From acac5a479b925bc8b981b42e395b8fa9279517b0 Mon Sep 17 00:00:00 2001 From: John Schember Date: Tue, 30 Nov 2010 18:27:09 -0500 Subject: [PATCH 1/3] Implement bug #7738: FB2 Output option to create section per HTML file. Implement FB2 options to wrap h1-3 tags with title elments. --- src/calibre/ebooks/fb2/fb2ml.py | 14 +++++++++++- src/calibre/ebooks/fb2/output.py | 14 ++++++++++++ src/calibre/gui2/convert/fb2_output.py | 4 +++- src/calibre/gui2/convert/fb2_output.ui | 30 +++++++++++++++++++++++++- 4 files changed, 59 insertions(+), 3 deletions(-) diff --git a/src/calibre/ebooks/fb2/fb2ml.py b/src/calibre/ebooks/fb2/fb2ml.py index 3987ffa1b8..2a9a92612e 100644 --- a/src/calibre/ebooks/fb2/fb2ml.py +++ b/src/calibre/ebooks/fb2/fb2ml.py @@ -91,6 +91,10 @@ class FB2MLizer(object): return u'\n%s' % etree.tostring(etree.fromstring(output), encoding=unicode, pretty_print=True) def clean_text(self, text): + text = re.sub(r'(?miu)
\s*
', '', text) + text = re.sub(r'(?miu)\s+', '', text) + text = re.sub(r'(?miu)
', '
\n\n
', text) + text = re.sub(r'(?miu)

\s*

', '', text) text = re.sub(r'(?miu)\s+

', '

', text) text = re.sub(r'(?miu)

', '

\n\n

', text) @@ -166,11 +170,15 @@ class FB2MLizer(object): def get_text(self): text = [] - for item in self.oeb_book.spine: + for i, item in enumerate(self.oeb_book.spine): + if self.opts.sectionize_chapters_using_file_structure and i is not 0: + text.append('

