Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Update Apple Daily

Browse Source 
Merge branch 'appledaily-patch-20170121' of https://github.com/Kennyl/calibre
This commit is contained in:
Kovid Goyal 2017-01-21 13:17:02 +05:30
commit fc63acac96
1 changed files with 14 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
recipes/apple_daily.recipe
View File

@ -9,6 +9,7 @@ from calibre.utils.date import now as nowf
import os import os
import datetime import datetime
import re import re
import time
from calibre.web.feeds.recipes import BasicNewsRecipe from calibre.web.feeds.recipes import BasicNewsRecipe
from contextlib import nested from contextlib import nested
from calibre.ebooks.BeautifulSoup import BeautifulSoup from calibre.ebooks.BeautifulSoup import BeautifulSoup
@ -33,7 +34,7 @@ class AppleDaily(BasicNewsRecipe):
no_stylesheets = True no_stylesheets = True
description = 'http://hkm.appledaily.com/' description = 'http://hkm.appledaily.com/'
category = 'Chinese, News, Hong Kong' category = 'Chinese, News, Hong Kong'
masthead_url = 'http://upload.wikimedia.org/wikipedia/zh/c/cf/AppleDailyLogo1.png' masthead_url = 'https://upload.wikimedia.org/wikipedia/zh/thumb/5/52/Apple_Daily_logo.svg/569px-Apple_Daily_logo.svg.png'
extra_css = 'img {display: block; margin-left: auto; margin-right: auto; margin-top: 10px; margin-bottom: 10px; max-height:90%;} h1 {font-size:200%; text-align:left; font-weight:bold;} p[class=video-caption] {font-size:50%; margin-left:auto; margin-right:auto;}' # noqa extra_css = 'img {display: block; margin-left: auto; margin-right: auto; margin-top: 10px; margin-bottom: 10px; max-height:90%;} h1 {font-size:200%; text-align:left; font-weight:bold;} p[class=video-caption] {font-size:50%; margin-left:auto; margin-right:auto;}' # noqa
keep_only_tags = [dict(name='div', attrs={'id': 'content-article'})] keep_only_tags = [dict(name='div', attrs={'id': 'content-article'})]
@ -104,9 +105,19 @@ class AppleDaily(BasicNewsRecipe):
relativea = li.find('a', href=True).get('href', False) relativea = li.find('a', href=True).get('href', False)
a = 'http://hkm.appledaily.com/' + relativea a = 'http://hkm.appledaily.com/' + relativea
title = li.find('a', text=True).strip() title = li.find('a', text=True).strip()
if (time.tzname != 'HKT'):
if (title == u'三藩市'):
continue
if (title == u'洛杉磯'):
continue
if (title == u'紐  約'):
continue
if (title == u'美  國'):
continue
if (not title == u'動新聞') and (relativea.startswith('list.php')): if (not title == u'動新聞') and (relativea.startswith('list.php')):
sectionList.append((title, a)) sectionList.append((title, a))
for title, url in sectionList: for title, url in sectionList:
title = title.replace(" ", "")
articles = self.parse_section(url) articles = self.parse_section(url)
if articles: if articles:
feeds.append((title, articles)) feeds.append((title, articles))
@ -116,6 +127,8 @@ class AppleDaily(BasicNewsRecipe):
soup = self.index_to_soup(url) soup = self.index_to_soup(url)
ul = soup.find(attrs={'class': 'list'}) ul = soup.find(attrs={'class': 'list'})
current_articles = [] current_articles = []
if ul is None :
return current_articles
for li in ul.findAll('li'): for li in ul.findAll('li'):
a = li.find('a', href=True) a = li.find('a', href=True)
title = li.find('p', text=True).strip() title = li.find('p', text=True).strip()