Skip to content

Latest commit

 

History

History
66 lines (51 loc) · 2.43 KB

README.md

File metadata and controls

66 lines (51 loc) · 2.43 KB
Raw

LayMix

Description:

LayMix - cli utility to generate Cartesian products from sample of image data. It should be useful for artists who need to produce few slightly-different versions of the same image (say, one character but with different hairstyles). Originally I've made it for a web visual novel, but then I've got permission to release the sources and further extend it.

Image Requirements:

  • All images must have the same (x, y) size
  • All images must be in png (for now).
  • Images that represent layers must have transparent background
  • All images (for now) must follow same naming convention, in order to automatically distinguish layers (see "Usage")

Dependencies:

  • Pillow==8.2.0

Installation:

  • pip install -r requirements.txt

Usage:

  • Open terminal of your choice in tool's directory
  • laymix_cli --prefixes first_layer second_layer third_layer (and so on) Where "{number}_layer" is either full name of layer or part of it. You can optionally pass path_to_dir_with_images before --prefixes - then tool will look for specific place where your files are located. Else it attempts to search for images in "./images" directory. Passing --savedir path_to_dir will set custom directory to save images into. Default is "./results"

Say, you have image with character's head drawn on it. And you want final images to have all custom hairstyles and facial expessions you've made. For this task, you will need (aside from base image) 2 image groups: one for hairs, other for facial expressions. You save these layers as separate images and name them like that: {background}_{name_of_layer}_{number}.png (without brakes) In case of our example, your images directory should look like that: lilu.png lilu_face_1.png lilu_face_2.png lilu_face_3.png lilu_hair_1.png lilu_hair_2.png and so on.

Then you run tool like ./laymix_cli --prefixes face hair If everything has been done correctly - you will get 6 different images in "results" directory. But what if original image has already included some hair or facial expression? Just add --include-background after command above - and you will keep them all (and amount of generated images will (in case of our example) increase to 11)

License:

GPLv3

This only affects this utility itself and not projects you use it in - you are free to produce pics under whatever license you want and use them in whatever projects you want, will they be FOSS or proprietary