def __iter__(self):
"""
Iterate over the pages in a Wikipedia articles database dump (*articles.xml.bz2),
yielding one (page id, title, page content) 3-tuple at a time.
"""
try:
for title, content, page_id in extract_pages(bzip_open(self.wikicorpus.fname, mode='rt'),
self.wikicorpus.filter_namespaces):
yield (page_id, title, content)
except ValueError: # Python 2 sucks and can't open bzip in text mode
for title, content, page_id in extract_pages(bzip_open(self.wikicorpus.fname, mode='r'),
self.wikicorpus.filter_namespaces):
yield (page_id, title, content)
def __iter__(self):
"""
Iterate over the pages in a Wikipedia articles database dump (*articles.xml.bz2),
yielding one (page id, title, page content) 3-tuple at a time.
"""
if PY2 is False:
for title, content, page_id in extract_pages(
open_sesame(self.wikicorpus.fname, mode='rt'),
self.wikicorpus.filter_namespaces):
yield (page_id, title, content)
else: # Python 2 sucks and can't open bzip in text mode
for title, content, page_id in extract_pages(
open_sesame(self.wikicorpus.fname, mode='rb'),
self.wikicorpus.filter_namespaces):
yield (page_id, title, content)
def fetch_wiki_texts(in_file,
namespaces_to_filter=WIKI_DEFAULT_NAMESPACES_TO_FILTER,
min_text_length=200):
return ((title, clean_text, page_id) for title, text, page_id in
extract_pages(bz2.BZ2File(in_file), namespaces_to_filter)
for clean_text in (filter_wiki(text), )
if len(clean_text.strip()) >= min_text_length)
def get_texts(self):
articles, articles_all = 0, 0
positions, positions_all = 0, 0
texts = ((text, self.lemmatize, title, pageid)
for title, text, pageid in extract_pages(
bz2.BZ2File(self.fname), self.filter_namespaces))
pool = multiprocessing.Pool(self.processes)
# process the corpus in smaller chunks of docs, because multiprocessing.Pool
# is dumb and would load the entire input into RAM at once...
for group in utils.chunkize(texts,
chunksize=10 * self.processes,
maxsize=1):
for tokens, title, pageid in pool.imap(process_article,
group): # chunksize=10):
articles_all += 1
positions_all += len(tokens)
# article redirects and short stubs are pruned here
if len(tokens) < ARTICLE_MIN_WORDS or any(
title.startswith(ignore + ':')
for ignore in IGNORED_NAMESPACES):
continue
articles += 1
positions += len(tokens)
if self.metadata:
yield (tokens, (pageid, title))
else:
yield tokens
pool.terminate()
logger.info(
"finished iterating over Wikipedia corpus of %i documents with %i positions"
" (total %i articles, %i positions before pruning articles shorter than %i words)",
articles, positions, articles_all, positions_all,
ARTICLE_MIN_WORDS)
self.length = articles # cache corpus length
def get_wiki2():
reobj1 = re.compile(
r"[ `~!@#$%^&*\(\)-_=+\[\]\{\}\\\|;:\'\",<.>/?a-zA-Z\d]+")
reobj2 = re.compile(r"\n+")
reobj3 = re.compile("(（）)|(“”)|(「」)|(《》)|(“”)|(‘’)|(【】)|[，。？——！]{2,}")
reuseful = re.compile('^[a-zA-Z]+:')
redirect = re.compile(r"^#")
def wiki_replace(s):
s = filter_wiki(s)
s = reobj1.sub("", s) # 为上传阿里云剔除竖线(|)符号
s = reobj2.sub("#", s)
s = reobj3.sub("", s)
return s
wiki = extract_pages(bz2file.open('zhwiki-latest-pages-articles.xml.bz2'))
with codecs.open('wiki-tw.csv', 'w', encoding='utf-8') as f:
i = 0
filelist = []
for d in tqdm(wiki):
if not reuseful.findall(d[0]) and not redirect.findall(d[1]):
i += 1
filelist.append(
reobj1.sub("", d[0]) + "|" + wiki_replace(d[1]) + "\n")
if i % 1000 == 0:
s = ("".join(filelist))
f.write(s)
filelist = []
if filelist:
s = ("".join(filelist))
f.write(s)
def preprocess_sentence():
i = 0
line = ''
wiki = extract_pages(
bz2file.open('./zhwiki-20190301-pages-articles.xml.bz2')
) # 用gensim的extract_pages来提取每个页面
with open('./zhwiki_sentence.txt', 'w') as f:
for text in wiki:
if not re.findall('^[a-zA-Z]+:',
text[0]) and text[0] and not re.findall(
u'^#', text[1]): # 去掉帮助页面以及重定向的页面
converted = opencc.convert(text[1]).strip() # 繁体转简体
converted = re.sub('\|\w*\]', '', converted)
for x in converted:
if len(x.encode(
'utf-8')) == 3 and x not in stop_punctuation(
'./stop_punctuation.txt'):
line += x
if x in ['\n', '。', '？', '！', '，', '；', '：'
] and line != '\n': # 以部分中文符号为分割换行
f.write(line.strip() + '\n') # 按行存入语料文件
line = ''
i += 1
if i == 10:
print("选取中文维基百科的文章篇数:", i)
break
def get_texts(self):
"""
Iterate over the dump, returning text version of each article as a list
of tokens.
