class Wanish():
def __init__(self,
url=None,
positive_keywords=None,
negative_keywords=None,
summary_sentences_qty=5,
headers=None):
"""
Initialization of the class. If url is set, it gets performed.
:param url: web-page url of the document
:param positive_keywords: list of keywords, which are likely to be seen in classes or ids of tags
:param negative_keywords: list of keywords, which are unlikely to be seen in classes or ids of tags
:param summary_sentences_qty: maximum quantity of summary sentences
:param headers: custom headers for GET request to obtain web page of the article
"""
self._article_extractor = ArticleExtractor(
positive_keywords=positive_keywords,
negative_keywords=negative_keywords)
self.url = None # source web-page url
self.title = None # document's title
self.image_url = None # document's image url
self.language = None # document's article language
self.clean_html = None # cleaned html of the article
self.description = None # summarized description (text only)
self.error_msg = None # error message
self._source_html = None # source html of the document (lxml doc)
self._charset = None # source html encoding
self._headers = headers if type(headers) == dict else {
} # custom headers for GET request
# summarized text sentences quantity
try:
self._summary_sentences_qty = summary_sentences_qty
except (TypeError, ValueError):
self._summary_sentences_qty = 5
# perform the url if defined
if url:
self.perform_url(url)
def perform_url(self, url):
"""
Perform an article document by designated url
:param url: web-page url of the document
"""
self.url = url
self.title = self.image_url = self.language = self.description = \
self.clean_html = self.error_msg = self._charset = None
if not self.url:
self.error_msg = 'Empty or null URL to perform'
return
# get the page (bytecode)
try:
web_page = requests.get(self.url, headers=self._headers)
# perform http status codes
if web_page.status_code not in [200, 301, 302]:
self.error_msg = str('HTTP Error. Status: %s' %
web_page.status_code)
return
self.url = web_page.url
raw_html = web_page.content
self._charset = get_encoding(raw_html)
raw_html_str = raw_html.decode(self._charset)
# getting and cleaning the document
self._source_html = document_fromstring(raw_html_str)
self._source_html = html_cleaner.clean_html(self._source_html)
# making links absolute
self._source_html.make_links_absolute(self.url,
resolve_base_href=True)
except (ConnectionError, Timeout, TypeError, Exception) as e:
self.error_msg = str(e)
finally:
if self.error_msg:
return
if self._source_html is not None:
# obtaining title
self.title = shorten_title(self._source_html)
# obtaining image url
self.image_url = get_image_url(self._source_html, self.url)
if self.image_url is not None:
image_url_node = "<meta itemprop=\"image\" content=\"%s\">" % self.image_url
image_url_img = "<img src=\"%s\" />" % self.image_url
else:
image_url_node = image_url_img = ""
# clean html
self.clean_html = self._article_extractor.get_clean_html(
source_html=self._source_html)
# summarized description, requires clean_html
if self.clean_html:
self.description, self.language = get_plain_text(
etree.XML(self.clean_html), self._summary_sentences_qty)
description_node = "<meta name=\"description\" content=\"%s\">" if self.description else ""
# filling the template
self.clean_html = ARTICLE_TEMPLATE % {
'language': self.language,
'title': self.title,
'image_url_node': image_url_node,
'image_url_img': image_url_img,
'description_node': description_node,
'clean_html': self.clean_html
}
class Wanish():
def __init__(self, url=None, positive_keywords=None, negative_keywords=None, summary_sentences_qty=5, headers=None):
"""
Initialization of the class. If url is set, it gets performed.
:param url: web-page url of the document
:param positive_keywords: list of keywords, which are likely to be seen in classes or ids of tags
:param negative_keywords: list of keywords, which are unlikely to be seen in classes or ids of tags
:param summary_sentences_qty: maximum quantity of summary sentences
:param headers: custom headers for GET request to obtain web page of the article
"""
self._article_extractor = ArticleExtractor(positive_keywords=positive_keywords,
negative_keywords=negative_keywords)
self.url = None # source web-page url
self.title = None # document's title
self.image_url = None # document's image url
self.language = None # document's article language
self.clean_html = None # cleaned html of the article
self.description = None # summarized description (text only)
self.error_msg = None # error message
self._source_html = None # source html of the document (lxml doc)
self._charset = None # source html encoding
self._headers = headers if type(headers) == dict else {} # custom headers for GET request
# summarized text sentences quantity
try:
self._summary_sentences_qty = summary_sentences_qty
except (TypeError, ValueError):
self._summary_sentences_qty = 5
# perform the url if defined
if url:
self.perform_url(url)
def perform_url(self, url):
"""
Perform an article document by designated url
:param url: web-page url of the document
"""
self.url = url
self.title = self.image_url = self.language = self.description = \
self.clean_html = self.error_msg = self._charset = None
if not self.url:
self.error_msg = 'Empty or null URL to perform'
return
# get the page (bytecode)
try:
web_page = requests.get(self.url, headers=self._headers)
# perform http status codes
if web_page.status_code not in [200, 301, 302]:
self.error_msg = str('HTTP Error. Status: %s' % web_page.status_code)
return
self.url = web_page.url
raw_html = web_page.content
self._charset = get_encoding(raw_html)
raw_html_str = raw_html.decode(self._charset)
# getting and cleaning the document
self._source_html = document_fromstring(raw_html_str)
