joram-berger / lydiff Goto Github PK

the lydiff module (i.e. the __init__.py and any other files in the folder) is intended to be importable in arbitrary contexts, e.g. from the CLI script or Frescobaldi (I can also imagine (and would like to create) a Python based web app to use that). Therefore it should not print() directly, at least not without further consideration.

Main communication should be handled by the calling program/script, so the functions should e.g. return values or raise exceptions.

For the intended debug output (show_output etc.) we have to find another way, e.g. with callback functions or the like.

The tool produces some yellow-ish output for non-changed areas for me. Looks like semi-green and semi-red are combined into semi-yellow instead of semi-gray. May be try switching off anti-aliasing, or using a different color space, or filter out yellow, or use "0, 255, 128" as green and "255, 0, 128" as red, etc...

Hi! The tool doesn't work with a score with multiple pages. It generates several png images for each page, then fails to combine them to a single image, then fails to locate that single image.

As convert-ly isn't intended to follow includes that step is somehow moot for moat real-world scores.

Seems like a complex issue. I think we should flatten the score to one huge file and keep references to the original locations for producing a reasonable text diff

Hi!

Is there a way to get a behavior of a classic diff program, i.e. skip non-changed portions of the score? For example, find a rectangular box that contains the difference and crop out everything else.