperties to handle videos and bounding boxes

We can update aperturedb/transformers/common_properties.py to iterate over commands like AddVideo and AddBoundingBox, allowing users to inject properties like annotation_source and annotation_mode dynamically.

• Pros: Simplifies the pipeline by reducing the number of transformer classes needed; makes CommonProperties truly universal for all data types.
• Cons: Users must manually specify properties like annotation_source="coco" during initialization rather than relying on dedicated defaults; might clutter CommonProperties if type-specific logic is needed later.

Recommended Path Forward

Approach 1 is recommended. Creating dedicated VideoProperties and BoundingBoxProperties aligns perfectly with the existing architecture (like ImageProperties in aperturedb/transformers/image_properties.py).

To implement this:

1. Update aperturedb/transformers/transformer.py (around line 60) to track _add_video_index and _add_bounding_box_index during command parsing.
2. Create aperturedb/transformers/bounding_box_properties.py which sets annotation_source and annotation_mode on AddBoundingBox properties.
3. Create aperturedb/transformers/video_properties.py for video-specific metadata.