def parse(s, tokenize=True, tags=True, chunks=True, relations=False, lemmata=False, encoding="utf-8", **kwargs):
""" Takes a string (sentences) and returns a tagged Unicode string.
Sentences in the output are separated by newlines.
"""
if tokenize:
s = _tokenize(s)
if isinstance(s, (list, tuple)):
s = [isinstance(s, basestring) and s.split(" ") or s for s in s]
if isinstance(s, basestring):
s = [s.split(" ") for s in s.split("\n")]
# Reuse the English parser:
kwargs.update({
"lemmata": False,
"light": False,
"lexicon": LEXICON,
"language": "de",
"default": "NN",
"map": kwargs.get("tagset", "") != STTS and stts2penntreebank or None,
})
# The German lexicon uses "ss" instead of "ß".
# Instead of simply replacing it, we keep a hash map of the normalized words.
# After parsing we restore the "ß" so the output stays identical to the input.
m = dict((token.replace(u"ß", "ss"), token) for sentence in s for token in sentence)
s = [[token.replace(u"ß", "ss") for token in sentence] for sentence in s]
s = _en_parse(s, False, tags, chunks, relations, **kwargs)
p = [[[m[token[0]]] + token[1:] for token in sentence] for sentence in s.split()]
p = "\n".join([" ".join(["/".join(token) for token in sentence]) for sentence in p])
s = TaggedString(p, tags=s.tags, language="de")
# Use pattern.de.inflect for lemmatization:
if lemmata:
p = [find_lemmata(sentence) for sentence in s.split()]
s = TaggedString(p, tags=s.tags+["lemma"], language="de")
return s
def parse(s,
tokenize=True,
tags=True,
chunks=True,
relations=False,
lemmata=False,
encoding="utf-8",
**kwargs):
""" Takes a string (sentences) and returns a tagged Unicode string.
Sentences in the output are separated by newlines.
"""
if tokenize:
s = [s.split(" ") for s in _tokenize(s)]
# Reuse the English parser:
kwargs.update({
"lemmata":
False,
"light":
False,
"lexicon":
LEXICON,
"map":
kwargs.get("tagset", "").lower() != WOTAN and wotan2penntreebank
or None,
"language":
"nl"
})
s = _en_parse(s, False, tags, chunks, relations, **kwargs)
if lemmata:
p = [find_lemmata(sentence) for sentence in s.split()]
p = "\n".join([
" ".join(["/".join(token) for token in sentence]) for sentence in p
])
s = TaggedString(p, tags=s.tags + ["lemma"], language="nl")
return s
def parse(s, tokenize=True, tags=True, chunks=True, relations=False, lemmata=False, encoding="utf-8", **kwargs):
""" Takes a string (sentences) and returns a tagged Unicode string.
Sentences in the output are separated by newlines.
"""
if tokenize:
s = _tokenize(s)
if isinstance(s, (list, tuple)):
s = [isinstance(s, basestring) and s.split(" ") or s for s in s]
if isinstance(s, basestring):
s = [s.split(" ") for s in s.split("\n")]
# Reuse the English parser:
kwargs.update({
"lemmata": False,
"light": False,
"lexicon": LEXICON,
"language": "fr",
"default": "NN",
"map": None,
})
s = _en_parse(s, False, tags, chunks, relations, **kwargs)
# Use pattern.fr.inflect for lemmatization:
if lemmata:
p = [find_lemmata(sentence) for sentence in s.split()]
s = TaggedString(p, tags=s.tags+["lemma"], language="fr")
return s
def parse(s,
tokenize=True,
tags=True,
chunks=True,
relations=False,
lemmata=False,
encoding="utf-8",
**kwargs):
""" Takes a string (sentences) and returns a tagged Unicode string.
Sentences in the output are separated by newlines.
"""
if tokenize:
s = _tokenize(s)
# Reuse the English parser:
kwargs.update({
"lemmata":
False,
"light":
False,
"lexicon":
LEXICON,
"language":
"nl",
"default":
"N(soort,ev,neut)",
"map":
kwargs.get("tagset", "") != WOTAN and wotan2penntreebank or None,
})
s = _en_parse(s, False, tags, chunks, relations, **kwargs)
# Use pattern.nl.inflect for lemmatization:
if lemmata:
p = [find_lemmata(sentence) for sentence in s.split()]
s = TaggedString(p, tags=s.tags + ["lemma"], language="nl")
return s