add-image-hash-to-detect-rename-and-optimize-reindexing

[prayas7102]

Here's a short note summarizing the changes made to address the issue of re-indexing renamed images:

Changes in db.py:

• Updated schema with hash column.
• Updated DB.VERSION to 3
• Updated ensure_version() with new migration logic and error handling
• Added get_image_by_hash() method
• Added update_image_filepath() method
• Added ensure_hashes() method to compute hashes for existing images
• Handled the case where the database schema is being upgraded from a version without the hash column in ensure_tables and ensure_version methods.
• Added error handling for UNIQUE constraint violations in update_image_filepath() method.
• Modified ensure_tables() to include 'hash' column and index. Also check if 'hash' column exists before creating the index

Changes in main.py:

• Added _compute_image_hash() method to RClip class
• Updated _index_files() method to compute and store hashes which implements logic to detect and handle renamed images.
• In _index_files() added error handling for IntegrityError when updating filepaths. Implemented a fallback to insert a new entry if updating fails.

These changes improve the efficiency of image indexing by avoiding re-indexing of renamed images and allow for faster queries using image hashes.