This repository contains the full source (for what it's worth) of Kawa's Asspull IIIx toolkit. Originally they were all made in Visual C# but every one of them has now been rewritten in Python.

This takes an AP3 ROM file, sets the checksum field in the header, and pads it out to the nearest power of two.

usage: assfix.py [-h] [-v] inFile [outFile]

Pad an A3X .ap3 ROM file and fix its checksum.

positional arguments:
  inFile         source .ap3 file
  outFile        target .ap3 file

optional arguments:
  -h, --help     show this help message and exit
  -v, --verbose  use verbose output

The A3X BIOS has its own image format. This little tool can convert them back to PNG (or GIF, or BMP, if you want to for whatever reason), or merely show them by not specifying an output file.

usage: ass2img.py [-h] [-v] inFile [outFile]

Display an A3X .api file or convert it to a regular image.

positional arguments:
  inFile         .api file to display or save
  outFile        if specified, file to save as

optional arguments:
  -h, --help     show this help message and exit
  -v, --verbose  use verbose output

This takes an image of a top-to-bottom gradient and converts it to an array of xBGR-1555 color values that you can use as HDMA gradient background data. Fair warning, that does cause banding. It takes up to three parameters, making assumptions if anything but the first is missing: the input image file, the output .s or .c file, and what to name the data. Given only a foo.png, it'll assume foo.s with identifier foo.

usage: hdma2ass.py [-h] inFile [outFile] [identifier]

Convert a color strip to HMDA color data.

positional arguments:
  inFile      source picture
  outFile     target .s or .c file
  identifier  identifier name

optional arguments:
  -h, --help  show this help message and exit

This takes a PNG image (or GIF, or BMP, if want to for whatever reason) of either four or eight bits per pixel and tries to convert it to the A3X BIOS' custom format. Basically It takes up to three parameters: the input image file, the output .api file, and whether or not to skip compression, the --raw/-r parameter. Again, given only a foo.png, it'll assume foo.api unless told otherwise.

Normally, either 16 or 256 colors are written. If the source image has significantly fewer than 256 but more than 16 colors, you may use the --truepal/-t parameter to store only the true amount of colors and significantly reduce the file size. Be aware that the A3X BIOS can't tell and will still try to load all 16 or 256 colors, so garbage data will appear in the system palette to fill up the difference.

If a foo-hN.png exists, that file will be converted to an HDMA gradient, like with hdma2ass, and embedded in the resulting file. N can be any number from 0 to 7, and the gradient will apply to that color index. Use the --nograds/-g parameter to skip this part.

usage: img2ass.py [-h] [-r] [-v] inFile [outFile]

Convert an image to the A3X .api format.

positional arguments:
  inFile         source image file
  outFile        target .api file

optional arguments:
  -h, --help       show this help message and exit
  -r, --raw        skip compression
  -v, --verbose    use verbose output
  -g, --nograds    skip HDMA gradients
  -c, --clipgrads  clip HDMA gradients
  -t, --truepal    store minimal palette data

This takes a JSON file describing a locale and keyboard layout, and converts it to a .loc file that you can use on the A3X by storing it on a disk and listing it in a start.cfg file on that disk, with a line like locale=foo.loc, or by selecting it in the Navigator.

usage: mkloc.py [-h] inFile [outFile]

Convert an A3X locale from JSON to binary or vice versa.

positional arguments:
  inFile      source file
  outFile     target .loc or .json file

optional arguments:
  -h, --help  show this help message and exit

This takes a Tiled map exported to uncompressed JSON and converts it to A3X tilemap format, layer by layer. Only tile layers are supported. Any hidden layers are skipped, and a custom tileOffset property can be used to adjust what tile indices are used in the result. Also, to set palette information you can bind a custom palette property to entries in the tileset. Tiles can be flipped. The resulting data will use the layer names as identifiers, so you may want to make sure you name them right. This takes two parameters: the input .json file and optionally the output .s or .c file, again assuming if only one is given.

usage: tiled2ass.py [-h] inFile [outFile]

Convert a Tiled JSON map to A3X map data.

positional arguments:
  inFile      source file
  outFile     target .s or .c file

optional arguments:
  -h, --help  show this help message and exit