From 81a501604128274d364c651ec7b37c3499fbd2d2 Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Sun, 8 May 2011 09:36:46 -0600
Subject: [PATCH] Updated recipes for Fronda and Rzeczpospolita

---
 recipes/fronda.recipe            |  17 +++++++++++------
 recipes/icons/rzeczpospolita.png | Bin 0 -> 1195 bytes
 recipes/rzeczpospolita.recipe    |  11 ++++++++++-
 3 files changed, 21 insertions(+), 7 deletions(-)
 create mode 100644 recipes/icons/rzeczpospolita.png

diff --git a/recipes/fronda.recipe b/recipes/fronda.recipe
index f57c7ecccb..452dca9068 100644
--- a/recipes/fronda.recipe
+++ b/recipes/fronda.recipe
@@ -21,14 +21,19 @@ class Fronda(BasicNewsRecipe):
 
     feeds          = [(u'Infformacje', u'http://fronda.pl/news/feed')]
 
-    keep_only_tags = [dict(name='h1', attrs={'class':'big'}),
-                      dict(name='ul', attrs={'class':'about clear'}),
-                      dict(name='div', attrs={'class':'content'})]
+    keep_only_tags = [dict(name='h2', attrs={'class':'news_title'}),
+                      dict(name='div', attrs={'class':'naglowek_tresc'}),
+                      dict(name='div', attrs={'id':'czytaj'}) ]
+
+    remove_tags = [dict(name='a', attrs={'class':'print'})]
+
     preprocess_regexps = [
         (re.compile(i[0], re.IGNORECASE | re.DOTALL), i[1]) for i in 
-            [ (r'<a href="#" class="print">Drukuj</a>', lambda match: ''),
-              (r'<p><a href="http://fronda.pl/sklepy">.*</a></p>', lambda match: ''),
+            [ (r'<p><a href="http://fronda.pl/sklepy">.*</a></p>', lambda match: ''),
               (r'<p><a href="http://fronda.pl/pasaz">.*</a></p>', lambda match: ''),
               (r'<h3><strong>W.* lektury.*</a></p></div>', lambda match: '</div>'),
-              (r'<h3>Zobacz t.*?</div>', lambda match: '</div>') ]
+              (r'<h3>Zobacz t.*?</div>', lambda match: '</div>'),
+              (r'<p[^>]*>&nbsp;</p>', lambda match: ''),
+              (r'<p><span style=".*?"><br /></span></p> ', lambda match: ''),
+              (r'<a style=\'float:right;margin-top:3px;\' href="http://www.facebook.com/share.php?.*?</a>', lambda match: '')]
     ]
diff --git a/recipes/icons/rzeczpospolita.png b/recipes/icons/rzeczpospolita.png
new file mode 100644
index 0000000000000000000000000000000000000000..3782c705b5567b888e60dd85c95d1da3e5ae71fe
GIT binary patch
literal 1195
zcmeAS@N?(olHy`uVBq!ia0vp^0w65F1SAhIZYc#)Y)RhkF8{%RGvfOlpa^GyM`SUO
z_5fqIli79*49qV*T^vI!PA?7d_YwA&s5QQuxA4NIFohtY*YZl82Pe-G_%FCuq9^@n
zLXRWATW{yXmZ}dIDhwrj4TPB4gDg4>k9c1cm|74Jn9SI`EzEcAg($8q0k__rd-vuK
zpY*5IhE*5JXQbTwzSrFNq<MeFts_UCELqliI5@b4A<z7uq!WX-_S2`;-()Ntf>tWK
zpPLgI!NFEtEz7rHPtDq=`HmgS7>_QxUj4Hue0s`ubCaA~PYM`PQ#a|H_Vr~*m{q?|
z-r>aiRV@s={yl6~F*V7#=c8F_HP=s(A@pkLF}qr!CPQCkmJK%<6k8&UQ_mQ*Tqr$u
zZN}DBTcfVcDdrR5a*}9U7ko}!Y}z!NdEDZE>bCzr()q3IGK-O<Z*or6Eg=D>fK!$e
z&dySsaGu#|hMHo7&g^CD)~_>K_S)FEjZsC&@A<j8zrTA2)z;?frs@cFy7?C8<z=qi
zWhJnq^73omRyq5!-TkuGrCo*!QlX)-^XJ9w|94Gy|NeD(IdA^ep5sl<^t@d1_pY^K
z%FJR{%j$2JRM&rhdwZgwt7ZAR#fJ^{)pj4BczWyh_`S1yqWJyPHb(|F#pD$fR8*8*
zdwOd}U`Ikh$(2XKou^JMTPAP+?X7iH$o1rn6?;Oq?t1lk>y|ZZO3KPgAO8N{*Z2Cm
zyE|h-+OnNH*ZTfxs`ztj>%4gn6Bz=oM&y*0<(+%KzkGc_$8m9SZ|mTd2M<o1I8l(n
zV|jJ8yv>0lM~-&Sof{j=Q1G0O;mG6KT7LOyr#9Uzv6{@;e*9S6J`>+=&!t9^p}JzB
zS6@A}nXN0v&c@Jx{5^X^#+Lo{86BsL99Ax3XgJQCbK}F8moG2p<~}SaFOQc`?(V*s
z6B=5&b&-kLwk;;+(T^iTL!(z~@VQ6MJiKt#=FOi!`_GfHxpC{1#DtSqt}-^L6zZ|)
z&dv2)v}Mtur$sw!_PHoY)clL-@Nv6bH}~86`1ki@rDn~ts{eOts(b&oyt`Yqr*gF(
zJ1*LK>FsUrdO_W(K>f+7seH$GmFAV_=K7v}e!kH`VZrOwtAEy5+1mYkb@h0^d0t6*
zy3xW5Pg1P@zPUNueATH4t*2$ox3_UJzNlKcJY?6zRHvAAHFb%Jb#+y{_W%EudyY}@
z(8GE2_SwyU|KZ`au-UvDZU(KqzCM1>g|APaCS80xZHaEQ#q&D`8mt<B-!d-aEXdm&
z_^LpI>%t9>4uKXCuFs!OrkF@^@bFx@^5(|I>hgsGohp;t7&B(w_49CLUD^_H&Az>3
z%IVslLfyK0>tqEx|5!Fpn|^;kKik`H|JSZru;G<h@=nq2P8TE1pEc_Ka`t|iwR?|0
zE;&_{`t*KkhUTZrV?{e>&2n!4_U<n~w?xV1MP7+%z}&A|;u=wsl30>zm0Xkxq!^40
z3=MS+Omz(`LJTaej7_afEOiactPBjS$}0DxXvob^$xN%nt-)LSATv+{gQu&X%Q~lo
FCIG#yBMJZj