Only articles of sufficient length are returned (short articles &
redirects etc are ignored).
Note that this iterates over the **texts**; if you want vectors,
just use the standard corpus interface instead of this function::
>>> for vec in wiki_corpus:
>>> print(vec)
"""
articles, articles_all = 0, 0
positions, positions_all = 0, 0
texts = ((text, title, pageid)
for title, text, pageid in extract_pages(
bz2.BZ2File(self.fname), self.filter_namespaces))
batch_idx = 0
pool = multiprocessing.Pool(self.processes)
# Process the corpus in smaller chunks of docs,
# because multiprocessing.Pool is dumb and would load the entire input
# into RAM at once...
for group in chunkize(texts, chunksize=10 * self.processes, maxsize=1):
for tokens, title, pageid in pool.imap(process_article, group):
articles_all += 1
positions_all += len(tokens)
# article redirects and short stubs are pruned here
to_ignored = any(
title.startswith(ignore + ':')
for ignore in IGNORED_NAMESPACES)
if len(tokens) < ARTICLE_MIN_WORDS or to_ignored:
continue
articles += 1
positions += len(tokens)
if self.metadata:
yield title, tokens
else:
yield tokens
batch_idx += 1
if self.max_batch and batch_idx == self.max_batch:
break
pool.terminate()
logger.info(
"Finished iterating over Wikipedia corpus of %i documents with "
"%i positions (total %i articles, %i positions before pruning "
"articles shorter than %i words)", articles, positions,
articles_all, positions_all, ARTICLE_MIN_WORDS)
self.length = articles
def get_wiki():
from opencc import OpenCC
# 参考这篇博客注释
# https://kexue.fm/archives/4176
opencc1 = OpenCC("t2s")
resub1 = re.compile(':*{\|[\s\S]*?\|}')
resub2 = re.compile('<gallery>[\s\S]*?</gallery>')
resub3 = re.compile('(.){{([^{}\n]*?\|[^{}\n]*?)}}')
resub4 = re.compile('\* *\n|\'{2,}')
resub5 = re.compile('\n+')
resub6 = re.compile('\n[:;]|\n +')
resub7 = re.compile('\n==')
refind1 = re.compile('^[a-zA-Z]+:')
refind2 = re.compile('^#')
p1 = re.compile(r'-\{.*?(zh-hans|zh-cn):([^;]*?)(;.*?)?\}-')
p2 = re.compile(r'[（\(][，；。？！\s]*[）\)]')
p3 = re.compile(r'[「『]')
p4 = re.compile(r'[」』]')
def wiki_replace(s):
s = filter_wiki(s)
s = resub1.sub('', s)
s = resub2.sub('', s)
s = resub3.sub('\\1[[\\2]]', s)
s = resub4.sub('', s)
s = resub5.sub('\n', s)
s = resub6.sub('\n', s)
s = resub7.sub('\n\n==', s)
s = p1.sub(r'\2', s)
s = p2.sub(r'', s)
s = p3.sub(r'“', s)
s = p4.sub(r'”', s)
return opencc1.convert(s).strip()
wiki = extract_pages(
bz2file.open(
r'E:\02program\python\nlp\data\corpus\zhwiki-latest-pages-articles.xml.bz2'
))
# wiki=WikiCorpus('zhwiki-latest-pages-articles.xml.bz2',lemmatize=False,dictionary={})
with codecs.open('wiki.csv', 'w', encoding='utf-8') as f:
i = 0
filelist = []
for d in tqdm(wiki):
if not refind1.findall(d[0]) and d[0] and not refind2.findall(
d[1]):
filelist.append(d[0] + "\n" + d[1])
line = d[1]
i += 1
if i % 100 == 0:
s = wiki_replace("\n\n".join(filelist))
f.write(s)
filelist = []
def _iterate_over_pages(fname):
"""
Iterate over the pages in a Wikipedia articles database dump (*articles.xml.bz2),
yielding one (page id, title, page content) 3-tuple at a time.