') self.log.debug('Converting %s to FictionBook2 XML' % item.href) stylizer = Stylizer(item.data, item.href, self.oeb_book, self.opts, self.opts.output_profile) text.append(self.add_page_anchor(item)) text += self.dump_text(item.data.find(XHTML('body')), stylizer, item) + if self.opts.sectionize_chapters_using_file_structure and i is not len(self.oeb_book.spine) - 1: + text.append('
') return ''.join(text) def fb2_body_footer(self): @@ -258,6 +266,10 @@ class FB2MLizer(object): if id_name: fb2_text.append(self.get_anchor(page, id_name)) + if tag == 'h1' and self.opts.h1_to_title or tag == 'h2' and self.opts.h2_to_title or tag == 'h3' and self.opts.h3_to_title: + fb2_text.append('') + tags.append('title') + fb2_tag = TAG_MAP.get(tag, None) if fb2_tag == 'p': if 'p' in tag_stack+tags: diff --git a/src/calibre/ebooks/fb2/output.py b/src/calibre/ebooks/fb2/output.py index d6c7a25a90..bacaf0da91 100644 --- a/src/calibre/ebooks/fb2/output.py +++ b/src/calibre/ebooks/fb2/output.py @@ -25,6 +25,20 @@ class FB2Output(OutputFormatPlugin): 'WARNING: ' \ 'This option is experimental. It can cause conversion ' \ 'to fail. It can also produce unexpected output.')), + OptionRecommendation(name='sectionize_chapters_using_file_structure', + recommended_value=False, level=OptionRecommendation.LOW, + help=_('Try to turn chapters into individual sections using the ' \ + 'internal structure of the ebook. This works well for EPUB ' \ + 'books that have been internally split by chapter.')), + OptionRecommendation(name='h1_to_title', + recommended_value=False, level=OptionRecommendation.LOW, + help=_('Wrap all h1 tags with fb2 title elements.')), + OptionRecommendation(name='h2_to_title', + recommended_value=False, level=OptionRecommendation.LOW, + help=_('Wrap all h2 tags with fb2 title elements.')), + OptionRecommendation(name='h3_to_title', + recommended_value=False, level=OptionRecommendation.LOW, + help=_('Wrap all h3 tags with fb2 title elements.')), ]) def convert(self, oeb_book, output_path, input_plugin, opts, log): diff --git a/src/calibre/gui2/convert/fb2_output.py b/src/calibre/gui2/convert/fb2_output.py index a3cbe0e647..5d927146a5 100644 --- a/src/calibre/gui2/convert/fb2_output.py +++ b/src/calibre/gui2/convert/fb2_output.py @@ -17,6 +17,8 @@ class PluginWidget(Widget, Ui_Form): ICON = I('mimetypes/fb2.png') def __init__(self, parent, get_option, get_help, db=None, book_id=None): - Widget.__init__(self, parent, ['inline_toc', 'sectionize_chapters']) + Widget.__init__(self, parent, ['inline_toc', 'sectionize_chapters', + 'sectionize_chapters_using_file_structure', 'h1_to_title', + 'h2_to_title', 'h3_to_title']) self.db, self.book_id = db, book_id self.initialize_options(get_option, get_help, db, book_id) diff --git a/src/calibre/gui2/convert/fb2_output.ui b/src/calibre/gui2/convert/fb2_output.ui index a43a8b72ea..a90ecd615e 100644 --- a/src/calibre/gui2/convert/fb2_output.ui +++ b/src/calibre/gui2/convert/fb2_output.ui @@ -14,7 +14,7 @@ <string>Form</string> </property> <layout class="QGridLayout" name="gridLayout"> - <item row="2" column="0"> + <item row="6" column="0"> <spacer name="verticalSpacer"> <property name="orientation"> <enum>Qt::Vertical</enum> @@ -41,6 +41,34 @@ </property> </widget> </item> + <item row="2" column="0"> + <widget class="QCheckBox" name="opt_sectionize_chapters_using_file_structure"> + <property name="text"> + <string>Sectionize Chapters using file structure</string> + </property> + </widget> + </item> + <item row="3" column="0"> + <widget class="QCheckBox" name="opt_h1_to_title"> + <property name="text"> + <string>Wrap h1 tags with <title> elements</string> + </property> + </widget> + </item> + <item row="4" column="0"> + <widget class="QCheckBox" name="opt_h2_to_title"> + <property name="text"> + <string>Wrap h2 tags with <title> elements</string> + </property> + </widget> + </item> + <item row="5" column="0"> + <widget class="QCheckBox" name="opt_h3_to_title"> + <property name="text"> + <string>Wrap h3 tags with <title> elements</string> + </property> + </widget> + </item> </layout> </widget> <resources/> From 37cde21c6d96b9b385f85d543227b0b3806db879 Mon Sep 17 00:00:00 2001 From: John Schember <john@nachtimwald.com> Date: Tue, 30 Nov 2010 18:48:21 -0500 Subject: [PATCH 2/3] Fix bug #7745: PML output should ignore external links as it is not support by the spec. Fix part of bug #7742: PML output extra % sign. --- src/calibre/ebooks/pml/pmlml.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/calibre/ebooks/pml/pmlml.py b/src/calibre/ebooks/pml/pmlml.py index 560a132ce1..f97f74f4a0 100644 --- a/src/calibre/ebooks/pml/pmlml.py +++ b/src/calibre/ebooks/pml/pmlml.py @@ -216,7 +216,7 @@ class PMLMLizer(object): w = '\\w' width = elem.get('width') if width: - w += '="%s%%"' % width + w += '="%s"' % width else: w += '="50%"' text.append(w) @@ -252,8 +252,8 @@ class PMLMLizer(object): if href not in self.link_hrefs.keys(): self.link_hrefs[href] = 'calibre_link-%s' % len(self.link_hrefs.keys()) href = '#%s' % self.link_hrefs[href] - text.append('\\q="%s"' % href) - tags.append('q') + text.append('\\q="%s"' % href) + tags.append('q') # Anchor ids id_name = elem.get('id') From c38eb08018ca9eb404247de0ccc84bf73196ed20 Mon Sep 17 00:00:00 2001 From: John Schember <john@nachtimwald.com> Date: Tue, 30 Nov 2010 19:04:26 -0500 Subject: [PATCH 3/3] PML Output: ensure \w always ends with a %. --- src/calibre/ebooks/pml/pmlml.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/calibre/ebooks/pml/pmlml.py b/src/calibre/ebooks/pml/pmlml.py index f97f74f4a0..ceb7f36124 100644 --- a/src/calibre/ebooks/pml/pmlml.py +++ b/src/calibre/ebooks/pml/pmlml.py @@ -216,6 +216,8 @@ class PMLMLizer(object): w = '\\w' width = elem.get('width') if width: + if not width.endswith('%'): + width += '%' w += '="%s"' % width else: w += '="50%"'