Widget for strucutre detection settings

2025-07-09 03:04:10 -04:00 · 2009-05-07 21:40:01 -07:00 · 2009-05-07 21:40:01 -07:00 · 9765d91570
commit 9765d91570
parent 2ac2aed136
10 changed files with 1842 additions and 2 deletions
--- a/src/calibre/ebooks/comic/input.py
+++ b/src/calibre/ebooks/comic/input.py
@ -304,6 +304,7 @@ class ComicInput(InputFormatPlugin):
        ('chapter', None, OptionRecommendation.HIGH),
        ('page_breaks_brefore', None, OptionRecommendation.HIGH),
        ('use_auto_toc', False, OptionRecommendation.HIGH),
+        ('page_breaks_before', None, OptionRecommendation.HIGH),
        ])

    def get_comics_from_collection(self, stream):
--- a/src/calibre/gui2/convert/init.py
+++ b/src/calibre/gui2/convert/init.py
@ -154,6 +154,7 @@ class Widget(QWidget):


    def get_value(self, g):
+        from calibre.gui2.convert.xpath_wizard import XPathEdit
        ret = self.get_value_handler(g)
        if ret != 'this is a dummy return value, xcswx1avcx4x':
            return ret
@ -169,11 +170,14 @@ class Widget(QWidget):
            return unicode(g.currentText())
        elif isinstance(g, QCheckBox):
            return bool(g.isChecked())
+        elif isinstance(g, XPathEdit):
+            return g.xpath
        else:
            raise Exception('Can\'t get value from %s'%type(g))


    def set_value(self, g, val):
+        from calibre.gui2.convert.xpath_wizard import XPathEdit
        if self.set_value_handler(g, val):
            return
        if isinstance(g, (QSpinBox, QDoubleSpinBox)):
@ -190,6 +194,8 @@ class Widget(QWidget):
            g.setCurrentIndex(idx)
        elif isinstance(g, QCheckBox):
            g.setCheckState(Qt.Checked if bool(val) else Qt.Unchecked)
+        elif isinstance(g, XPathEdit):
+            g.edit.setText(val if val else '')
        else:
            raise Exception('Can\'t set value %s in %s'%(repr(val), type(g)))
        self.post_set_value(g, val)
@ -223,6 +229,9 @@ class Widget(QWidget):
    def post_get_value(self, g):
        pass

+    def pre_commit_check(self):
+        return True
+
    def commit(self, save_defaults=False):
        return self.commit_options(save_defaults)

--- a/src/calibre/gui2/convert/single.py
+++ b/src/calibre/gui2/convert/single.py
@ -17,6 +17,8 @@ from calibre.gui2.convert.single_ui import Ui_Dialog
 from calibre.gui2.convert.metadata import MetadataWidget
 from calibre.gui2.convert.look_and_feel import LookAndFeelWidget
 from calibre.gui2.convert.page_setup import PageSetupWidget
+from calibre.gui2.convert.structure_detection import StructureDetectionWidget
+

 from calibre.ebooks.conversion.plumber import Plumber, supported_input_formats, \
                    INPUT_FORMAT_PREFERENCES, OUTPUT_FORMAT_PREFERENCES
@ -115,6 +117,7 @@ class Config(ResizableDialog, Ui_Dialog):


    def setup_pipeline(self, *args):
+        oidx = self.groups.currentIndex().row()
        input_format = self.input_format
        output_format = self.output_format
        input_path = self.db.format_abspath(self.book_id, input_format,
@ -134,6 +137,7 @@ class Config(ResizableDialog, Ui_Dialog):
        self.setWindowTitle(_('Convert')+ ' ' + unicode(self.mw.title.text()))
        lf = widget_factory(LookAndFeelWidget)
        ps = widget_factory(PageSetupWidget)
+        sd = widget_factory(StructureDetectionWidget)

        output_widget = None
        name = 'calibre.gui2.convert.%s' % self.plumber.output_plugin.name.lower().replace(' ', '_')
@ -162,7 +166,7 @@ class Config(ResizableDialog, Ui_Dialog):
            if not c: break
            self.stack.removeWidget(c)

