From c4b36e8ecb70b0912c3172c90ed44fc5cb4debdc Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Tomasz=20D=C5=82ugosz?= Date: Mon, 10 Oct 2016 23:12:40 +0200 Subject: [PATCH 1/3] recipes: fix dwutygodnik --- recipes/dwutygodnik.recipe | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/recipes/dwutygodnik.recipe b/recipes/dwutygodnik.recipe index 2a911e9cc3..ce6ab2bab2 100644 --- a/recipes/dwutygodnik.recipe +++ b/recipes/dwutygodnik.recipe @@ -32,7 +32,7 @@ class dwutygodnik(BasicNewsRecipe): browser.open('http://www.dwutygodnik.com/') # find the link - epublink = browser.find_link(text_regex=re.compile('Wersja ePub')) + epublink = browser.find_link(text_regex=re.compile('Wydanie EPUB')) # download ebook self.report_progress(0, _('Downloading ePUB')) From 53d1ebb3b894896782ae130447b4f5582dcb3543 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Tomasz=20D=C5=82ugosz?= Date: Mon, 10 Oct 2016 23:15:30 +0200 Subject: [PATCH 2/3] recipes: fix biweekly --- recipes/biweekly.recipe | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/recipes/biweekly.recipe b/recipes/biweekly.recipe index 8a9d280635..09b7af5bcd 100644 --- a/recipes/biweekly.recipe +++ b/recipes/biweekly.recipe @@ -30,7 +30,7 @@ class biweekly(BasicNewsRecipe): browser.open('http://www.biweekly.pl/') # find the link - epublink = browser.find_link(text_regex=re.compile('ePUB VERSION')) + epublink = browser.find_link(text_regex=re.compile('EPUB')) # download ebook self.report_progress(0, _('Downloading ePUB')) From c18552a30ef206b4e34cf85eee4d36de2da1c02b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Tomasz=20D=C5=82ugosz?= Date: Mon, 10 Oct 2016 23:58:54 +0200 Subject: [PATCH 3/3] recipes: slightly improve DziennikBaltycki --- recipes/dziennik_baltycki.recipe | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/recipes/dziennik_baltycki.recipe b/recipes/dziennik_baltycki.recipe index 104822e3a2..858921239a 100644 --- a/recipes/dziennik_baltycki.recipe +++ b/recipes/dziennik_baltycki.recipe @@ -7,7 +7,6 @@ class DziennikBaltycki(BasicNewsRecipe): description = u'Gazeta Regionalna Dziennik Bałtycki. Najnowsze Wiadomości Trójmiasto i Wiadomości Pomorskie. Czytaj!' category = 'newspaper' language = 'pl' - encoding = 'iso-8859-2' masthead_url = 'http://s.polskatimes.pl/g/logo_naglowek/dziennikbaltycki.png?24' oldest_article = 7 max_articles_per_feed = 100 @@ -15,8 +14,10 @@ class DziennikBaltycki(BasicNewsRecipe): no_stylesheets = True use_embedded_content = False ignore_duplicate_articles = {'title', 'url'} - remove_tags_after = dict(attrs={'src': 'http://nm.dz.com.pl/dz.png'}) - remove_tags = [dict(id='mat-podobne'), dict(name='a', attrs={ + keep_only_tags = [dict(name='section', attrs={'class': 'zajawka'}), + dict(name='section', attrs={'id': 'tresc'}) +] + remove_tags = [dict(name='div', attrs={'class':['materialyZKategorii','materialyPodobne']}), dict(name='a', attrs={ 'class': 'czytajDalej'}), dict(attrs={'src': 'http://nm.dz.com.pl/dz.png'})] feeds = [