Adding separate methods to extract the timestamp at which the keyframes occur.

[AkashWudali12]

Description: This PR introduces separate methods and helper functions to return both the extracted keyframes and their associated temporal data. Specifically:

These methods maintain the original functionality while also returning:

• The time (in seconds) from the start of the video when each keyframe occurs.
• The interval (in seconds) during which each keyframe appears.

This enhancement enables future features, such as matching video transcripts or audio to keyframes, thereby improving the contextual understanding of the keyframes by adding more detailed metadata.

Bug Fix: Additionally, this PR addresses a bug found in the image_selector.py file, where the variance_laplacians list was occasionally empty.

• A temporary solution has been implemented to handle cases where this list is empty.
• However, this likely points to an issue further upstream that may require additional investigation and resolution.

Why This is Useful

• The added temporal data for keyframes allows for better integration of visual information with audio or transcripts, paving the way for richer contextual analysis of videos.
• The bug fix ensures the stability of the current functionality while highlighting an area for potential improvement.