"""
dictionary = Dictionary()
wiki = WikiCorpus(fname, lemmatize=False, dictionary=dictionary)
for title, content, page_id in extract_pages(bz2.BZ2File(wiki.fname), wiki.filter_namespaces):
yield (page_id, title, content)
def load_doc(dump_file, title, filter_namespaces=('0', )):
"""
Load a wikipedia article from its title.
"""
bz2_file = bz2.BZ2File(dump_file)
for page_title, text, pageid in extract_pages(bz2_file, filter_namespaces):
if page_title == title:
text = filter_wiki(text)
tokens = extract_jp_entities(text)
return tokens, pageid
return None, None
def extract_text_content(xml_dump):
# article_count = 0
with open('wiki.en.txt', 'w') as file:
for title, content, pageid in tqdm(extract_pages(xml_dump)):
try:
file.write(filter_wiki(content).strip() + "\n")
# article_count += 1
# if article_count % 10000 == 0:
# logging.info(f'{article_count} articles processed')
except Exception as e:
logging.warning(str(e))
def __iter__(self):
with bz2.BZ2File(self._dump_file) as f:
c = 0
for (title, wiki_text, wiki_id) in wikicorpus.extract_pages(f):
if any([title.lower().startswith(ns) for ns in self._ignored_ns]):
continue
c += 1
yield WikiPage(six.text_type(title), self._language, six.text_type(wiki_text))
if c % 10000 == 0:
logger.info('Processed: %d', c)
def _extract_article_onebyone(self):
wiki_pages = extract_pages(
bz2file.open(self.download_wiki_articles_dump()))
counter = 0
w = tqdm(wiki_pages, desc=u'get 0 article')
for d in w:
if not re.findall('^[a-zA-Z]+:', d[0]) and d[0] and not re.findall(
u'^#', d[1]):
yield d
counter += 1
if counter % 100 == 0:
w.set_description(u'processed %s article' % counter)
def _iterate_over_pages(fname):
"""
Iterate over the pages in a Wikipedia articles database dump (*articles.xml.bz2),
yielding one (page id, title, page content) 3-tuple at a time.
"""
dictionary = Dictionary()
wiki = WikiCorpus(fname,
lemmatize=False,
dictionary=dictionary,
filter_namespaces={'0'})
for title, content, page_id in extract_pages(bz2.BZ2File(wiki.fname),
wiki.filter_namespaces):
yield (page_id, title, content)
def main(args):
os.makedirs(args.outdir, exist_ok=True)
filter_namespaces = ('0', )
out_i = 0
for in_fname in glob.glob(args.inglob):
for title, text, pageid in extract_pages(
bz2.BZ2File(in_fname), filter_namespaces=filter_namespaces):
if out_i % args.skip == 0:
with open(os.path.join(args.outdir, f'{pageid}.txt'),
'w') as f:
f.write(filter_wiki(text))
out_i += 1
def iterate_wiki(input_path):
lemmatize = utils.has_pattern()
filter_namespaces = ('0',)
texts = ((text, lemmatize, title, pageid) for title, text, pageid in
extract_pages(bz2.BZ2File(input_path), filter_namespaces))
for article in texts:
text, lemmatize, title, pageid = article
text = utils.to_unicode(text, 'utf8', errors='ignore')
text = utils.decode_htmlentities(text) # '&nbsp;' --> '\xa0'
text = remove_markup(text)
tokens = get_all_words(text)
if len(tokens) < ARTICLE_MIN_WORDS or any(title.startswith(ignore + ':') for ignore in IGNORED_NAMESPACES):
continue
yield title, tokens
def loadSimpleWiki():
#file_name = "C:/Users/Admin/Anaconda2/envs/py27/corpora/wiki2017/simplewiki-20170820-pages-meta-current.xml.bz2"
#counter =0
#... 179,620 articles found
file_name = "C:/Users/Admin/Anaconda2/envs/py27/corpora/wiki2017/simplewiki-20170820-pages-meta-current.xml.bz2"
wiki = WikiCorpus(file_name, lemmatize=False,
dictionary={}) #vocab dict not needed
i = 0
#for text in wiki.get_texts():
allwords = defaultdict(lambda: 0)
allsents = []
print("starting...")