literal 0
HcmV?d00001

diff --git a/recipes/rzeczpospolita.recipe b/recipes/rzeczpospolita.recipe
index 836ee95027..05c7361bbb 100644
--- a/recipes/rzeczpospolita.recipe
+++ b/recipes/rzeczpospolita.recipe
@@ -2,7 +2,7 @@ from calibre.web.feeds.news import BasicNewsRecipe
 
 class RzeczpospolitaRecipe(BasicNewsRecipe):
     __license__  = 'GPL v3'
-    __author__ = 'kwetal'
+    __author__ = u'kwetal and Tomasz Dlugosz'
     language = 'pl'
     version = 1
 
@@ -38,6 +38,8 @@ class RzeczpospolitaRecipe(BasicNewsRecipe):
     remove_tags.append(dict(name = 'div', attrs = {'class' : 'clr'}))
     remove_tags.append(dict(name = 'div', attrs = {'id' : 'share_bottom'}))
     remove_tags.append(dict(name = 'div', attrs = {'id' : 'copyright_law'}))
+    remove_tags.append(dict(name = 'div', attrs = {'class' : 'more'}))
+    remove_tags.append(dict(name = 'div', attrs = {'class' : 'editorPicks'}))
 
     extra_css = '''
                     body {font-family: verdana, arial, helvetica, geneva, sans-serif ;}
@@ -48,6 +50,13 @@ class RzeczpospolitaRecipe(BasicNewsRecipe):
                     .fot{font-size: x-small; color: #666666;}
                     '''
 
+    def skip_ad_pages(self, soup):
+        if ('advertisement' in soup.find('title').string.lower()):
+            href = soup.find('a').get('href')
+            return self.index_to_soup(href, raw=True)
+        else:
+            return None
+
     def print_version(self, url):
         start, sep, rest = url.rpartition('/')
         forget, sep, index = rest.rpartition(',')