import re
import sys
import types
import fnmatch
from os.path import basename
from pip._vendor.pygments.lexers._mapping import LEXERS
from pip._vendor.pygments.modeline import get_filetype_from_buffer
from pip._vendor.pygments.plugin import find_plugin_lexers
from pip._vendor.pygments.util import ClassNotFound, guess_decode
COMPAT = {
'Python3Lexer': 'PythonLexer',
'Python3TracebackLexer': 'PythonTracebackLexer',
}
__all__ = ['get_lexer_by_name', 'get_lexer_for_filename', 'find_lexer_class',
'guess_lexer', 'load_lexer_from_file'] + list(LEXERS) + list(COMPAT)
_lexer_cache = {}
_pattern_cache = {}
def _fn_matches(fn, glob):
if glob not in _pattern_cache:
pattern = _pattern_cache[glob] = re.compile(fnmatch.translate(glob))
return pattern.match(fn)
return _pattern_cache[glob].match(fn)
def _load_lexers(module_name):
mod = __import__(module_name, None, None, ['__all__'])
for lexer_name in mod.__all__:
cls = getattr(mod, lexer_name)
_lexer_cache[cls.name] = cls
def get_all_lexers(plugins=True):
for item in LEXERS.values():
yield item[1:]
if plugins:
for lexer in find_plugin_lexers():
yield lexer.name, lexer.aliases, lexer.filenames, lexer.mimetypes
def find_lexer_class(name):
if name in _lexer_cache:
return _lexer_cache[name]
for module_name, lname, aliases, _, _ in LEXERS.values():
if name == lname:
_load_lexers(module_name)
return _lexer_cache[name]
for cls in find_plugin_lexers():
if cls.name == name:
return cls
def find_lexer_class_by_name(_alias):
if not _alias:
raise ClassNotFound('no lexer for alias %r found' % _alias)
for module_name, name, aliases, _, _ in LEXERS.values():
if _alias.lower() in aliases:
if name not in _lexer_cache:
_load_lexers(module_name)
return _lexer_cache[name]
for cls in find_plugin_lexers():
if _alias.lower() in cls.aliases:
return cls
raise ClassNotFound('no lexer for alias %r found' % _alias)
def get_lexer_by_name(_alias, **options):
if not _alias:
raise ClassNotFound('no lexer for alias %r found' % _alias)
for module_name, name, aliases, _, _ in LEXERS.values():
if _alias.lower() in aliases:
if name not in _lexer_cache:
_load_lexers(module_name)
return _lexer_cache[name](**options)
for cls in find_plugin_lexers():
if _alias.lower() in cls.aliases:
return cls(**options)
raise ClassNotFound('no lexer for alias %r found' % _alias)
def load_lexer_from_file(filename, lexername="CustomLexer", **options):
try:
custom_namespace = {}
with open(filename, 'rb') as f:
exec(f.read(), custom_namespace)
if lexername not in custom_namespace:
raise ClassNotFound('no valid %s class found in %s' %
(lexername, filename))
lexer_class = custom_namespace[lexername]
return lexer_class(**options)
except OSError as err:
raise ClassNotFound('cannot read %s: %s' % (filename, err))
except ClassNotFound:
raise
except Exception as err:
raise ClassNotFound('error when loading custom lexer: %s' % err)
def find_lexer_class_for_filename(_fn, code=None):
matches = []
fn = basename(_fn)
for modname, name, _, filenames, _ in LEXERS.values():
for filename in filenames:
if _fn_matches(fn, filename):
if name not in _lexer_cache:
_load_lexers(modname)
matches.append((_lexer_cache[name], filename))
for cls in find_plugin_lexers():
for filename in cls.filenames:
if _fn_matches(fn, filename):
matches.append((cls, filename))
if isinstance(code, bytes):
code = guess_decode(code)
def get_rating(info):
cls, filename = info
bonus = '*' not in filename and 0.5 or 0
if code:
return cls.analyse_text(code) + bonus, cls.__name__
return cls.priority + bonus, cls.__name__
if matches:
matches.sort(key=get_rating)
return matches[-1][0]
def get_lexer_for_filename(_fn, code=None, **options):
res = find_lexer_class_for_filename(_fn, code)
if not res:
raise ClassNotFound('no lexer for filename %r found' % _fn)
return res(**options)
def get_lexer_for_mimetype(_mime, **options):
for modname, name, _, _, mimetypes in LEXERS.values():
if _mime in mimetypes:
if name not in _lexer_cache:
_load_lexers(modname)
return _lexer_cache[name](**options)
for cls in find_plugin_lexers():
if _mime in cls.mimetypes:
return cls(**options)
raise ClassNotFound('no lexer for mimetype %r found' % _mime)
def _iter_lexerclasses(plugins=True):
for key in sorted(LEXERS):
module_name, name = LEXERS[key][:2]
if name not in _lexer_cache:
_load_lexers(module_name)
yield _lexer_cache[name]
if plugins:
yield from find_plugin_lexers()
def guess_lexer_for_filename(_fn, _text, **options):
fn = basename(_fn)
primary = {}
matching_lexers = set()
for lexer in _iter_lexerclasses():
for filename in lexer.filenames:
if _fn_matches(fn, filename):
matching_lexers.add(lexer)
primary[lexer] = True
for filename in lexer.alias_filenames:
if _fn_matches(fn, filename):
matching_lexers.add(lexer)
primary[lexer] = False
if not matching_lexers:
raise ClassNotFound('no lexer for filename %r found' % fn)
if len(matching_lexers) == 1:
return matching_lexers.pop()(**options)
result = []
for lexer in matching_lexers:
rv = lexer.analyse_text(_text)
if rv == 1.0:
return lexer(**options)
result.append((rv, lexer))
def type_sort(t):
return (t[0], primary[t[1]], t[1].priority, t[1].__name__)
result.sort(key=type_sort)
return result[-1][1](**options)
def guess_lexer(_text, **options):
if not isinstance(_text, str):
inencoding = options.get('inencoding', options.get('encoding'))
if inencoding:
_text = _text.decode(inencoding or 'utf8')
else:
_text, _ = guess_decode(_text)
ft = get_filetype_from_buffer(_text)
if ft is not None:
try:
return get_lexer_by_name(ft, **options)
except ClassNotFound:
pass
best_lexer = [0.0, None]
for lexer in _iter_lexerclasses():
rv = lexer.analyse_text(_text)
if rv == 1.0:
return lexer(**options)
if rv > best_lexer[0]:
best_lexer[:] = (rv, lexer)
if not best_lexer[0] or best_lexer[1] is None:
raise ClassNotFound('no lexer matching the text found')
return best_lexer[1](**options)
class _automodule(types.ModuleType):
def __getattr__(self, name):
info = LEXERS.get(name)
if info:
_load_lexers(info[0])
cls = _lexer_cache[info[1]]
setattr(self, name, cls)
return cls
if name in COMPAT:
return getattr(self, COMPAT[name])
raise AttributeError(name)
oldmod = sys.modules[__name__]
newmod = _automodule(__name__)
newmod.__dict__.update(oldmod.__dict__)
sys.modules[__name__] = newmod
del newmod.newmod, newmod.oldmod, newmod.sys, newmod.types