dfcounter_stemmed = defaultdict(lambda: 0)
dfcounter_nostem = defaultdict(lambda: 0)
for (title, article, pageid) in extract_pages(
bz2.BZ2File(file_name),
filter_namespaces=('0', )): #filter_namespaces=["0"]):
if len(article) == 0:
continue
text = process_article((filter_wiki(article), False, title, pageid),
tokenizer_func=tokenizer)
text_orig = set(
text[0][0]
) #... We are ONLY interested in whether or not the term appeared in this document NOT how many times or where
text_stemmed = set(text[0][1])
for term in text_orig:
dfcounter_nostem[term] += 1
for term in text_stemmed:
dfcounter_stemmed[term] += 1
i += 1
if i % 1000 == 0:
print(i)
#break
print(i)
handle = open("simplewiki_docfreqs_stemmed.txt", "w+")
for key, val in dfcounter_stemmed.items():
handle.write(str(key) + "\t" + str(val) + "\n")
handle.close()
handle = open("simplewiki_docfreqs_nostem.txt", "w+")
for key, val in dfcounter_nostem.items():
handle.write(str(key) + "\t" + str(val) + "\n")
handle.close()
def wiki_bz_process(self,language):
wiki = extract_pages(bz2file.open(self.corpus))
f = codecs.open(os.path.join(os.path.dirname(self.corpus), 'wiki.txt'),
'w', encoding='utf-8')
w = tqdm(wiki, desc="Currently get 0 files!")
if language=='zh':
for i, d in enumerate(w):
if not re.findall('^[a-zA-Z]+:', d[0]) and not re.findall(u'^#', d[1]):
s = self.wiki_replace(d)
f.write(s + '\n\n\n')
i += 1
if i % 100 == 0:
w.set_description('Currently got %s files' % i)
elif language=='en':
pass
def __iter__(self):
with bz2.BZ2File(self._dump_file) as f:
c = 0
for (title, wiki_text, wiki_id) in wikicorpus.extract_pages(f):
if any(
[title.lower().startswith(ns) for ns in self._ignored_ns]
):
continue
c += 1
yield WikiPage(
unicode(title), self._language, unicode(wiki_text)
)
if c % 10000 == 0:
logger.info('Processed: %d', c)
def build_dict(N):
tuple = extract_pages("enwiki-20181220-pages-articles-multistream.xml")
page_dict = {}
elect_id = random.randint(1, 500)
id = 0
cnt = 1
for t in tuple:
if (cnt > N):
break
id += 1
if (id == elect_id):
title = t[0]
interlinks = find_interlinks(str(t))
outlink_num = len(interlinks)
page_dict[title] = [cnt, outlink_num, list(interlinks.keys())]
cnt += 1
elect_id += random.randint(1, 150)
return page_dict
def wiki_process(input_file, save_path):
# wikicorpus解析
wiki = extract_pages(bz2file.open(input_file))
# 处理并导出
i = 0
f = codecs.open(save_path, 'w', encoding='utf-8')
w = tqdm(wiki, desc=u'已获取0篇文章')
openCC = OpenCC('t2s')
for d in w:
if not re.findall('^[a-zA-Z]+:', d[0]) and d[0] and not re.findall(
u'^#', d[1]):
s = wiki_replace(d)
f.write(s + '\n\n\n')
i += 1
if i % 100 == 0:
w.set_description(u'已获取%s篇文章' % i)
f.close()
def main():
parser = ArgumentParser()
parser.add_argument('-w', '--wiki-file')
args = parser.parse_args()
if args.wiki_file:
f = bz2.BZ2File(args.wiki_file)
else:
f = sys.stdin
texts = ((text, title, pageid) for title, text, pageid in extract_pages(f))
for text, title, pageid in texts:
text.replace('\n', ' ')
sys.stdout.write(text.encode('utf-8'))
sys.stdout.write('\n')
f.close()
def prepare_data(filename, destname):
pages = wc.extract_pages(bz2.BZ2File(filename), ('0',))
corpus = []
x = []
y = []
count = 0
for p in pages:
text = wc.filter_wiki(p[1])
tokens = [token.encode('utf8') for token in utils.tokenize(text, errors='ignore')
if len(token) <= 15 and not token.startswith('_')]
if len(tokens) >= 50:
length = 0
old_i = 0
for i, token in enumerate(tokens):
length += len(token)
if length > MAX_CHAR_LENGTH:
corpus.append(tokens[old_i: i])
length = len(token)
old_i = i
if i == len(tokens) - 1:
corpus.append(tokens[i:])
count = 0
for sent in corpus:
count += 1
if count >= 100000:
break
sent_y = []
sent_x = []
for token in sent:
if all([65 <= c <= 90 or 97 <= c <= 122 for c in token]):
sent_y.extend([False] * (len(token) - 1) + [True])
sent_x.extend([c - 64 if c <= 90 else c - 70 for c in token])
sent_y.extend([False] * (MAX_CHAR_LENGTH - len(sent_x)))
sent_x.extend([0] * (MAX_CHAR_LENGTH - len(sent_x)))
y.append(sent_y)
x.append(sent_x)