-        widgets = [self.mw, lf, ps]
+        widgets = [self.mw, lf, ps, sd]
        if input_widget is not None:
            widgets.append(input_widget)
        if output_widget is not None:
@ -175,7 +179,8 @@ class Config(ResizableDialog, Ui_Dialog):
        self._groups_model = GroupModel(widgets)
        self.groups.setModel(self._groups_model)

-        self.groups.setCurrentIndex(self._groups_model.index(0))
+        idx = oidx if -1 < oidx < self._groups_model.rowCount() else 0
+        self.groups.setCurrentIndex(self._groups_model.index(idx))


    def setup_input_output_formats(self, db, book_id, preferred_input_format,
@ -213,6 +218,8 @@ class Config(ResizableDialog, Ui_Dialog):
    def accept(self):
        recs = GuiRecommendations()
        for w in self._groups_model.widgets:
+            if not w.pre_commit_check():
+                return
            x = w.commit(save_defaults=False)
            recs.update(x)
        self.opf_path, self.cover_path = self.mw.opf_file, self.mw.cover_file
--- a/src/calibre/gui2/convert/structure_detection.py
+++ b/src/calibre/gui2/convert/structure_detection.py
@ -0,0 +1,42 @@
+#!/usr/bin/env python
+# vim:fileencoding=UTF-8:ts=4:sw=4:sta:et:sts=4:ai
+from __future__ import with_statement
+
+__license__   = 'GPL v3'
+__copyright__ = '2009, Kovid Goyal <kovid@kovidgoyal.net>'
+__docformat__ = 'restructuredtext en'
+
+
+from calibre.gui2.convert.structure_detection_ui import Ui_Form
+from calibre.gui2.convert import Widget
+from calibre.gui2 import error_dialog
+
+class StructureDetectionWidget(Widget, Ui_Form):
+
+    TITLE = _('Structure\nDetection')
+    ICON  = ':/images/chapters.svg'
+    HELP  = _('Fine tune the detection of chapter headings and '
+            'other document structure.')
+
+    def __init__(self, parent, get_option, get_help, db=None, book_id=None):
+        Widget.__init__(self, parent, 'structure_detection',
+                ['chapter', 'chapter_mark',
+                'remove_first_image',
+                'insert_metadata', 'page_breaks_before',
+                'preprocess_html']
+                )
+        self.db, self.book_id = db, book_id
+        self.initialize_options(get_option, get_help, db, book_id)
+        self.opt_chapter.set_msg(_('Detect chapters at (XPath expression):'))
+        self.opt_page_breaks_before.set_msg(_('Insert page breaks before '
+            '(XPath expression):'))
+
+
+    def pre_commit_check(self):
+        for x in ('chapter', 'page_breaks_before'):
+            x = getattr(self, 'opt_'+x)
+            if not x.check():
+                error_dialog(self, _('Invalid XPath'),
+                _('The XPath expression %s is invalid.')%x.xpath).exec_()
+                return False
+        return True
--- a/src/calibre/gui2/convert/structure_detection.ui
+++ b/src/calibre/gui2/convert/structure_detection.ui
@ -0,0 +1,103 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<ui version="4.0">
+ <class>Form</class>
+ <widget class="QWidget" name="Form">
+  <property name="geometry">
+   <rect>
+    <x>0</x>
+    <y>0</y>
+    <width>657</width>
+    <height>479</height>
+   </rect>
+  </property>
+  <property name="windowTitle">
+   <string>Form</string>
+  </property>
+  <layout class="QGridLayout" name="gridLayout">
+   <item row="1" column="0">
+    <widget class="QLabel" name="label">
+     <property name="text">
+      <string>Chapter &amp;mark:</string>
+     </property>
+     <property name="buddy">
+      <cstring>opt_chapter_mark</cstring>
+     </property>
+    </widget>
+   </item>
+   <item row="1" column="1">
