def _decorate_article(self, article):
"""在 parse_response 後執行，後處理其輸出"""
# html post-process
from lxml.html import tostring, fromstring
from bs4 import BeautifulSoup
from lib.util.net import normalize_url
from lib.util.text import pack_string
# article['content'] may be list of lxml doms
if type(article['content']) is list:
article['content'] = \
fromstring('\n'.join([tostring(x, encoding=unicode) for x in article['content']]))
# remove unwanted tags
self.css_sel_drop_tree(article['content'], ['script'])
# prettify html with BeautifulSoup
html_bs4 = BeautifulSoup(tostring(article['content'],
encoding=unicode)).body.next
article['text'] = pack_string(html_bs4.text)
article['html'] = pack_string(unicode(html_bs4))
article["ctlr_classname"] = str(self.__class__)
article['url'] = normalize_url(article['url'])
article['url_read'] = normalize_url(article['url_read'])
article['url_canonical'] = normalize_url(article['url_canonical'])
self.move_out_of_meta(article, 'title')
return article
def _decorate_article(self, article):
"""在 parse_response 後執行，後處理其輸出"""
# html post-process
from lxml.html import tostring, fromstring
from bs4 import BeautifulSoup
from lib.util.net import normalize_url
from lib.util.text import pack_string
# article['content'] may be list of lxml doms
if type(article['content']) is list:
article['content'] = \
fromstring('\n'.join([tostring(x, encoding=unicode) for x in article['content']]))
# remove unwanted tags
self.css_sel_drop_tree(article['content'], ['script'])
# prettify html with BeautifulSoup
html_bs4 = BeautifulSoup(tostring(article['content'], encoding=unicode)).body.next
article['text'] = pack_string(html_bs4.text)
article['html'] = pack_string(unicode(html_bs4))
article["ctlr_classname"] = str(self.__class__)
article['url'] = normalize_url(article['url'])
article['url_read'] = normalize_url(article['url_read'])
article['url_canonical'] = normalize_url(article['url_canonical'])
self.move_out_of_meta(article, 'title')
return article
def get_fresh_urls(urls, dbi = None):
"""篩選出 urls 中未曾成功抓取過者並回傳"""
from lib.util.text import md5
from lib.util.net import normalize_url
if (len(urls) == 0): return set()
url_md5 = [{'url': x, 'md5': md5(normalize_url(x))} for x in urls]
hashes = "(" + (",".join(["UNHEX('%s')" % x['md5'] for x in url_md5 ])) + ")"
sql = "SELECT HEX(`hash`) FROM `article__urls` WHERE `hash` IN %s" % hashes
ret = set(DB.query(sql, dbi = dbi))
output = []
for x in url_md5:
if not (x['md5'].upper(),) in ret:
output.append(x['url'])
return output
