How to customize the code-block language of Sphinx?

Question

Consider the below situation:

.. code-block:: my_lang

    ...

If I want to make my_lang as some stuff such as Python, how can I do it?

Answer 1

First, you need to create a script (_ext/mylanglexer.py), and then add on the extensions on the conf.py

(_ext is a convention but not a must.)

_ext/mylanglexer.py

# _ext/mylanglexer.py

from pygments.lexers import get_lexer_by_name  # refer LEXERS
from pygments.lexers._mapping import LEXERS
from pygments.lexers.python import PythonLexer


def setup(app):
    # choose one, both ok
    app.add_lexer('my_lang', get_lexer_by_name('py'))
    # app.add_lexer('my_lang', PythonLexer)

conf.py

# conf.py

extensions = [
    ...
    '_ext.mylanglexer',  # types.ModuleType, they are likely the_module = __import__('sphinx.ext.autodoc')
]

at some tutorial the tell you better add sys.path.append(os.path.abspath("./_ext")) then extensions = ['mylanglexer']

Anyway, as long as you know that is a module, all extensions should be able to be import ..., so if your module is not in the default path, of course, you must append.

Now, it's working!

---

How does it work

see the pygements.lexers.init.py

# pygements.lexers.__init__.py

def get_lexer_by_name(_alias, **options):
    ...

    # lookup builtin lexers
    for module_name, name, aliases, _, _ in LEXERS.values():  # <-- Be focus on this line.
        if _alias.lower() in aliases:
            return _lexer_cache[name](**options)  # The class object (module_name+key_name), for example: pygments.lexers.python.PythonLexer(**options)
    # continue with lexers from setuptools entrypoints
    for cls in find_plugin_lexers():

        ...
        return cls(**options)

    raise ClassNotFound('no lexer for alias %r found' % _alias)

In which LEXERS is some kind of the following stuff.

# pygments.lexers._mapping.py

LEXERS = {
    # key_name: module_name, name, aliases: Tuple[str], _, _
    ...
    'ObjectiveCLexer': ('pygments.lexers.objective', 'Objective-C', ('objective-c', 'objectivec', 'obj-c', 'objc'), ('*.m', '*.h'), ('text/x-objective-c',)),
    'ObjectiveCppLexer': ('pygments.lexers.objective', 'Objective-C++', ('objective-c++', 'objectivec++', 'obj-c++', 'objc++'), ('*.mm', '*.hh'), ('text/x-objective-c++',)),
    'ObjectiveJLexer': ('pygments.lexers.javascript', 'Objective-J', ('objective-j', 'objectivej', 'obj-j', 'objj'), ('*.j',), ('text/x-objective-j',)),
    'OcamlLexer': ('pygments.lexers.ml', 'OCaml', ('ocaml',), ('*.ml', '*.mli', '*.mll', '*.mly'), ('text/x-ocaml',)),
    'OctaveLexer': ('pygments.lexers.matlab', 'Octave', ('octave',), ('*.m',), ('text/octave',)),
    'JsonLexer': ('pygments.lexers.data', 'JSON', ('json',), ('*.json', 'Pipfile.lock'), ('application/json',)),
    'PythonLexer': ('pygments.lexers.python', 'Python', ('python', 'py', 'sage', 'python3', 'py3'), ('*.py', '*.pyw', '*.jy', '*.sage', '*.sc', 'SConstruct', 'SConscript', '*.bzl', 'BUCK', 'BUILD', 'BUILD.bazel', 'WORKSPACE', '*.tac'), ('text/x-python', 'application/x-python', 'text/x-python3', 'application/x-python3')),
    ...
}

There you know, your _alias in the aliases, then will work!

How can I customize my style?

you can copy and do some modify the Lexer, such as ObjectiveCLexer, JsonLexer ...

finally app.add_lexer('my_lang', YourLexer) that's kind of the problem.