def get_tokenizer(tag,chunkers=None,filters=None):
"""Locate an appropriate tokenizer by language tag.
This requires importing the function 'tokenize' from an
appropriate module. Modules tried are named after the
language tag, tried in the following order:
* the entire tag (e.g. "en_AU.py")
* the base country code of the tag (e.g. "en.py")
If a suitable function cannot be found, raises TokenizerNotFoundError.
If given and not None, 'chunkers' and 'filters' must be lists of chunker
classes and filter classes resectively. These will be applied to the
tokenizer during creation.
"""
# "filters" used to be the second argument. Try to catch cases
# where it is given positionally and issue a DeprecationWarning.
if chunkers is not None and filters is None:
chunkers = list(chunkers)
if chunkers:
try:
chunkers_are_filters = issubclass(chunkers[0],Filter)
except TypeError:
pass
else:
if chunkers_are_filters:
msg = "passing 'filters' as a non-keyword argument "\
"to get_tokenizer() is deprecated"
warnings.warn(msg,category=DeprecationWarning)
filters = chunkers
chunkers = None
# Ensure only '_' used as separator
tag = tag.replace("-","_")
# First try the whole tag
tkFunc = _try_tokenizer(tag)
if tkFunc is None:
# Try just the base
base = tag.split("_")[0]
tkFunc = _try_tokenizer(base)
if tkFunc is None:
msg = "No tokenizer found for language '%s'" % (tag,)
raise TokenizerNotFoundError(msg)
# Given the language-specific tokenizer, we now build up the
# end result as follows:
# * chunk the text using any given chunkers in turn
# * begin with basic whitespace tokenization
# * apply each of the given filters in turn
# * apply language-specific rules
tokenizer = basic_tokenize
if chunkers is not None:
chunkers = list(chunkers)
for i in xrange(len(chunkers)-1,-1,-1):
tokenizer = wrap_tokenizer(chunkers[i],tokenizer)
if filters is not None:
for f in filters:
tokenizer = f(tokenizer)
tokenizer = wrap_tokenizer(tokenizer,tkFunc)
return tokenizer
def get_tokenizer(tag=None, chunkers=None, filters=None):
"""Locate an appropriate tokenizer by language tag.
This requires importing the function 'tokenize' from an appropriate
module. Modules tried are named after the language tag, tried in the
following order:
* the entire tag (e.g. "en_AU.py")
* the base country code of the tag (e.g. "en.py")
If the language tag is None, a default tokenizer (actually the English
one) is returned. It's unicode aware and should work OK for most
latin-derived languages.
If a suitable function cannot be found, raises TokenizerNotFoundError.
If given and not None, 'chunkers' and 'filters' must be lists of chunker
classes and filter classes respectively. These will be applied to the
tokenizer during creation.
"""
if tag is None:
tag = "en"
# "filters" used to be the second argument. Try to catch cases
# where it is given positionally and issue a DeprecationWarning.
if chunkers is not None and filters is None:
chunkers = list(chunkers)
if chunkers:
try:
chunkers_are_filters = issubclass(chunkers[0], Filter)
except TypeError:
pass
else:
if chunkers_are_filters:
msg = "passing 'filters' as a non-keyword argument "\
"to get_tokenizer() is deprecated"
warnings.warn(msg,
category=DeprecationWarning,
stacklevel=2)
filters = chunkers
chunkers = None
# Ensure only '_' used as separator
tag = tag.replace("-", "_")
# First try the whole tag
tkFunc = _try_tokenizer(tag)
if tkFunc is None:
# Try just the base
base = tag.split("_")[0]
tkFunc = _try_tokenizer(base)
if tkFunc is None:
msg = "No tokenizer found for language '%s'" % (tag, )
raise TokenizerNotFoundError(msg)
# Given the language-specific tokenizer, we now build up the
# end result as follows:
# * chunk the text using any given chunkers in turn
# * begin with basic whitespace tokenization
# * apply each of the given filters in turn
# * apply language-specific rules
tokenizer = basic_tokenize
if chunkers is not None:
chunkers = list(chunkers)
for i in xrange(len(chunkers) - 1, -1, -1):
tokenizer = wrap_tokenizer(chunkers[i], tokenizer)
if filters is not None:
for f in filters:
tokenizer = f(tokenizer)
tokenizer = wrap_tokenizer(tokenizer, tkFunc)
return tokenizer
