Handwriting generation pipeline
The Handwriting generation pipeline extends Tom Haines's My Text in Your Handwriting project with the surrounding tools needed to move from scanned samples to a printable result.
The repository groups training-sheet preparation, scan cleanup, sample splitting, text formatting, limited Unicode-to-ASCII mapping, synthesis, and printable-page layout. Its modern-hst branch also ports the active HST and LET paths to Python 3 on Ubuntu 24.04.
The optimized compositor reduced a validated page benchmark from 1.695 seconds to 0.190 seconds and peak memory from 1625 MiB to 362 MiB while preserving byte-for-byte output.