+    <widget class="QComboBox" name="opt_chapter_mark">
+     <item>
+      <property name="text">
+       <string>pagebreak</string>
+      </property>
+     </item>
+     <item>
+      <property name="text">
+       <string>rule</string>
+      </property>
+     </item>
+     <item>
+      <property name="text">
+       <string>both</string>
+      </property>
+     </item>
+     <item>
+      <property name="text">
+       <string>none</string>
+      </property>
+     </item>
+    </widget>
+   </item>
+   <item row="2" column="0">
+    <widget class="QCheckBox" name="opt_remove_first_image">
+     <property name="text">
+      <string>Remove first &amp;image</string>
+     </property>
+    </widget>
+   </item>
+   <item row="3" column="0">
+    <widget class="QCheckBox" name="opt_insert_metadata">
+     <property name="text">
+      <string>Insert &amp;metadata as page at start of book</string>
+     </property>
+    </widget>
+   </item>
+   <item row="4" column="0" colspan="2">
+    <widget class="QCheckBox" name="opt_preprocess_html">
+     <property name="text">
+      <string>&amp;Preprocess input file to possibly improve structure detection</string>
+     </property>
+    </widget>
+   </item>
+   <item row="5" column="0" colspan="2">
+    <widget class="XPathEdit" name="opt_page_breaks_before" native="true"/>
+   </item>
+   <item row="0" column="0" colspan="2">
+    <widget class="XPathEdit" name="opt_chapter" native="true"/>
+   </item>
+   <item row="6" column="0">
+    <spacer name="verticalSpacer">
+     <property name="orientation">
+      <enum>Qt::Vertical</enum>
+     </property>
+     <property name="sizeHint" stdset="0">
+      <size>
+       <width>20</width>
+       <height>40</height>
+      </size>
+     </property>
+    </spacer>
+   </item>
+  </layout>
+ </widget>
+ <customwidgets>
+  <customwidget>
+   <class>XPathEdit</class>
+   <extends>QWidget</extends>
+   <header>convert/xpath_wizard.h</header>
+   <container>1</container>
+  </customwidget>
+ </customwidgets>
+ <resources/>
+ <connections/>
+</ui>
--- a/src/calibre/gui2/convert/xpath_edit.ui
+++ b/src/calibre/gui2/convert/xpath_edit.ui
@ -0,0 +1,59 @@
+<ui version="4.0" >
+ <class>Form</class>
+ <widget class="QWidget" name="Form" >
+  <property name="geometry" >
+   <rect>
+    <x>0</x>
+    <y>0</y>
+    <width>400</width>
+    <height>64</height>
+   </rect>
+  </property>
+  <property name="windowTitle" >
+   <string>Form</string>
+  </property>
+  <layout class="QHBoxLayout" name="horizontalLayout" >
+   <item>
+    <layout class="QVBoxLayout" name="verticalLayout" >
+     <item>
+      <widget class="QLabel" name="msg" >
+       <property name="text" >
+        <string>TextLabel</string>
+       </property>
+       <property name="wordWrap" >
+        <bool>true</bool>
+       </property>
+      </widget>
+     </item>
+     <item>
+      <widget class="QLineEdit" name="edit" />
+     </item>
+    </layout>
+   </item>
+   <item>
+    <widget class="QToolButton" name="button" >
+     <property name="toolTip" >
+      <string>Use a wizard to help construct the XPath expression</string>
+     </property>
+     <property name="text" >
+      <string>...</string>
+     </property>
+     <property name="icon" >
+      <iconset resource="../images.qrc" >
+       <normaloff>:/images/wizard.svg</normaloff>:/images/wizard.svg</iconset>
+     </property>
+     <property name="iconSize" >
+      <size>
+       <width>40</width>
+       <height>40</height>
+      </size>
+     </property>
+    </widget>
+   </item>
+  </layout>