if len(sent_x) != MAX_CHAR_LENGTH:
print(len(sent_x))
x = np.array(x)
y = np.array(y)
pickle.dump(x, open(os.path.abspath(destname + '_x'), 'wb'))
pickle.dump(y, open(os.path.abspath(destname + '_y'), 'wb'))
def main():
parser = ArgumentParser()
parser.add_argument('-w', '--wiki-file')
args = parser.parse_args()
if args.wiki_file:
f = bz2.BZ2File(args.wiki_file)
else:
f = sys.stdin
texts = ((text, title, pageid) for title, text, pageid
in extract_pages(f))
for text, title, pageid in texts:
text.replace('\n', ' ')
sys.stdout.write(text.encode('utf-8'))
sys.stdout.write('\n')
f.close()
def get_wiki2():
from opencc import OpenCC
# 参考这篇博客注释
# https://kexue.fm/archives/4176
opencc1 = OpenCC("t2s")
reobj1 = re.compile(
"[ `~!@#$%^&*\(\)-_=+\[\]\{\}\\\|;:\'\",<.>/?a-zA-Z\d]+")
reobj2 = re.compile(r"\n+")
reobj3 = re.compile("(（）)|(“”)|(「」)|(《》)|(“”)|(‘’)|(【】)|[，。？——！]{2,}")
reuseful = re.compile('^[a-zA-Z]+:')
redirect = re.compile(r"^#")
def wiki_replace(s):
s = filter_wiki(s)
s = reobj1.sub("", s) # 为上传阿里云剔除竖线(|)符号
s = reobj2.sub("#", s)
s = reobj3.sub("", s)
return opencc1.convert(s).strip()
wiki = extract_pages(
bz2file.open(
r'E:\02program\python\nlp\data\corpus\zhwiki-latest-pages-articles.xml.bz2'
))
with codecs.open('wiki-tw.csv', 'w', encoding='utf-8') as f:
i = 0
filelist = []
for d in tqdm(wiki):
if not reuseful.findall(d[0]) and not redirect.findall(d[1]):
i += 1
try:
filelist.append(
reobj1.sub("", d[0]) + "|" + wiki_replace(d[1]) + "\n")
except Exception as e:
print(d[0], '=' * 10, d[1])
if i % 1000 == 0:
s = ("".join(filelist))
f.write(s)
filelist = []
if filelist:
s = ("".join(filelist))
f.write(s)
def PreProcessing():
print('begin time of the program is: ', time.ctime())
tuPle = WIKI.extract_pages(xml_path)
global tot_word
#read the xml file into the tuple which is read as type yield
cnt_time = 0
while cnt_time < tot_number:
curr_page = next(tuPle)
redirects = [
redirect for keyword, redirect in WIKI.find_interlinks(
curr_page[1]).items()
]
cnt_time += 1
# extract the title and the redirect title
curr_title = curr_page[0]
if curr_title not in WordDict:
WordDict[curr_title] = tot_word
NumDict[tot_word] = curr_title
tot_word += 1
org_id = WordDict[curr_title]
# set the id of the word
# sum_redirect = len(redirects)
for redirect_title in redirects:
if redirect_title not in WordDict:
WordDict[redirect_title] = tot_word
NumDict[tot_word] = redirect_title
#link_id = WordDict[redirect_title]
tot_word += 1
link_id = WordDict[redirect_title]
if org_id not in OutLink:
OutLink[org_id] = []
OutLink[org_id].append(link_id)
if link_id not in InLink:
InLink[link_id] = []
InLink[link_id].append(org_id)
#addtwodimdict(RankScore,org_id,link_id,1/sum_redirect)
print('end time of the pre-processing is: ', time.ctime())
def wikipedia_extract(input_file, output_dir):
#character chunk
chunk_size = 50000000
os.makedirs(output_dir, exist_ok=True)
fin = xutils.open_file(input_file, 'rt')
extractor = extract_pages(fin, ['0'])
fout, counter, chunk = None, chunk_size, -1
for page in extractor:
if page[1]:
text = filter_wiki(page[1])
if counter >= chunk_size:
if fout: fout.close()
counter, chunk = 0, chunk + 1
output_file = '%s/%s_%d.txt.gz' % (
output_dir, os.path.basename(input_file), chunk)
fout = xutils.open_file(output_file, 'wt')
print(output_file)
counter += len(text)
fout.write(text)
fout.write('\n\n\n\n')
fin.close()
def summarize_wiki():
ignore_namespaces = 'Wikipedia Category File Portal Template MediaWiki User Help Book Draft'.split(
)
t1 = time.time()
successful = 0
failed = 0
timedout = 0
extracted_pages = extract_pages(bz2.BZ2File(wiki_fname), ('0', ))
for title, text, pageid in extracted_pages:
if any(title.startswith(ignore + ':') for ignore in ignore_namespaces):
continue
try:
summarize_timeout(text)
successful += 1
except TimeoutError:
print "Timeout summarizing article", title, "with id", pageid
timedout += 1
except RuntimeError:
failed += 1
if (successful + failed) % 1000 == 0:
print "Article", successful + failed, "summarized."
