From e311bc02a842f1ee4914a662630adeb39bf047a7 Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Tue, 23 Jun 2009 11:03:55 -0700
Subject: [PATCH] Implement #2690 (Various Belgian recipes)

---
 src/calibre/gui2/images/news/gva_be.png       | Bin 0 -> 621 bytes
 src/calibre/gui2/images/news/hln.png          | Bin 0 -> 295 bytes
 src/calibre/gui2/images/news/tijd.png         | Bin 0 -> 381 bytes
 src/calibre/web/feeds/recipes/__init__.py     |   1 +
 .../web/feeds/recipes/recipe_degentenaar.py   |  75 ++++++++++++++++++
 .../web/feeds/recipes/recipe_gva_be.py        |  63 +++++++++++++++
 src/calibre/web/feeds/recipes/recipe_hln.py   |  52 ++++++++++++
 src/calibre/web/feeds/recipes/recipe_tijd.py  |  70 ++++++++++++++++
 8 files changed, 261 insertions(+)
 create mode 100644 src/calibre/gui2/images/news/gva_be.png
 create mode 100644 src/calibre/gui2/images/news/hln.png
 create mode 100644 src/calibre/gui2/images/news/tijd.png
 create mode 100644 src/calibre/web/feeds/recipes/recipe_degentenaar.py
 create mode 100644 src/calibre/web/feeds/recipes/recipe_gva_be.py
 create mode 100644 src/calibre/web/feeds/recipes/recipe_hln.py
 create mode 100644 src/calibre/web/feeds/recipes/recipe_tijd.py

diff --git a/src/calibre/gui2/images/news/gva_be.png b/src/calibre/gui2/images/news/gva_be.png
new file mode 100644
index 0000000000000000000000000000000000000000..ca753a5bbda65ec3b65a210ac62b352f56ee02d6
GIT binary patch
literal 621
zcmeAS@N?(olHy`uVBq!ia0vp^0wB!61|;P_|4#%`Y)RhkE)4%caKYZ?lYt_f1s;*b
zK-vS0-A-oPfdtD69Mgd`SU*F|v9*U87#OE{x;TbdoPK+0ueWre%<+%k-<R(^60^vi
z*Ed8!L{zl<#e)DAO(njivz-rKbS)5czy6Q&oF1Q_=#m!~{FXKgiMrM}D5|R2Et|!z
z_w3y1rw?T=B|W#=@#Q!Bo*$o{F`u_#e6VZX@AVBe@7VSED;QQCnEqJC-uPR5A<L|A
zgOAH5dIYd0C>NgJKEpBI<#BTTW8n=IFE(td^>#nuenESp(}eBXbMD=IaOli!j<TJn
zF2&7qIhXkT)S~?-{qt|cGuO=#mB?J5{x8<JON2q9to7zrmFnwa498cp>+!Z<^qVW}
z{V-bxq%u2T-L5TrKP~cb)X6p8Dx^MdtK|Djp3k~umg;QN_Z8_sl+u6f!uRuUj$E4Z
z&*e*{iu!gl%hf)w_PN+Zt7L~tyv|X05FByuyue-!gBwB%Lc?FK)=r90-Db4Ek^in^
z*?OO4;lesO|D3oQgI5ait>Q7sbUE+4j<<c|H-<&$&VN{L>9oEyjQ`AX)jT!sIjP+o
zjH{F+-exE)soZhWdd++(lk2PZyuUbc(KU^#IEBnp|AX1&7@IfDd#^vAwF($rswJ)w
zCCNppi6yDJDL{h3$iTo-*T78I$T-B%z{=Fj%EVCDz}(8f;7Gf|LYRi!{FKbJN;D0I
fAw~vPhQ?NgMi34D{J(^N8W=oX{an^LB{Ts5KJWJP