+ </widget>
+ <resources>
+  <include location="../images.qrc" />
+ </resources>
+ <connections/>
+</ui>
--- a/src/calibre/gui2/convert/xpath_wizard.py
+++ b/src/calibre/gui2/convert/xpath_wizard.py
@ -0,0 +1,89 @@
+#!/usr/bin/env python
+# vim:fileencoding=UTF-8:ts=4:sw=4:sta:et:sts=4:ai
+from __future__ import with_statement
+
+__license__   = 'GPL v3'
+__copyright__ = '2009, Kovid Goyal <kovid@kovidgoyal.net>'
+__docformat__ = 'restructuredtext en'
+
+from PyQt4.Qt import QDialog, QWidget, SIGNAL, Qt, QDialogButtonBox, QVBoxLayout
+
+from calibre.gui2.convert.xpath_wizard_ui import Ui_Form
+from calibre.gui2.convert.xpath_edit_ui import Ui_Form as Ui_Edit
+
+
+class WizardWidget(QWidget, Ui_Form):
+
+    def __init__(self, parent=None):
+        QWidget.__init__(self, parent)
+        self.setupUi(self)
+
+    @property
+    def xpath(self):
+        tag = unicode(self.tag.currentText()).strip()
+        if tag != '*':
+            tag = 'h:'+tag
+        attr, val = map(unicode, (self.attribute.text(), self.value.text()))
+        attr, val = attr.strip(), val.strip()
+        q = ''
+        if attr:
+            if val:
+                q = '[re:test(@%s, "%s", "i")]'%(attr, val)
+            else:
+                q = '[@%s]'%attr
+        expr = '//'+tag + q
+        return expr
+
+class Wizard(QDialog):
+
+    def __init__(self, parent=None):
+        QDialog.__init__(self, parent)
+        self.resize(400, 300)
+        self.verticalLayout = QVBoxLayout(self)
+        self.widget = WizardWidget(self)
+        self.verticalLayout.addWidget(self.widget)
+        self.buttonBox = QDialogButtonBox(self)
+        self.buttonBox.setOrientation(Qt.Horizontal)
+        self.buttonBox.setStandardButtons(QDialogButtonBox.Cancel|QDialogButtonBox.Ok)
+        self.verticalLayout.addWidget(self.buttonBox)
+
+        self.connect(self.buttonBox, SIGNAL("accepted()"), self.accept)
+        self.connect(self.buttonBox, SIGNAL("rejected()"), self.reject)
+        self.setModal(Qt.WindowModal)
+
+    @property
+    def xpath(self):
+        return self.widget.xpath
+
+
+class XPathEdit(QWidget, Ui_Edit):
+
+    def __init__(self, parent=None):
+        QWidget.__init__(self, parent)
+        self.setupUi(self)
+        self.connect(self.button, SIGNAL('clicked()'), self.wizard)
+
+    def wizard(self):
+        wiz = Wizard(self)
+        if wiz.exec_() == wiz.Accepted:
+            self.edit.setText(wiz.xpath)
+
+
+    def set_msg(self, msg):
+        self.msg.setText(msg)
+
+    @property
+    def text(self):
+        return unicode(self.edit.text())
+
+    def check(self):
+        from calibre.ebooks.oeb.base import XPNSMAP
+        from lxml.etree import XPath
+        try:
+            XPath(self.text, namespace=XPNSMAP)
+        except:
+            return False
+        return True
+
+
+
--- a/src/calibre/gui2/convert/xpath_wizard.ui
+++ b/src/calibre/gui2/convert/xpath_wizard.ui
@ -0,0 +1,143 @@
+<ui version="4.0" >
+ <class>Form</class>
+ <widget class="QWidget" name="Form" >
+  <property name="geometry" >
+   <rect>
+    <x>0</x>
+    <y>0</y>
+    <width>400</width>
+    <height>381</height>
+   </rect>
+  </property>
+  <property name="windowTitle" >
+   <string>Form</string>
+  </property>
+  <layout class="QGridLayout" name="gridLayout" >
+   <item row="0" column="0" colspan="2" >
+    <widget class="QLabel" name="label" >
+     <property name="text" >
+      <string>Match HTML &amp;tags with tag name:</string>
+     </property>
+     <property name="buddy" >
+      <cstring>tag</cstring>
+     </property>
