Goals of Kar Program:

• Install, refresh, and remove documents with minimal info (only URL).
• Query documents with full-text search and filters.

Design Proposal:

• API inspired by Pacman with top-level operations: -S (sync), -Q (query), -R (remove), -V (version), -h (help).
• Data stored in statedir (default: ${XDG_STATE_HOME}/kar/).

Sync Operation:

• Install new documents by downloading content and updating database and index.
• Refresh existing documents to update content and search index.
• Validates URLs according to [URL Standard] and applies normalizations.
• Follows redirects based on specific conditions.
• Identifies duplicates based on normalized URLs and uses ext_equiv_url for extended equivalency.
• Can rebuild metadata database from downloaded content.

Query Operation:

• Centers around full-text search of document content with additional filters and projections.
• Supports filters like --url, --domain, and --limit.
• Projections include --only-url.
• Query results are printed as static text.

Appendix - Future Work:

• Extend support for more mime types like text/plain and application/pdf.
• Better handle rate limiting per domain.
• Improve search architecture for better query performance.
• Extend support for document display options like adding filters and projections.
• Explore a browser extension to install documents with the active page's URL.