literal 0
HcmV?d00001

diff --git a/src/calibre/gui2/images/news/hln.png b/src/calibre/gui2/images/news/hln.png
new file mode 100644
index 0000000000000000000000000000000000000000..d7824d4ccc1d339ded30047d10438f11c4ebe109
GIT binary patch
literal 295
zcmeAS@N?(olHy`uVBq!ia0vp^0wB!61|;P_|4#%`Y)RhkE)4%caKYZ?lYt_f1s;*b
zK-vS0-A-oPfdtD69Mgd`SU*F|v9*VRoB&T3#}JFt$q5pl8YU(_oB5vy1YWA`xCx|o
z{P3BwtD%HBK~M2E!yG*aV@Y)fK6w%5ghuDj1`eBh_%=Ey<Y<%~JivU+n1|=cQF}fH
z2H7vtEB>%s&jXsRTH+c}l3bLUSdyxn0wfrW3=AxF4a{_nj6)0!tW3?UObm4m%&iOz
tj<hQ*glWjlPsvQHMAKjxVq{=t0#s@S(ZC;eX*o~>gQu&X%Q~loCII{bQf2@E

literal 0
HcmV?d00001

diff --git a/src/calibre/gui2/images/news/tijd.png b/src/calibre/gui2/images/news/tijd.png
new file mode 100644
index 0000000000000000000000000000000000000000..fd55a4e5adc5e4e6b3777d7e209f5ea408ca7793
GIT binary patch
literal 381
zcmeAS@N?(olHy`uVBq!ia0vp^0wB!61|;P_|4#%`Y)RhkE)4%caKYZ?lYt_f1s;*b
zK-vS0-A-oPfdtD69Mgd`SU*F|v9*VRoMoOajv*GOlT*UJ{y*=)n6Svjm>CGTdlo9`
zA1z4EwG5nP(qt|9&ZegA6_a(bE*lWE&FQ<!qoei{ggK%Y`~h+`vI2Pw$_r9g#Fb7@
zJT^<&B6r@#XB^H4Up?nhdzLmM&SHPQaokx0gQc6B6_t;$ol8F=&Ru`br=^-ve1V0S
z8}EWvMj@p?JkjFY4lSP0$S5=y$TITi_>e4dz~aG9L5Asdx)=UwHLU@<Shd78q9nN}
zHL)aBHw8#A7#SE?>Kd5o8X1Qe8d#Z{S(zB>8kk!d7#wL=SP0XQo1c=IR*9y;FvQ5f
b%E;Wx)B>X6zfJC4paup{S3j3^P6<r_O9FXk

literal 0
HcmV?d00001

diff --git a/src/calibre/web/feeds/recipes/__init__.py b/src/calibre/web/feeds/recipes/__init__.py
index 2fd1c74ba8..af776493f8 100644
--- a/src/calibre/web/feeds/recipes/__init__.py
+++ b/src/calibre/web/feeds/recipes/__init__.py
@@ -50,6 +50,7 @@ recipe_modules = ['recipe_' + r for r in (
            'marca', 'kellog_faculty', 'kellog_insight',
            'theeconomictimes_india', '7dias', 'buenosaireseconomico',
            'diagonales', 'miradasalsur', 'newsweek_argentina', 'veintitres',
+           'gva_be', 'hln', 'tijd', 'degentenaar',
           )]
 
