Formerly known as tilegen.py, tiletools is a collection of tile-managing tools grouped together in a single repository.
Splits big image into smaller tiles, checks MD5 signature for repeated images, and deletes them if found. After the purge, a new smaller tileset is created according to specified dimensions. Requires PIL and Python 3.10 to work. Usage is best explained by example:
tilegen.py ./sample/maniac.png -- 16x16 result 256x192
will create a 256x192 tileset called result_256x192.png composed of the original sample/maniac.png image, cut in 16x16 tiles. Identical tiles are removed if found, so the actual number of tiles may be smaller than the total number of tiles in the original. A Tiled-compatible result_map.json that uses the tileset will also be created. You can use the result prefix as the directory where both files will be created.
Now you can specify multiple images as input and this script will generate unique tiles from all of them, but it's the last image used that will create a map.json tilemap. These multiple files are separated from the rest of the parameters by a --. If no -- is found, the previous behaviour of a single input file is considered for compatibility.
map.py [-h] [--version] [--room-width RW] [--room-height RH] [--max-ents MAX_ENTS] [--max-bytes MAX_BYTES] [-b] [-d DIR] [-c CONF] [--aplib] [-r] [-t] [-q] map_json id
ubox version of map.py is also included, but with the additional option -t or --transpose to encode the map in a transposed matrix layout in memory. This is necessary if your game needs to scroll the screen horizontally (you will probably need a new rendering function too). map.py creates a C or assembly version of the Tiled-compatible map.json file.
vwrap_up.py <input_images...> -- <output_image> <width> <step>
vwrap_down.py <input_images...> -- <output_image> <width> <step>
vwrap_{up|down}.py scripts create vertically wrapped around copies of the input_images and stack them side by side in the output file. Wrapping around is useful for pixel-level scrolling in games on the MSX1, like Konami's Pippols. width defines the output image width. If the tiles won't fit, vwrap_{up|down}.py will wrap to the next 8 lines. If step is 1, the resulting image contains 8 times the amount of input tiles. Step can be 1 (8 times), 2 (4 times) and 4 (2 times).
There are two types of tiles in a tileset:
- horizontal aligned tiles will wrap around themselves, not considering transition tiles;
- vertical aligned tiles will be considered as transitional tiles and
vwrap_{up|down}.pywill create transitional tiles between them;
You shouldn't mix vertical and horizontal tiles in the same file, because it creates unused wraping tiles.
For instance, considering a file called input.png composed of two horizontally aligned tiles:
calling:
vwrap_down.py input.png -- output.png 136 1
will result in the following output.png file:
Some tilegen.py features I may add some day:
- Remove ImageMagick dependency;
- Allow multiple input files;
- Create separated tilebanks when tilesets are bigger than 256 and use the screen region to decide which tilebank to use;
- Detection of near-identical tiles;
- Detection of color-swapped tiles;