def htmls_to_docs(urls, htmls, text_params):
""" Given a list of urls and a dictionary with text preprocessing
parameters, return a list of extracted text for each html using.
"""
docs = []
if text_params['cache_docs']:
db = shelve.open(text_params['docs_shelve'])
for url, html in zip(urls, htmls):
if text_params['cache_docs']:
if str(url) in db:
docs.append(db[str(url)])
else:
doc = text_extraction.extract_texts(
html,
title_weight=text_params['title_weight'],
header_weights=text_params['header_weights'],
use_pdf=text_params['use_pdf'],
use_stmmer=text_params['use_stemmer'],
ukkonen_len=text_params['ukkonen_len'],
homepage_weight=text_params['homepage_weight'])
doc = ' '.join(doc)
docs.append(doc)
db[str(url)] = doc
if text_params['cache_docs']:
db.close()
return docs
def __test_site_kw(url, kws):
if len(kws) > 0:
html = crawl.extract_html_rec(url, max_depth=4)
txt = text_extraction.extract_texts(html)
txts = ' '.join(txt)
failed = False
for kw in kws:
if txts.find(text_extraction.clean_text(kw).lower()) == -1:
print 'error, keyword not found in txt', url, kw
failed = True
if not failed:
print 'OK', url
def htmls2docs(self, htmls):
""" Convert HTMLs to text documents using the params specified in
`text_params`.
"""
docs = []
for i, html in enumerate(htmls):
doc = extract_texts(
html,
title_weight=self.text_params['title_weight'],
header_weights=self.text_params['header_weights'],
use_pdf=self.text_params['use_pdf'],
use_stemmer=self.text_params['use_stemmer'],
ukkonen_len=self.text_params['ukkonen_len'],
homepage_weight=self.text_params['homepage_weight'])
docs.append(' '.join(doc))
return docs