 import re, imp, inspect, time, os
diff --git a/src/calibre/web/feeds/recipes/recipe_degentenaar.py b/src/calibre/web/feeds/recipes/recipe_degentenaar.py
new file mode 100644
index 0000000000..03b324cc2d
--- /dev/null
+++ b/src/calibre/web/feeds/recipes/recipe_degentenaar.py
@@ -0,0 +1,75 @@
+#!/usr/bin/env  python
+
+__license__   = 'GPL v3'
+__copyright__ = '2009, Darko Miletic <darko.miletic at gmail.com>'
+'''
+www.nieuwsblad.be
+'''
+from calibre.web.feeds.news import BasicNewsRecipe
+from calibre.ebooks.BeautifulSoup import Tag
+
+class DeGentenaarOnline(BasicNewsRecipe):
+    title                 = 'De Gentenaar Online'
+    __author__            = 'Darko Miletic'
+    description           = 'News from Belgium in Dutch'
+    publisher             = 'De Gentenaar'
+    category              = 'news, politics, Belgium'
+    oldest_article        = 2
+    max_articles_per_feed = 100
+    no_stylesheets        = True
+    use_embedded_content  = False
+    encoding              = 'utf-8'
+    language              = _('Dutch')
+    lang                  = 'nl-BE'
+    direction             = 'ltr'
+
+    html2lrf_options = [
+                          '--comment'  , description
+                        , '--category' , category
+                        , '--publisher', publisher
+                        ]
+
+    html2epub_options = 'publisher="' + publisher + '"\ncomments="' + description + '"\ntags="' + category + '"\noverride_css=" p {text-indent: 0cm; margin-top: 0em; margin-bottom: 0.5em} "'
+
+    keep_only_tags = [dict(name='span', attrs={'id':['lblArticleTitle','lblArticleIntroduction','lblArticleMainText']})]
+    remove_tags    = [dict(name=['embed','object'])]
+
+
+
+    feeds = [
+              (u'Snelnieuws' , u'http://feeds.nieuwsblad.be/nieuws/snelnieuws'     )
+             ,(u'Binnenland' , u'http://feeds.nieuwsblad.be/nieuws/binnenland'     )
+             ,(u'Buitenland' , u'http://feeds.nieuwsblad.be/nieuwsblad/buitenland' )
+             ,(u'Economie'   , u'http://feeds.nieuwsblad.be/economie/home'         )
+             ,(u'Economie'   , u'http://feeds.nieuwsblad.be/economie/home'         )
+             ,(u'Algemeen'   , u'http://feeds.nieuwsblad.be/life/algemeen'         )
+             ,(u'Film'       , u'http://feeds.nieuwsblad.be/life/film'             )
+             ,(u'Boek'       , u'http://feeds.nieuwsblad.be/life/boeken'           )
+             ,(u'Muziek'     , u'http://feeds.nieuwsblad.be/life/muziek'           )
+             ,(u'Podium'     , u'http://feeds.nieuwsblad.be/life/podium'           )
+             ,(u'TV & radio' , u'http://feeds.nieuwsblad.be/life/tv'               )
+            ]
+
+    def print_version(self, url):
+        return url.replace('/Detail.aspx?articleid','/PrintArticle.aspx?ArticleID')
+
+    def get_article_url(self, article):
+        return article.get('guid',  None)
+
+    def preprocess_html(self, soup):
+        del soup.body['onload']
+        for item in soup.findAll(style=True):
+            del item['style']
+        for item in soup.findAll('span'):
+            item.name='div'
+            if item.has_key('id') and item['id'] == 'lblArticleTitle':
+               item.name='h3'
+
+        soup.html['lang']     = self.lang
+        soup.html['dir' ]     = self.direction
+        mlang = Tag(soup,'meta',[("http-equiv","Content-Language"),("content",self.lang)])
+        mcharset = Tag(soup,'meta',[("http-equiv","Content-Type"),("content","text/html; charset=utf-8")])
+        soup.head.insert(0,mlang)
+        soup.head.insert(1,mcharset)
+        return soup
+
diff --git a/src/calibre/web/feeds/recipes/recipe_gva_be.py b/src/calibre/web/feeds/recipes/recipe_gva_be.py
new file mode 100644
index 0000000000..a02807d8d2
--- /dev/null
+++ b/src/calibre/web/feeds/recipes/recipe_gva_be.py
@@ -0,0 +1,63 @@
+#!/usr/bin/env  python
+
+__license__   = 'GPL v3'
+__copyright__ = '2009, Darko Miletic <darko.miletic at gmail.com>'
+'''
+www.gva.be
+'''
+from calibre.web.feeds.news import BasicNewsRecipe
+from calibre.ebooks.BeautifulSoup import Tag
+
+class GazetvanAntwerpen(BasicNewsRecipe):
+    title                 = 'Gazet van Antwerpen'
+    __author__            = 'Darko Miletic'
+    description           = 'News from Belgium in Dutch'
