glyphsLib

This Python library provides a bridge from Glyphs source files (.glyphs) to UFOs and Designspace files via defcon and designspaceLib.

The main methods for conversion are found in __init__.py. Intermediate data can be accessed without actually writing UFOs, if needed.

Write and return UFOs

The following code will write UFOs and a Designspace file to disk.

import glyphsLib

master_dir = "master_ufos"
ufos, designspace_path = glyphsLib.build_masters("MyFont.glyphs", master_dir)

If you want to interpolate instances, please use fontmake instead. It uses this library under the hood when dealing with Glyphs files.

Load UFO objects without writing

import glyphsLib

ufos = glyphsLib.load_to_ufos("MyFont.glyphs")

Read and write Glyphs data as Python objects

from glyphsLib import GSFont

font = GSFont(glyphs_file)
font.save(glyphs_file)

The glyphsLib.classes module aims to provide an interface similar to Glyphs.app's Python Scripting API.

Note that currently not all the classes and methods may be fully implemented. We try to keep up to date, but if you find something that is missing or does not work as expected, please open a issue.

Go back and forth between UFOs and Glyphs

1. You can use the ufo2glyphs and glyphs2ufo command line scripts to round-trip your source files. By default, the scripts try to preserve as much metadata as possible.

   # Generate master UFOs and Designspace file
   glyphs2ufo Example.glyphs
   
   # Go back
   ufo2glyphs Example.designspace
   
   # You can also combine single UFOs into a Glyphs source file.
   ufo2glyphs Example-Regular.ufo Example-Bold.ufo
   

2. Without a designspace file, using for example the Inria fonts by Black[Foundry]:

   import glob
   from defcon import Font
   from glyphsLib import to_glyphs
   
   ufos = [Font(path) for path in glob.glob("*Italic.ufo")]
   # Sort the UFOs because glyphsLib will create masters in the same order
   ufos = sorted(ufos, key=lambda ufo: ufo.info.openTypeOS2WeightClass)
   font = to_glyphs(ufos)
   font.save("InriaSansItalic.glyphs")
   

   Here is the resulting glyphs file

3. With a designspace, using Spectral from Production Type:

   import glob
   from fontTools.designspaceLib import DesignSpaceDocument
   from glyphsLib import to_glyphs
   
   doc = DesignSpaceDocument()
   doc.read("spectral-build-roman.designspace")
   font = to_glyphs(doc)
   font.save("SpectralRoman.glyphs")
   

   Here is the resulting glyphs file

4. In both programmatic cases, if you intend to go back to UFOs after modifying the file with Glyphs, you should use the minimize_ufo_diffs parameter to minimize the amount of diffs that will show up in git after the back and forth. To do so, the glyphsLib will add some bookkeeping values in various userData fields. For example, it will try to remember which GSClass came from groups.plist or from the feature file.

The same option exists for people who want to do Glyphs->UFOs->Glyphs: minimize_glyphs_diffs, which will add some bookkeeping data in UFO lib. For example, it will keep the same UUIDs for Glyphs layers, and so will need to store those layer UUIDs in the UFOs.

import glob
import os
from fontTools.designspaceLib import DesignSpaceDocument
from glyphsLib import to_glyphs, to_designspace, GSFont

doc = DesignSpaceDocument()
doc.read("spectral-build-roman.designspace")
font = to_glyphs(doc, minimize_ufo_diffs=True)
doc2 = to_designspace(font, propagate_anchors=False)
# UFOs are in memory only, attached to the doc via `sources`
# Writing doc2 over the original doc should generate very few git diffs (ideally none)
doc2.write(doc.path)
for source in doc2.sources:
    path = os.path.join(os.path.dirname(doc.path), source.filename)
    # You will want to use ufoNormalizer after
    source.font.save(path)

font = GSFont("SpectralRoman.glyphs")
doc = to_designspace(font, minimize_glyphs_diffs=True, propagate_anchors=False)
font2 = to_glyphs(doc)
# Writing font2 over font should generate very few git diffs (ideally none):
font2.save(font.filepath)

In practice there are always a few diffs on things that don't really make a difference, like optional things being added/removed or whitespace changes or things getting reordered...

Make a release

Use git tag -a to make a new annotated tag, or git tag -s for a GPG-signed annotated tag, if you prefer.

Name the new tag with with a leading ‘v’ followed by three MAJOR.MINOR.PATCH digits, like in semantic versioning. Look at the existing tags for examples.

In the tag message write some short release notes describing the changes since the previous tag.

Finally, push the tag to the remote repository (e.g. assuming your upstream is called origin):

$ git push origin v0.4.3

This will trigger the CI to build the distribution packages and upload them to the Python Package Index automatically, if all the tests pass successfully.