Fix cover fetching for The Atlantic and fix regressions in epub-meta and ebook-viewer

setup.py

@@ -144,6 +144,7 @@ if __name__ == '__main__':
             ncx_template    = 'ebooks/metadata/ncx.xml',
             fb2_xsl         = 'ebooks/lrf/fb2/fb2.xsl',
             metadata_sqlite = 'library/metadata_sqlite.sql',
+            jquery          = 'gui2/viewer/jquery.js',
          )
         
         DEST = os.path.join('src', APPNAME, 'resources.py')

src/calibre/ebooks/metadata/epub.py

@@ -8,7 +8,6 @@ __copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
 import sys, os, time
 from cStringIO import StringIO
 from contextlib import closing
-from ctypes import c_long, byref
 
 from PyQt4.Qt import QUrl, QEventLoop, QSize, QByteArray, QBuffer, \
                      SIGNAL, QPainter, QImage, QObject, QApplication, Qt, QPalette
@@ -19,7 +18,7 @@ from calibre.ebooks.BeautifulSoup import BeautifulStoneSoup
 from calibre.ebooks.metadata import get_parser, MetaInformation
 from calibre.ebooks.metadata.opf2 import OPF
 from calibre.ptempfile import TemporaryDirectory
-from calibre import CurrentDir
+from calibre import CurrentDir, fit_image
 
 class EPubException(Exception):
     pass
@@ -120,11 +119,11 @@ class CoverRenderer(QObject):
     def render_html(self, ok):
         self.rendered = True
         try:
-            from calibre.utils.PythonMagickWand import ImageMagick, NewMagickWand, MagickGetImageBlob, \
-                                           MagickReadImageBlob, MagickTrimImage, MagickSetFormat
             if not ok:
                 return
-            self.page.setViewportSize(QSize(1280, 1024))
+            size = self.page.mainFrame().contentsSize()
+            width, height = fit_image(size.width(), size.height(), 1280, 1024)[1:]
+            self.page.setViewportSize(QSize(width, height))
             image = QImage(self.page.viewportSize(), QImage.Format_ARGB32)
             image.setDotsPerMeterX(96*(100/2.54))
             image.setDotsPerMeterY(96*(100/2.54))
@@ -136,21 +135,7 @@ class CoverRenderer(QObject):
             buf = QBuffer(ba)
             buf.open(QBuffer.WriteOnly)
             image.save(buf, 'JPEG')
-            raw = str(ba.data())
+            self.image_data = str(ba.data())
-            with ImageMagick():
-                wand = NewMagickWand()
-                if not MagickReadImageBlob(wand, raw, len(raw)):
-                    raise ValueError('Failed to load cover image')
-                if not MagickTrimImage(wand, 10.):
-                    raise ValueError('Failed to process cover image')
-                x = c_long(0)
-                if not MagickSetFormat(wand, 'JPEG'):
-                    raise Exception()
-                dat = MagickGetImageBlob(wand, byref(x))
-                obuf = StringIO()
-                for i in xrange(x.value):
-                    obuf.write(chr(dat[i]))
-                self.image_data = obuf.getvalue()
         finally:
             self.loop.exit(0)
         

src/calibre/gui2/viewer/documentview.py

@@ -14,6 +14,7 @@ from PyQt4.QtWebKit import QWebPage, QWebView, QWebSettings
 from calibre.utils.config import Config, StringConfig
 from calibre.gui2.viewer.config_ui import Ui_Dialog
 
+
 def load_builtin_fonts():
     from calibre.ebooks.lrf.fonts.liberation import LiberationMono_BoldItalic
     QFontDatabase.addApplicationFontFromData(QByteArray(LiberationMono_BoldItalic.font_data))
@@ -128,6 +129,14 @@ class Document(QWebPage):
         # Miscellaneous
         settings.setAttribute(QWebSettings.LinksIncludedInFocusChain, True)
         
+        # Load jQuery
+        self.connect(self.mainFrame(), SIGNAL('javaScriptWindowObjectCleared()'), 
+                     self.load_javascript_libraries)
+
+    def load_javascript_libraries(self):
+        from calibre.resources import jquery
+        self.javascript(jquery)
+    
     def javascript(self, string, typ=None):
         ans = self.mainFrame().evaluateJavaScript(string)
         if typ == 'int':
@@ -259,7 +268,7 @@ class DocumentView(QWebView):
         self.connect(self.scrollbar, SIGNAL('valueChanged(int)'), self.scroll_horizontally)
         
     def scroll_horizontally(self, amount):
-        self.document.scroll_to(x=amount)
+        self.document.scroll_to(y=self.document.ypos, x=amount)
         
     def link_hovered(self, link, text, context):
         link, text = unicode(link), unicode(text)
@@ -340,6 +349,7 @@ class DocumentView(QWebView):
             if scrolled:
                 self.manager.scrolled(self.document.scroll_fraction)
         
+        
     @classmethod
     def test_line(cls, img, y):
         start = img.pixel(0, y)

src/calibre/gui2/viewer/main.py

@@ -363,9 +363,15 @@ class EbookViewer(MainWindow, Ui_EbookViewer):
         self.view.load_path(path, pos=pos)
     
     def viewport_resized(self, frac):
-        pos = self.pos.value()
+        new_page = self.pos.value()
-        self.view.scroll_to(pos, notify=False)
+        if self.current_page is not None:
-        #self.set_page_number(frac)
+            try:
+                frac = float(new_page-self.current_page.start_page)/(self.current_page.pages-1)
+            except ZeroDivisionError:
+                frac = 0
+            self.view.scroll_to(frac, notify=False)
+        else:
+            self.set_page_number(frac)
     
     def close_progress_indicator(self):
         self.pi.stop()

src/calibre/web/feeds/news.py

@@ -79,6 +79,9 @@ class BasicNewsRecipe(object, LoggingInterface):
     #: If True stylesheets are not downloaded and processed
     no_stylesheets         = False
     
+    #: Convenient flag to strip all javascripts tags from the downloaded HTML
+    remove_javascript      = True
+    
     #: If True the GUI will ask the user for a username and password 
     #: to use while downloading
     #: @type: boolean
@@ -485,6 +488,9 @@ class BasicNewsRecipe(object, LoggingInterface):
         if self.no_stylesheets:
             for link in list(soup.findAll('link', type=re.compile('css'))):
                 link.extract()
+        if self.remove_javascript:
+            for script in list(soup.findAll('script')):
+                script.extract()
         return self.postprocess_html(soup, first_fetch)
         
     

src/calibre/web/feeds/recipes/atlantic.py

@@ -5,7 +5,7 @@ __copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
 '''
 theatlantic.com
 '''
-
+import re
 from calibre.web.feeds.news import BasicNewsRecipe
 
 class TheAtlantic(BasicNewsRecipe):
@@ -28,7 +28,7 @@ class TheAtlantic(BasicNewsRecipe):
         if issue:
             self.timefmt = ' [%s]'%self.tag_to_string(issue).rpartition('|')[-1].strip().replace('/', '-')
             
-        cover = soup.find('img', alt='feature image', src=True)
+        cover = soup.find('img', alt=re.compile('Cover'), src=True)
         if cover is not None:
             self.cover_url = 'http://theatlantic.com'+cover['src']