+    publisher             = 'Gazet van Antwerpen'
+    category              = 'news, politics, Belgium'
+    oldest_article        = 2
+    max_articles_per_feed = 100
+    no_stylesheets        = True
+    use_embedded_content  = False
+    encoding              = 'utf-8'
+    language              = _('Dutch')
+    lang                  = 'nl-BE'
+    direction             = 'ltr'
+
+    html2lrf_options = [
+                          '--comment'  , description
+                        , '--category' , category
+                        , '--publisher', publisher
+                        ]
+
+    html2epub_options = 'publisher="' + publisher + '"\ncomments="' + description + '"\ntags="' + category + '"\noverride_css=" p {text-indent: 0cm; margin-top: 0em; margin-bottom: 0.5em} "'
+
+    keep_only_tags = [dict(name='div', attrs={'id':'article'})]
+    remove_tags    = [
+                         dict(name=['embed','object'])
+                       , dict (name='div',attrs={'class':['note NotePortrait','note']})
+                     ]
+    remove_tags_after  = dict(name='span', attrs={'class':'author'})
+
+    feeds = [
+              (u'Overzicht & Blikvanger', u'http://www.gva.be/syndicationservices/artfeedservice.svc/rss/overview/overzicht'       )
+             ,(u'Stad & Regio'          , u'http://www.gva.be/syndicationservices/artfeedservice.svc/rss/mostrecent/stadenregio'   )
+             ,(u'Economie'              , u'http://www.gva.be/syndicationservices/artfeedservice.svc/rss/mostrecent/economie'      )
+             ,(u'Binnenland'            , u'http://www.gva.be/syndicationservices/artfeedservice.svc/rss/mostrecent/binnenland'    )
+             ,(u'Buitenland'            , u'http://www.gva.be/syndicationservices/artfeedservice.svc/rss/mostrecent/buitenland'    )
+             ,(u'Media & Cultur'        , u'http://www.gva.be/syndicationservices/artfeedservice.svc/rss/mostrecent/mediaencultuur')
+             ,(u'Wetenschap'            , u'http://www.gva.be/syndicationservices/artfeedservice.svc/rss/mostrecent/mediaencultuur')
+             ,(u'Sport'                 , u'http://www.gva.be/syndicationservices/artfeedservice.svc/rss/mostrecent/sport'         )
+            ]
+
+    def preprocess_html(self, soup):
+        del soup.body['onload']
+        for item in soup.findAll(style=True):
+            del item['style']
+        soup.html['lang']     = self.lang
+        soup.html['dir' ]     = self.direction
+        mlang = Tag(soup,'meta',[("http-equiv","Content-Language"),("content",self.lang)])
+        mcharset = Tag(soup,'meta',[("http-equiv","Content-Type"),("content","text/html; charset=utf-8")])
+        soup.head.insert(0,mlang)
+        soup.head.insert(1,mcharset)
+        return soup
+
diff --git a/src/calibre/web/feeds/recipes/recipe_hln.py b/src/calibre/web/feeds/recipes/recipe_hln.py
new file mode 100644
index 0000000000..c2b6b2375a
--- /dev/null
+++ b/src/calibre/web/feeds/recipes/recipe_hln.py
@@ -0,0 +1,52 @@
+#!/usr/bin/env  python
+
+__license__   = 'GPL v3'
+__copyright__ = '2009, Darko Miletic <darko.miletic at gmail.com>'
+'''
+www.hln.be
+'''
+from calibre.web.feeds.news import BasicNewsRecipe
+from calibre.ebooks.BeautifulSoup import Tag
+
+class HLN_be(BasicNewsRecipe):
+    title                 = 'Het Belang Van Limburg'
+    __author__            = 'Darko Miletic'
+    description           = 'News from Belgium in Dutch'
+    publisher             = 'Het Belang Van Limburg'
+    category              = 'news, politics, Belgium'
+    oldest_article        = 2
+    max_articles_per_feed = 100
+    no_stylesheets        = True
+    use_embedded_content  = False
+    encoding              = 'utf-8'
+    language              = _('Dutch')
+    lang                  = 'nl-BE'
+    direction             = 'ltr'
+
+    html2lrf_options = [
+                          '--comment'  , description
+                        , '--category' , category
+                        , '--publisher', publisher
+                        ]
+
+    html2epub_options = 'publisher="' + publisher + '"\ncomments="' + description + '"\ntags="' + category + '"\noverride_css=" p {text-indent: 0cm; margin-top: 0em; margin-bottom: 0.5em} "'
+
+    keep_only_tags = [dict(name='div', attrs={'class':'art_box2'})]
+    remove_tags    = [