time.sleep(1)
t2 = time.time()
print "Successful summaries:", successful
print "Failed summaries:", failed
print "Timeout summaries:", timedout
print "t1:", t1
print "t2", t2
print "dt", t2 - t1
proper_names_dict = {
'ουσιαστικό': 'nouns',
'επίθετο': 'adjectives',
'άρθρο': 'dets',
'επίρρημα': 'adverbs',
'κύριο όνομα': 'proper_names',
'μετοχή': 'participles',
'ρήμα': 'verbs'
}
expected_parts_dict = {}
for expected_part in expected_parts:
expected_parts_dict[expected_part] = []
other_parts = {}
for title, text, pageid in extract_pages(wiktionary_file_path):
if text.startswith('#REDIRECT'):
continue
title = title.lower()
all_regex = regex.findall(text)
all_regex.extend(regex2.findall(text))
for a in all_regex:
if a in expected_parts:
expected_parts_dict[a].append(title)
for i in expected_parts_dict:
with open('_{0}.py'.format(proper_names_dict[i]), 'w') as f:
f.write('from __future__ import unicode_literals\n')
f.write('{} = set(\"\"\"\n'.format(proper_names_dict[i].upper()))
words = sorted(expected_parts_dict[i])
line = ''
argparser = argparse.ArgumentParser(description='Wikipedia Dump Extractor')
argparser.add_argument('-input_path', type=str, required=True, help='Path to the raw Wikipedia dump')
argparser.add_argument('-output_path', type=str, required=True, help='Write path for extracted text content')
return argparser.parse_args()
if __name__ == '__main__':
arguments = config_argparser()
program = os.path.basename(sys.argv[0])
logger = logging.getLogger(program)
logging.basicConfig(format='%(asctime)s: %(levelname)s: %(message)s')
logging.root.setLevel(level=logging.INFO)
lemmatize = utils.has_pattern()
filter_namespaces = ('0',)
texts = ((text, lemmatize, title, pageid) for title, text, pageid in
extract_pages(bz2.BZ2File(arguments.input_path), filter_namespaces))
parsed_article_counter = 0
space = u' '
output = codecs.open(arguments.output_path, 'w', 'utf-8')
for article in texts:
text, lemmatize, title, pageid = article
text = utils.to_unicode(text, 'utf8', errors='ignore')
text = utils.decode_htmlentities(text) # '&nbsp;' --> '\xa0'
text = remove_markup(text)
tokens = get_all_words(text)
if len(tokens) < ARTICLE_MIN_WORDS or any(title.startswith(ignore + ':') for ignore in IGNORED_NAMESPACES):
continue
output.write("{}\n".format(space.join(tokens) + "\n"))
parsed_article_counter += 1
print('Parsed articles: {}', parsed_article_counter)
import bz2file
import re
from opencc import OpenCC
from tqdm import tqdm
import codecs
import sys
if len(sys.argv) < 3:
print("argc less 3")
sys.exit(1)
inp, outp = sys.argv[1:3]
print("inp file:%s" % inp)
print("outp file:%s" % outp)
wiki = extract_pages(bz2file.open(inp))
c = OpenCC('t2s')
def wiki_replace(d):
s = d[1]
s = re.sub(':*{\|[\s\S]*?\|}', '', s)
s = re.sub('<gallery>[\s\S]*?</gallery>', '', s)
s = re.sub('(.){{([^{}\n]*?\|[^{}\n]*?)}}', '\\1[[\\2]]', s)
s = filter_wiki(s)
s = re.sub('\* *\n|\'{2,}', '', s)
s = re.sub('\n+', '\n', s)
s = re.sub('\n[:;]|\n +', '\n', s)
s = re.sub('\n==', '\n\n==', s)
s = u'【' + d[0] + u'】\n' + s
#print("string:%s" % s)
import pickle
def fantojan_savepickle(input):
#this function will save every line into a pickle
i = 0
w = tqdm(input, desc=u'已获取0篇文章') #tqdm is for progressbar prompt
