s3-parallel-put

Uses multiple processes (walker, putter, statter) to upload files concurrently, significantly reducing transfer times for many small files, as detailed in the architecture section.

Supports modes like 'update', 'add', and 'stupid' to optimize uploads based on file existence and changes, avoiding unnecessary HEAD requests or MD5 calculations for speed.

Allows resuming interrupted uploads from a log file with the --resume option, which is crucial for handling network failures in large batch operations.

Can guess Content-Type via 'magic' or filename, apply gzip compression to text files, and set custom headers, improving efficiency and S3 metadata management.