+                         dict(name=['embed','object'])
+                     ]
+
+    feeds = [(u'Alle nieuws', u'http://www.hln.be/rss.xml')]
+
+    def preprocess_html(self, soup):
+        del soup.body['onload']
+        for item in soup.findAll(style=True):
+            del item['style']
+        soup.html['lang']     = self.lang
+        soup.html['dir' ]     = self.direction
+        mlang = Tag(soup,'meta',[("http-equiv","Content-Language"),("content",self.lang)])
+        mcharset = Tag(soup,'meta',[("http-equiv","Content-Type"),("content","text/html; charset=utf-8")])
+        soup.head.insert(0,mlang)
+        soup.head.insert(1,mcharset)
+        return soup
+
diff --git a/src/calibre/web/feeds/recipes/recipe_tijd.py b/src/calibre/web/feeds/recipes/recipe_tijd.py
new file mode 100644
index 0000000000..7e8fe62449
--- /dev/null
+++ b/src/calibre/web/feeds/recipes/recipe_tijd.py
@@ -0,0 +1,70 @@
+#!/usr/bin/env  python
+
+__license__   = 'GPL v3'
+__copyright__ = '2009, Darko Miletic <darko.miletic at gmail.com>'
+'''
+www.tijd.be
+'''
+from calibre.web.feeds.news import BasicNewsRecipe
+from calibre.ebooks.BeautifulSoup import Tag
+
+class DeTijd(BasicNewsRecipe):
+    title                 = 'De Tijd'
+    __author__            = 'Darko Miletic'
+    description           = 'News from Belgium in Dutch'
+    publisher             = 'De Tijd'
+    category              = 'news, politics, Belgium'
+    oldest_article        = 2
+    max_articles_per_feed = 100
+    no_stylesheets        = True
+    use_embedded_content  = False
+    encoding              = 'utf-8'
+    language              = _('Dutch')
+    lang                  = 'nl-BE'
+    direction             = 'ltr'
+
+    html2lrf_options = [
+                          '--comment'  , description
+                        , '--category' , category
+                        , '--publisher', publisher
+                        ]
+
+    html2epub_options = 'publisher="' + publisher + '"\ncomments="' + description + '"\ntags="' + category + '"\noverride_css=" p {text-indent: 0cm; margin-top: 0em; margin-bottom: 0.5em} "'
+
+    keep_only_tags = [dict(name='div', attrs={'id':'lcol'})]
+    remove_tags    = [
+                         dict(name=['embed','object'])
+                       , dict (name='div',attrs={'id':'art_reactwrap'})
+                     ]
+    remove_tags_after  = dict(name='div', attrs={'id':'art_author'})
+
+    feeds = [
+              (u'Volledig nieuwsaanbod', u'http://www.tijd.be/rss/nieuws.xml'        )
+             ,(u'Markten'              , u'http://www.tijd.be/rss/markten.xml'       )
+             ,(u'Ondernemingen'        , u'http://www.tijd.be/rss/ondernemingen.xml' )
+             ,(u'Chemie-Farma'         , u'http://www.tijd.be/rss/chemie_farma.xml'  )
+             ,(u'Consumptie'           , u'http://www.tijd.be/rss/consumptie.xml'    )
+             ,(u'Diensten'             , u'http://www.tijd.be/rss/diensten.xml'      )
+             ,(u'Energie'              , u'http://www.tijd.be/rss/energie.xml'       )
+             ,(u'Financen'             , u'http://www.tijd.be/rss/financien.xml'     )
+             ,(u'Industrie'            , u'http://www.tijd.be/rss/industrie.xml'     )
+             ,(u'Media'                , u'http://www.tijd.be/rss/media_telecom.xml' )
+             ,(u'Technologie'          , u'http://www.tijd.be/rss/technologie.xml'   )
+             ,(u'Economie & Financien' , u'http://www.tijd.be/rss/economie.xml'      )
+             ,(u'Binnenland'           , u'http://www.tijd.be/rss/binnenland.xml'    )
+             ,(u'Buitenland'           , u'http://www.tijd.be/rss/buitenland.xml'    )
+             ,(u'De wijde wereld'      , u'http://www.tijd.be/rss/cultuur.xml'       )
+            ]
+
+    def preprocess_html(self, soup):
+        del soup.body['onload']
+        for item in soup.findAll(style=True):
+            del item['style']
+        soup.html['lang']     = self.lang
+        soup.html['dir' ]     = self.direction
+        mlang = Tag(soup,'meta',[("http-equiv","Content-Language"),("content",self.lang)])
+        mcharset = Tag(soup,'meta',[("http-equiv","Content-Type"),("content","text/html; charset=utf-8")])
+        soup.head.insert(0,mlang)
+        soup.head.insert(1,mcharset)
+        return soup
+