self._source_html = html_cleaner.clean_html(self._source_html)
# making links absolute
self._source_html.make_links_absolute(self.url, resolve_base_href=True)
except (ConnectionError, Timeout, TypeError, Exception) as e:
self.error_msg = str(e)
finally:
if self.error_msg:
return
if self._source_html is not None:
# obtaining title
self.title = shorten_title(self._source_html)
# obtaining image url
self.image_url = get_image_url(self._source_html, self.url)
if self.image_url is not None:
image_url_node = "<meta itemprop=\"image\" content=\"%s\">" % self.image_url
image_url_img = "<img src=\"%s\" />" % self.image_url
else:
image_url_node = image_url_img = ""
# clean html
self.clean_html = self._article_extractor.get_clean_html(source_html=self._source_html)
# summarized description, requires clean_html
if self.clean_html:
self.description, self.language = get_plain_text(etree.XML(self.clean_html),
self._summary_sentences_qty)
description_node = "<meta name=\"description\" content=\"%s\">" if self.description else ""
# filling the template
self.clean_html = ARTICLE_TEMPLATE % {
'language': self.language,
'title': self.title,
'image_url_node': image_url_node,
'image_url_img': image_url_img,
'description_node': description_node,
'clean_html': self.clean_html
}
class Wanish(object):
def __init__(self, url=None, positive_keywords=None, negative_keywords=None, summary_sentences_qty=5, headers=None):
"""
Initialization of the class. If url is set, it gets performed.
:param url: web-page url of the document
:param positive_keywords: list of keywords, which are likely to be seen in classes or ids of tags
:param negative_keywords: list of keywords, which are unlikely to be seen in classes or ids of tags
:param summary_sentences_qty: maximum quantity of summary sentences
:param headers: custom headers for GET request to obtain web page of the article
"""
self._article_extractor = ArticleExtractor(positive_keywords=positive_keywords,
negative_keywords=negative_keywords)
self.url = None # source web-page url
self.canonical_url = None # canonical web-page url if present, otherwise same as url
self.title = None # document's title
self.image_url = None # document's image url
self.language = None # document's article language
self.clean_html = None # cleaned html of the article
self.description = None # summarized description (text only)
self.error_msg = None # error message
self._source_html = None # source html of the document (lxml doc)
self._charset = None # source html encoding
self._headers = headers if type(headers) == dict else {} # custom headers for GET request
# summarized text sentences quantity
try:
self._summary_sentences_qty = summary_sentences_qty
except (TypeError, ValueError):
self._summary_sentences_qty = 5
# perform the url if defined
if url:
self.perform_url(url)
def perform_url(self, url):
"""
Perform an article document by designated url
:param url: web-page url of the document
"""
self.url = url
self.title = self.image_url = self.language = self.description = \
self.clean_html = self.error_msg = self._charset = None
if not self.url:
self.error_msg = 'Empty or null URL to perform'
return
# get the page (bytecode)
try:
web_page = requests.get(self.url, headers=self._headers)
# perform http status codes
if web_page.status_code not in [200, 301, 302]:
self.error_msg = str('HTTP Error. Status: %s' % web_page.status_code)
return
self.url = web_page.url
raw_html = web_page.content
self._charset = get_encoding(raw_html)
our_parser = XMLParser(encoding=self._charset, recover=True)
# getting and cleaning the document
self._source_html = parse(BytesIO(raw_html), parser=our_parser)
self._source_html = fromstring(tostring(self._source_html))
# searching for canonical url
link_canonicals = self._source_html.xpath("//link[normalize-space(@rel)='canonical']/@href")
self.canonical_url = link_canonicals[0] if len(link_canonicals) > 0 else self.url
self._source_html = html_cleaner.clean_html(self._source_html)
# making links absolute
self._source_html.make_links_absolute(self.url, resolve_base_href=True)
strip_elements(self._source_html, 'blockquote', 'code', 'table', 'ol', 'ul',
'embedded', 'input', 'address', 'iframe', 'textarea', 'dl')
except (ConnectionError, Timeout, TypeError, Exception) as e:
self.error_msg = str(e)
finally:
if self.error_msg:
return
if self._source_html is not None:
# obtaining title
self.title = shorten_title(self._source_html)
# obtaining image url
self.image_url = get_image_url(self._source_html, self.url)
if self.image_url is not None:
image_url_node = "<meta itemprop=\"image\" content=\"%s\">" % self.image_url
image_url_img = "<img src=\"%s\" />" % self.image_url
else:
image_url_node = image_url_img = ""
# clean html
self.clean_html = self._article_extractor.get_clean_html(source_html=self._source_html)
# summarized description, requires clean_html
if self.clean_html:
self.description, self.language = get_plain_text(etree.XML(self.clean_html),
self._summary_sentences_qty)
description_node = ""
if self.description:
# Replacing \xc2\xa0 and \xa0 in result with space
self.description = self.description.replace(u'\xc2\xa0', u' ').replace(u'\xa0', u' ')
description_node = "<meta name=\"description\" content=\"%s\">" if self.description else ""
# filling the template
self.clean_html = ARTICLE_TEMPLATE % {
'language': self.language,
'title': self.title,
'image_url_node': image_url_node,
'image_url_img': image_url_img,
'description_node': description_node,
'clean_html': self.clean_html
}
class Wanish(object):
def __init__(self,
url=None,
positive_keywords=None,
negative_keywords=None,
summary_sentences_qty=5,
headers=None):
"""
Initialization of the class. If url is set, it gets performed.
