CODEWITHSUNDEEP
pythonfontsttx-fonttools
AlexZheda
AlexZheda

Reputation: 475

How to use pyftsubset of Fonttools inside of the python environment, not from the command line

I need to subset very many font files and I need to do that from within the python environment. Yet, Fonttools is very poorly documented and I cannot find a module and the proper function syntax to perform subsetting based on unicode from within python, not as a command line tool (pyftsubset). Some of my files contain various errors when read by the Fonttools and I cannot catch exceptions using !command inside jupyter.

Upvotes: 1

Views: 5908

Answers (3)

quick
quick

Reputation: 1204

In 2023 the proper way how to call fonttools subset from python is to use standard interface, similar to command line:

from fontTools import subset

args = [
    "font.woff2",
    "--unicodes=5f-7a,30-39,e8a6,e1b1,e5cf,e15b,e5c4,e8fd",
    "--no-layout-closure",
    "--output-file=./out.woff2",
    "--flavor=woff2",
]

subset.main(args)

Upvotes: 2

mark2mark
mark2mark

Reputation: 61

I think that should be a pythonic way to do it properly:

from fontTools import subset
subsetter = subset.Subsetter()
subsetter.populate(unicodes=["U+0020", "U+0021"])
subsetter.subset(font)

While font is your TTFont and you might need to check the documentation for how to exactly pass in the the list of unicodes. I didn’t test this exact code, but I tested it with subsetter.populate(glyphs=["a", "b"]) which does a similar job, but with glyphNames instead. The populate method can take these arguments as documented: populate(self, glyphs=[], gids=[], unicodes=[], text='')

I found a clue to that in this discussion.

Upvotes: 4

djangodude
djangodude

Reputation: 5690

pyftsubset is itself just a Python script, which calls fontTools.subset.main, which in turn parses sys.argv (command-line args) to perform subsetting. You can do the same thing pretty easily in your own script, for example:

import sys
from fontTools.subset import main as ss

sys.argv = [None, '/path/to/font/file.ttf', '--unicodes=U+0020-002F']
ss()  # this is what actually does the subsetting and writes the output file

Obviously you'll want to use your own values for --unicodes plus the numerous other pyftsubset options, but in general this scheme should work. Possible caveat is if you have other parts of your program that use/rely on sys.argv; if that's the case you might want to capture the initial values in another variable before modifying sys.argv and calling the subsetter, then re-set it to the initial values after.

Upvotes: 9

Related Questions