• Supported Programs
• Usage
• Requirements
• Versions
• Roadmap

wall-to-term converts images to color themes by finding the closest colors in the picture to the corresponding ideal colors. For instance, it looks for the closest color to red for determining the red color for the theme.

• Alacritty
• Konsole
• XFCE4 Terminal

• Doom Emacs. Largely untested. Preliminary tests show that some color combinations defined by the theme template result in reduced visibility due to the colors being too close. Will be fine-tuned with time and further testing.

I can add support for others. Just raise an issue and I'll see what I can do!

For general usage, just run ./wtt2.py --help. All the options and flags are described there.

Note that I tend to get the best results with the following settings (of course, this could and probably will be different for each image and your personal preference):

• Play quite a bit with the --threshold. This is especially important for images which do not have, for instance, a green color.
• Use --kmeans. There is nothing inherently better about it, it just gives slightly different results, which look better to me.
• Experiment with different palettes. This changes what colors wtt2 regards as "best-case" colors, and therefore can drastically affect output.
• Use the swatch theme type. It will generate a color swatch image. This is extremely useful for quickly testing other parameters, such as --threshold or --palette.

wall-to-term works by trying to find the colors in an image that are closest to a set of ideal colors. By default, it tries to find the closest color to absolute red (255,0,0), green, blue, etc. This set of colors is considered a palette. It is essentially a collection of "goal" colors for the program to try to reach.

With the --palette option, you can supply your own palette in the form of a text file. It must follow the following syntax:

• 8 colors and 8 lines, each line corresponding to a color.
• Color names may be used, such as black or yellow. Case does not matter.
• For exact colors, write a comma-separated triplet of 0-255 RGB values, such as 255,0,0.

Each line corresponds to a specific color, in this order:

black
red
green
yellow
blue
magenta
cyan
white

An example palette file:

black
255,0,0
0,255,0
yellow
BLUE
255,0,255
0,255,255
white

• OpenCV python3 -m pip install opencv-python
• Numpy python3 -m pip install numpy
• scikit-learn python3 -m pip install scikit-learn
• matplotlib python3 -m pip install matplotlib

The current working/featureful version is wtt2. It has more advanced, more accurate, and faster image analysis algorithms than wtt (and more features). wtt3 is a WIP for advanced scripting options, i.e. batch processing/generation. It is currently unstable to the point of being unusable. Same thing goes for wtt-numba. It is a slightly radioactive experiment which cannot be disposed of as it serves as an interesting example of performance testing. Note that, according to very basic tests, wtt2 with the --parallel flag tends to be as fast as wtt-numba, but without the extra library dependency.

• Add Gnome Terminal support.
• Add Windows Terminal support.