for d in w:
if not re.findall('^[a-zA-Z]+:', d[0]) and d[0] and not re.findall(
u'^#', d[1]):
outputfile = codecs.open(
'/Users/wisdombeat/PycharmProjects/wiki_txt/wikis/line' +
str(i) + '.pkl', 'wb+')
s = wiki_replace(d)
pickle.dump(s, outputfile)
outputfile.close()
i += 1
if i % 100 == 0:
w.set_description(u'已获取%s篇文章' % i)
w.set_description(u'All have done, there are %s papers' % i)
return
if __name__ == "__main__":
openbz2 = extract_pages(
bz2file.open(
'/Users/wisdombeat/Desktop/zhwiki-latest-pages-articles.xml.bz2'))
#resultname = '/Users/wisdombeat/PycharmProjects/wiki_txt/wiki.txt'
#fantojan_savetxt(input=openbz2, output=resultname)
fantojan_savepickle(openbz2)
proper_names_dict={
'ουσιαστικό':'nouns',
'επίθετο':'adjectives',
'άρθρο':'dets',
'επίρρημα':'adverbs',
'κύριο όνομα': 'proper_names',
'μετοχή': 'participles',
'ρήμα': 'verbs'
}
expected_parts_dict = {}
for expected_part in expected_parts:
expected_parts_dict[expected_part] = []
other_parts = {}
for title, text, pageid in extract_pages(wiktionary_file_path):
if text.startswith('#REDIRECT'):
continue
title = title.lower()
all_regex = regex.findall(text)
all_regex.extend(regex2.findall(text))
for a in all_regex:
if a in expected_parts:
expected_parts_dict[a].append(title)
for i in expected_parts_dict:
with open('_{0}.py'.format(proper_names_dict[i]), 'w') as f:
f.write('from __future__ import unicode_literals\n')
f.write('{} = set(\"\"\"\n'.format(proper_names_dict[i].upper()))
words = sorted(expected_parts_dict[i])
def get_pos_from_wiktionary():
import re
from gensim.corpora.wikicorpus import extract_pages
regex = re.compile(r"==={{(\w+)\|el}}===")
regex2 = re.compile(r"==={{(\w+ \w+)\|el}}===")
# get words based on the Wiktionary dump
# check only for specific parts
# ==={{κύριο όνομα|el}}===
expected_parts = [
"μετοχή",
"ρήμα",
"επίθετο",
"επίρρημα",
"ουσιαστικό",
"κύριο όνομα",
"άρθρο",
]
wiktionary_file_path = (
"/data/gsoc2018-spacy/spacy/lang/el/res/elwiktionary-latest-pages-articles.xml"
)
proper_names_dict = {
"ουσιαστικό": "nouns",
"επίθετο": "adjectives",
"άρθρο": "dets",
"επίρρημα": "adverbs",
"κύριο όνομα": "proper_names",
"μετοχή": "participles",
"ρήμα": "verbs",
}
expected_parts_dict = {}
for expected_part in expected_parts:
expected_parts_dict[expected_part] = []
for title, text, pageid in extract_pages(wiktionary_file_path):
if text.startswith("#REDIRECT"):
continue
title = title.lower()
all_regex = regex.findall(text)
all_regex.extend(regex2.findall(text))
for a in all_regex:
if a in expected_parts:
expected_parts_dict[a].append(title)
for i in expected_parts_dict:
with open("_{0}.py".format(proper_names_dict[i]), "w") as f:
f.write("from __future__ import unicode_literals\n")
f.write('{} = set("""\n'.format(proper_names_dict[i].upper()))
words = sorted(expected_parts_dict[i])
line = ""
to_write = []
for word in words:
if len(line + " " + word) > 79:
to_write.append(line)
line = ""
else:
line = line + " " + word
f.write("\n".join(to_write))
f.write('\n""".split())')
from db_connect import get_cursor
dbc = get_cursor()
'''
CREATE TABLE `wiki_pages` (
`id` varchar(255) NOT NULL,
`title` varchar(255) NOT NULL,
`content` mediumtext NOT NULL,