:param url: web-page url of the document
:param positive_keywords: list of keywords, which are likely to be seen in classes or ids of tags
:param negative_keywords: list of keywords, which are unlikely to be seen in classes or ids of tags
:param summary_sentences_qty: maximum quantity of summary sentences
:param headers: custom headers for GET request to obtain web page of the article
"""
# TODO: customizable redirects limit?
self._article_extractor = ArticleExtractor(
positive_keywords=positive_keywords,
negative_keywords=negative_keywords)
self.url = None # source web-page url
self.canonical_url = None # canonical web-page url if present, otherwise same as url
self.title = None # document's title
self.image_url = None # document's image url
self.language = None # document's article language
self.clean_html = None # cleaned html of the article
self.description = None # summarized description (text only)
self.error_msg = None # error message
self._source_html = None # source html of the document (lxml doc)
self._charset = None # source html encoding
self._headers = headers if type(headers) == dict else {
} # custom headers for GET request
# summarized text sentences quantity
try:
self._summary_sentences_qty = summary_sentences_qty
except (TypeError, ValueError):
self._summary_sentences_qty = 5
# perform the url if defined
if url:
self.perform_url(url)
def perform_url(self, url):
"""
Perform an article document by designated url
:param url: web-page url of the document
"""
self.url = url
self.title = self.image_url = self.language = self.description = \
self.clean_html = self.error_msg = self._charset = None
if not self.url:
self.error_msg = 'Empty or null URL to perform'
return
# get the page (bytecode)
try:
web_page = requests.get(self.url, headers=self._headers)
# perform http status codes
if web_page.status_code not in [200, 301, 302]:
self.error_msg = str('HTTP error. Status: %s' %
web_page.status_code)
return
self.url = web_page.url
raw_html = web_page.content
# getting content_type from headers to obtain encoding, will use it if it is not specified on page
page_encodings = get_encodings(raw_html)
if len(page_encodings) > 0:
self._charset = page_encodings[0]
elif web_page.encoding is not None:
self._charset = web_page.encoding
else:
res = chardet.detect(raw_html)
self._charset = res['encoding']
string_source = raw_html.decode(self._charset, "ignore")
self._source_html = fromstring(string_source)
# searching for canonical url
link_canonicals = self._source_html.xpath(
"//link[normalize-space(@rel)='canonical']/@href")
self.canonical_url = link_canonicals[0] if len(
link_canonicals) > 0 else self.url
self._source_html = html_cleaner.clean_html(self._source_html)
# making links absolute
self._source_html.make_links_absolute(self.url,
resolve_base_href=True)
strip_elements(self._source_html, 'blockquote', 'code', 'table',
'ol', 'ul', 'embedded', 'input', 'address',
'iframe', 'textarea', 'dl')
except (ConnectionError, Timeout, TypeError, Exception) as e:
self.error_msg = str(e)
finally:
if self.error_msg:
return
if self._source_html is not None:
# clean html of the article and its starting node
self.clean_html, starting_node = self._article_extractor.get_clean_html(
source_html=self._source_html)
# obtaining title
short_title, title_node = shorten_title(self._source_html,
starting_node)
self.title = clean_entities(short_title)
# obtaining image url
self.image_url = get_image_url(self._source_html, self.url,
self._headers, starting_node,
title_node)
if self.image_url is not None:
image_url_node = "<meta itemprop=\"image\" content=\"%s\">" % self.image_url
image_url_img = "<img src=\"%s\" />" % self.image_url
else:
image_url_node = image_url_img = ""
# summarized description, requires clean_html
if self.clean_html:
self.description, self.language = get_plain_text(
etree.XML(self.clean_html), self._summary_sentences_qty)
description_node = ""
if self.description:
# Replacing \xc2\xa0 and \xa0 in result with space
self.description = self.description.replace(
u'\xc2\xa0', u' ').replace(u'\xa0', u' ')
self.description = clean_entities(self.description)
self.description = ' '.join(self.description.split())
description_node = "<meta name=\"description\" content=\"%s\">" if self.description else ""
# filling the template
self.clean_html = ARTICLE_TEMPLATE % {
'language': self.language,
'title': self.title,
'image_url_node': image_url_node,
'image_url_img': image_url_img,
'description_node': description_node,
'clean_html': self.clean_html
}