Implement Text-Based Multi-Person Face Search

[rohan-pandeyy]

Description

Implement natural language multi-person face search through the existing search bar.

This feature will allow users to search for media containing multiple people by entering queries such as:

• John and Jane
• Mom, Dad, Son
• Alice, Bob
• John, Jane and Sarah

For reliable results, users must have labeled their face clusters in the People section.

The search system should interpret the entered names, match them against labeled face clusters, and return media containing all specified people.

This approach expands upon the newly introduced UI based multi-person search #1302.

Proposed Solution

• Query parsing for multi-person searches
• Name-to-face-cluster resolution
• Backend search implementation
• Frontend integration
• Search result handling and validation
• Testing and documentation

Acceptance Criteria

• Users can search for multiple people using the existing search bar.
• Queries containing multiple names are correctly parsed.
• Names are matched against labeled face clusters.
• Results only include media containing all requested people.
• Invalid or unknown names are handled gracefully.
• Existing search functionality remains unaffected.