`is_artist` tinyint(1) NOT NULL,
`size` int(11) NOT NULL
) DEFAULT CHARSET=utf8;
ALTER TABLE `wiki_pages`
ADD PRIMARY KEY (`id`),
ADD KEY `is_artist` (`is_artist`),
ADD KEY `size` (`size`),
ADD KEY `title` (`title`);
'''
with open('enwiki-latest-pages-articles.xml') as fh:
gen = corpus.extract_pages(fh)
i = 0
for title, text, pgid in gen:
text = text.lower()
if 'infobox musical artist' in text and ':' not in title:
dbc.execute('INSERT INTO wiki_pages (id, title, content, is_artist) VALUES(%s, %s, %s, 1)',
[pgid, title, text])
print i, title
i += 1
import sys
sys.path.remove(
'/System/Library/Frameworks/Python.framework/Versions/2.7/Extras/lib/python'
)
#sys.path.remove('/System/Library/Frameworks/Python.framework/Versions/2.7/Extras/lib/python/PyObjC')
sys.path.append('/Users/xujiaxing/anaconda/lib/python3.6/site-packages')
#sys.path.append('/Users/xujiaxing/anaconda/lib/python3.6/site-packages/opencc.py')
from gensim.corpora.wikicorpus import extract_pages, filter_wiki
import bz2file
import re
import opencc
from tqdm import tqdm
import codecs
wiki = extract_pages(bz2file.open('zhwiki-latest-pages-articles.xml.bz2'))
def wiki_replace(d):
s = d[1]
s = re.sub(':*{\|[\s\S]*?\|}', '', s)
s = re.sub('<gallery>[\s\S]*?</gallery>', '', s)
s = re.sub('(.){{([^{}\n]*?\|[^{}\n]*?)}}', '\\1[[\\2]]', s)
s = filter_wiki(s)
s = re.sub('\* *\n|\'{2,}', '', s)
s = re.sub('\n+', '\n', s)
s = re.sub('\n[:;]|\n +', '\n', s)
s = re.sub('\n==', '\n\n==', s)
s = u'【' + d[0] + u'】\n' + s
return opencc.convert(s).strip()
# -*- coding: utf-8 -*-
"""
Created on Sat Nov 16 21:09:51 2019
@author: us
"""
from gensim.corpora.wikicorpus import extract_pages,filter_wiki
import bz2file
import re
#import opencc
from opencc import OpenCC
from tqdm import tqdm
import codecs
cc = OpenCC('t2s') # convert from Simplified Chinese to Traditional Chinese
wiki = extract_pages(bz2file.open('zhwiki-20190720-pages-articles.xml.bz2'))
def wiki_replace(d):
s = d[1]
s = re.sub(':*{\|[\s\S]*?\|}', '', s)
s = re.sub('<gallery>[\s\S]*?</gallery>', '', s)
s = re.sub('(.){{([^{}\n]*?\|[^{}\n]*?)}}', '\\1[[\\2]]', s)
s = filter_wiki(s)
s = re.sub('\* *\n|\'{2,}', '', s)
s = re.sub('\n+', '\n', s)
s = re.sub('\n[:;]|\n +', '\n', s)
s = re.sub('\n==', '\n\n==', s)
s = u'【' + d[0] + u'】\n' + s
return cc.convert(s).strip()
i = 0
wiki = gensim.corpora.WikiCorpus(input_file, lemmatize=False, dictionary={})
for text in tqdm(list(wiki.get_texts())):
str_line = bytes.join(b' ', text).decode()
f.write(str_line + '\n')
# from https://spaces.ac.cn/archives/4176/
from gensim.corpora.wikicorpus import extract_pages, filter_wiki
import bz2file
import re
import opencc
from tqdm import tqdm
import codecs
input_file = "E:\matt\get\wiki\zhwiki-20180301-pages-articles-multistream.xml.bz2"
wiki = extract_pages(bz2file.open(input_file))
def wiki_replace(d):
s = d[1]
s = re.sub(':*{\|[\s\S]*?\|}', '', s)
s = re.sub('<gallery>[\s\S]*?</gallery>', '', s)
s = re.sub('(.){{([^{}\n]*?\|[^{}\n]*?)}}', '\\1[[\\2]]', s)
s = filter_wiki(s)
s = re.sub('\* *\n|\'{2,}', '', s)
s = re.sub('\n+', '\n', s)
s = re.sub('\n[:;]|\n +', '\n', s)
s = re.sub('\n==', '\n\n==', s)
s = u'【' + d[0] + u'】\n' + s
return opencc.convert(s).strip() #