125 lines
4.2 KiB
Python
125 lines
4.2 KiB
Python
# -*- coding: utf-8 -*-
|
||
import spacy
|
||
import textacy
|
||
from spacy.tokens import Doc
|
||
|
||
# -*- coding: utf-8 -*-
|
||
import re
|
||
import spacy
|
||
import functools
|
||
|
||
import textacy
|
||
|
||
|
||
class TextCleaner:
|
||
|
||
def __init__(self, parser, thesaurus=None, customClass_symbols=None, customClass_words=None, keep4All=None):
|
||
"""
|
||
:param parser: spacy-parser
|
||
:param thesaurus: [[syn1, syn2, ...],[syn1, syn2, ...], ...]
|
||
:param customClass_symbols:[str]
|
||
:param customClass_words:[str]
|
||
:param customClassPOS:[str]
|
||
:param keep4All: [str]
|
||
"""
|
||
if thesaurus is None:
|
||
DATAPATH_thesaurus = "openthesaurus.csv"
|
||
|
||
## !!!!!! list wichtig, da sonst nicht die gleichen Synonyme zurückgegeben werden, weil ein generator während der laufzeit pickt
|
||
self.thesaurus = list(textacy.fileio.read_csv(DATAPATH_thesaurus, delimiter=";"))
|
||
else:
|
||
self.thesaurus = thesaurus
|
||
|
||
self.parser = parser
|
||
|
||
|
||
|
||
self.whitespaceFinder = re.compile(r'(\r\n|\r|\n|(\s)+)', re.IGNORECASE)
|
||
self.mentionFinder = re.compile(r"@[a-z0-9_]{1,15}", re.IGNORECASE)
|
||
self.emailFinder = re.compile(r"\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,}\b", re.IGNORECASE)
|
||
self.urlFinder = re.compile(r"^(?:https?:\/\/)?(?:www\.)?[a-zA-Z0-9./]+$", re.IGNORECASE)
|
||
|
||
|
||
|
||
# to remove
|
||
self.symbols = ["-----", "---", "...", "“", "”", ".", "-", "<", ">", ",", "?", "!", "..", "n’t", "n't", "|", "||",
|
||
";", ":",
|
||
"…", "’s", "'s", ".", "(", ")", "[", "]", "#"] + (customClass_symbols if customClass_symbols is not None else [])
|
||
self.stop_words = list(__import__("spacy." + self.parser.lang, globals(), locals(), ['object']).STOP_WORDS)+ (customClass_words if customClass_words is not None else [])
|
||
|
||
|
||
|
||
# to keep
|
||
self.entities2keep = ["WORK_OF_ART", "ORG", "PRODUCT", "LOC"] # ,"PERSON"]
|
||
self.pos2keep = ["NOUN"] # , "NUM" ]#,"VERB","ADJ"] #fürs TopicModeling nur Nomen http://aclweb.org/anthology/U15-1013
|
||
|
||
self.entities2keep = self.entities2keep + (keep4All if keep4All is not None else [])
|
||
self.pos2keep = self.pos2keep + (keep4All if keep4All is not None else [])
|
||
|
||
|
||
keep = (keep4All if hasattr(keep4All, '__iter__') else []) + self.pos2keep + self.entities2keep
|
||
|
||
|
||
# modify those to remove with those to keep
|
||
for sym in keep:
|
||
try:
|
||
self.symbols.remove(sym)
|
||
except ValueError:
|
||
pass
|
||
for sym in keep:
|
||
try:
|
||
self.stop_words.remove(sym)
|
||
except ValueError:
|
||
pass
|
||
|
||
|
||
def loadString(self,string):
|
||
self.currentDoc = self.parser(string)
|
||
|
||
|
||
def removeWhitespace(self, string):
|
||
return " ".join([tok.text for tok in self.currentDoc if not tok.is_space])
|
||
|
||
|
||
def removePunctuation(self, string, custom_symbols=None, keep=None):
|
||
symbols = self.symbols + (custom_symbols if custom_symbols is not None else [])
|
||
if hasattr(keep, '__iter__'):
|
||
for k in keep:
|
||
try:
|
||
symbols.remove(k)
|
||
except ValueError:
|
||
pass
|
||
|
||
return " ".join([tok.text for tok in self.currentDoc if not tok.is_punct and tok.text not in symbols])
|
||
|
||
|
||
#todo funzt irgendwie nich wie's soll: https://mathieularose.com/function-composition-in-python/
|
||
parser = spacy.load('de')
|
||
cleaner = TextCleaner(parser)
|
||
corpus = textacy.Corpus(parser)
|
||
|
||
|
||
def compose(self,*functions):
|
||
return functools.reduce(lambda f, g: lambda x: f(g(x)), functions, lambda x: x)
|
||
|
||
def composeo(*functions):
|
||
return functools.reduce(lambda f, g: lambda x: f(g(x)), functions)
|
||
|
||
#pipeline = compose(functools.partial(removeWhitespace,lemmatize=True))#, cleaner.normalizeSynonyms)
|
||
|
||
pipeline = composeo(cleaner.removePunctuation, cleaner.removeWhitespace, cleaner.loadString)
|
||
|
||
def pipe1(string):
|
||
cleaner.loadString(string)
|
||
string = cleaner.removeWhitespace(string)
|
||
string = cleaner.removePunctuation(string)
|
||
return string
|
||
|
||
|
||
string = "Frau Hinrichs überdenkt die tu Situation und 545453 macht ' dann neue Anträge. \n Dieses Ticket wird geschlossen \n \n test"
|
||
print(pipe1(string))
|
||
corpus.add_text(pipeline(string))
|
||
|
||
print(corpus[0].text)
|
||
|