2.8 KiB
Fonts
This code sample demonstrates a way to get from a system font or other font file to Python code which may be used normally or stored as frozen bytecode. The latter enables major savings in scarce RAM.
Alas this process uses a closed source Windows program. It is available here but it is free (as in beer) and it will run under Wine.
To convert a system font to Python follow these steps.
Start the Font Creator. Select File - New Font - Import an Existing System Font and select a font.
Accept the defaults. Assuming you have no desire to modify it click on the button "Export for GLCD".
Select the microC tab and press Save, following the usual file creation routine.
On a PC with Python 3 installed issue (to convert a file Ubuntu_Medium17x19.c to Python ubuntu.py)
python3 CfontToPython.py -i Ubuntu_Medium17x19.c -o ubuntu.py
To test issue
import ubuntu
ubuntu.font.test('ghi')
The machine generated Python file can be run under cPython, copied to the Pyboard and run, or copied to the scripts directory and used to build firmware with the file as a frozen module.
This assumes Linux but CfontToBinary.py is plain Python3 and should run on other platforms.
The font matrix
In the C file each character is stored as a fixed size array of bytes, the first byte being the character width. When rendering a font to a device, fonts designed as variable pitch should use this byte as the width. Monospaced fonts should be rendered using the font's width (see pyfont.py).
Organisation
This describes the layout of the data in the C file and the Python bytes object, for a font with characters of X by Y pixels. Individual pixels are denoted by x and y. Displays usually use a cordinate system where (0, 0) represents the top left hand corner, with points on the screen represented by positive values of (x, y). These notes assume this organisation.
For a font of X by Y pixels, each column uses (Y % 8) +1 bytes. Thus a 5 x 10 font will use 2 bytes per column by 5 columns = 10 bytes. An additional byte at the start signifies the width of the individual character giving 11 bytes per character.
The first byte after the width holds the vertical data for pixels (0, 0) to (0, 7), with (0, 0) in the LSB. For a 5 x 10 font the second byte will hold vertical data for pixels (0, 8) and (0, 9) in its LSB's, with the next byte holding (1, 0) to (1, 7) and so on, through to (4, 8) and (4, 9).
Thus, if in our 5 x 10 font we have a character consisting of a one pixel horizontal line at
y == 0 the bytes will be
05 01 00 01 00 01 00 01 00 01 00
Note
If anyone knows a Python way of converting a font file (a.g. ttf) to a bitmap, please let me know. My own efforts were not good enough hence my reluctant advocacy of the above closed source program.