+    </widget>
+   </item>
+   <item row="1" column="0" >
+    <widget class="QComboBox" name="tag" >
+     <property name="editable" >
+      <bool>true</bool>
+     </property>
+     <item>
+      <property name="text" >
+       <string>*</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>a</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>br</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>div</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>h1</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>h2</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>h3</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>h4</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>h5</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>h6</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>hr</string>
+      </property>
+     </item>
+     <item>
+      <property name="text" >
+       <string>span</string>
+      </property>
+     </item>
+    </widget>
+   </item>
+   <item row="2" column="0" colspan="2" >
+    <widget class="QLabel" name="label_2" >
+     <property name="text" >
+      <string>Having the &amp;attribute:</string>
+     </property>
+     <property name="buddy" >
+      <cstring>attribute</cstring>
+     </property>
+    </widget>
+   </item>
+   <item row="3" column="0" colspan="2" >
+    <widget class="QLineEdit" name="attribute" />
+   </item>
+   <item row="4" column="0" >
+    <widget class="QLabel" name="label_3" >
+     <property name="text" >
+      <string>With &amp;value:</string>
+     </property>
+     <property name="buddy" >
+      <cstring>value</cstring>
+     </property>
+    </widget>
+   </item>
+   <item row="5" column="0" >
+    <widget class="QLineEdit" name="value" />
+   </item>
+   <item row="5" column="1" >
+    <widget class="QLabel" name="label_4" >
+     <property name="text" >
+      <string>(A regular expression)</string>
+     </property>
+    </widget>
+   </item>
+   <item row="6" column="0" colspan="2" >
+    <widget class="QLabel" name="label_5" >
+     <property name="text" >
+      <string>&lt;p>For example, to match all h2 tags that have class="chapter", set tag to &lt;i>h2&lt;/i>, attribute to &lt;i>class&lt;/i> and value to &lt;i>chapter&lt;/i>.&lt;/p>&lt;p>Leaving attribute blank will match any attribute and leaving value blank will match any value. Setting tag to * will match any tag.&lt;/p>&lt;p>To learn more advanced usage of XPath see the &lt;a href="http://calibre.kovidgoyal.net/user_manual/xpath.html">XPath Tutorial&lt;/a>.</string>
+     </property>
+     <property name="wordWrap" >
+      <bool>true</bool>
+     </property>
+     <property name="openExternalLinks" >
+      <bool>true</bool>
+     </property>
+    </widget>
+   </item>
+  </layout>
+ </widget>
+ <resources/>
+ <connections/>
+</ui>
--- a/src/calibre/gui2/images/wizard.svg
+++ b/src/calibre/gui2/images/wizard.svg
--- a/upload.py
+++ b/upload.py
@ -312,6 +312,8 @@ class gui(OptionlessCommand):
                dat = dat.replace('import images_rc', 'from calibre.gui2 import images_rc')
                dat = dat.replace('from library import', 'from calibre.gui2.library import')
                dat = dat.replace('from widgets import', 'from calibre.gui2.widgets import')
+                dat = dat.replace('from convert.xpath_wizard import',
+                    'from calibre.gui2.convert.xpath_wizard import')
                dat = re.compile(r'QtGui.QApplication.translate\(.+?,\s+"(.+?)(?<!\\)",.+?\)', re.DOTALL).sub(r'_("\1")', dat)

                # Workaround bug in Qt 4.4 on Windows