Use a wizard for XPath input

src/calibre/ebooks/html_transform_rules.py

@@ -3,7 +3,11 @@
 # License: GPLv3 Copyright: 2016, Kovid Goyal <kovid at kovidgoyal.net>
 
 
+from functools import partial
+
 from calibre.utils.serialize import json_dumps, json_loads
+from calibre.ebooks.oeb.base import XPath
+from css_selectors.select import get_parsed_selector
 
 
 class Action:
@@ -43,20 +47,46 @@ ACTION_MAP = {a.name: a for a in (
 )}
 
 
+def non_empty_validator(label, val):
+    if not val:
+        return _('{} must not be empty').format(label)
+
+
+def always_valid(*a):
+    pass
+
+
+def validate_css_selector(val):
+    try:
+        get_parsed_selector(val)
+    except Exception:
+        return _('{} is not a valid CSS selector').format(val)
+
+
+def validate_xpath_selector(val):
+    try:
+        XPath(val)
+    except Exception:
+        return _('{} is not a valid XPath selector').format(val)
+
+
 class Match:
 
-    def __init__(self, name, text, placeholder=''):
+    def __init__(self, name, text, placeholder='', validator=None):
         self.name = name
         self.text = text
         self.placeholder = placeholder
+        if validator is None and placeholder:
+            validator = partial(non_empty_validator, self.placeholder)
+        self.validator = validator or always_valid
 
 
 MATCH_TYPE_MAP = {m.name: m for m in (
     Match('is', _('is'), _('Tag name')),
     Match('has_class', _('has class'), _('Class name')),
     Match('not_has_class', _('does not have class'), _('Class name')),
-    Match('css', _('matches CSS selector'), _('CSS selector')),
+    Match('css', _('matches CSS selector'), _('CSS selector'), validate_css_selector),
-    Match('xpath', _('matches XPath selector'), _('XPath selector')),
+    Match('xpath', _('matches XPath selector'), _('XPath selector'), validate_xpath_selector),
     Match('*', _('is any tag')),
 )}
 

src/calibre/gui2/convert/regex_builder.py

@@ -6,14 +6,14 @@ import os
 from contextlib import suppress
 from qt.core import (
     QApplication, QBrush, QByteArray, QDialog, QDialogButtonBox, Qt, QTextCursor,
-    QTextEdit, QWidget, pyqtSignal
+    QTextEdit, pyqtSignal
 )
 
 from calibre.constants import iswindows
 from calibre.ebooks.conversion.search_replace import compile_regular_expression
 from calibre.gui2 import choose_files, error_dialog, gprefs
 from calibre.gui2.convert.regex_builder_ui import Ui_RegexBuilder
-from calibre.gui2.convert.xexp_edit_ui import Ui_Form as Ui_Edit
+from calibre.gui2.convert.xpath_wizard import XPathEdit
 from calibre.gui2.dialogs.choose_format import ChooseFormatDialog
 from calibre.ptempfile import TemporaryFile
 from calibre.utils.icu import utf16_length
@@ -212,20 +212,19 @@ class RegexBuilder(QDialog, Ui_RegexBuilder):
         return str(self.preview.toPlainText())
 
 
-class RegexEdit(QWidget, Ui_Edit):
+class RegexEdit(XPathEdit):
 
     doc_update = pyqtSignal(str)
 
     def __init__(self, parent=None):
-        QWidget.__init__(self, parent)
+        super().__init__(parent)
-        self.setupUi(self)
         self.edit.completer().setCaseSensitivity(Qt.CaseSensitivity.CaseSensitive)
-
         self.book_id = None
         self.db = None
         self.doc_cache = None
 
-        self.button.clicked.connect(self.builder)
+    def wizard(self):
+        return self.builder()
 
     def builder(self):
         if self.db is None:
@@ -244,7 +243,7 @@ class RegexEdit(QWidget, Ui_Edit):
         return self.doc_cache
 
     def setObjectName(self, *args):
-        QWidget.setObjectName(self, *args)
+        super().setObjectName(*args)
         if hasattr(self, 'edit'):
             self.edit.initialize('regex_edit_'+str(self.objectName()))
 

src/calibre/gui2/convert/xexp_edit.ui

@@ -1,89 +0,0 @@
-<?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>430</width>
-    <height>74</height>
-   </rect>
-  </property>
-  <property name="windowTitle">
-   <string>Form</string>
-  </property>
-  <layout class="QGridLayout" name="gridLayout">
-   <item row="0" column="0">
-    <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>
-       <property name="buddy">
-        <cstring>edit</cstring>
-       </property>
-      </widget>
-     </item>
-     <item>
-      <widget class="HistoryLineEdit" name="edit">
-       <property name="sizePolicy">
-        <sizepolicy hsizetype="Expanding" vsizetype="Fixed">
-         <horstretch>100</horstretch>
-         <verstretch>0</verstretch>
-        </sizepolicy>
-       </property>
-       <property name="minimumSize">
-        <size>
-         <width>350</width>
-         <height>0</height>
-        </size>
-       </property>
-       <property name="sizeAdjustPolicy">
-        <enum>QComboBox::AdjustToMinimumContentsLengthWithIcon</enum>
-       </property>
-       <property name="minimumContentsLength">
-        <number>30</number>
-       </property>
-      </widget>
-     </item>
-    </layout>
-   </item>
-   <item row="0" column="1">
-    <widget class="QToolButton" name="button">
-     <property name="toolTip">
-      <string>Use a wizard to help construct the Regular expression</string>
-     </property>
-     <property name="text">
-      <string>...</string>
-     </property>
-     <property name="icon">
-      <iconset resource="../../../../resources/images.qrc">
-       <normaloff>:/images/wizard.png</normaloff>:/images/wizard.png</iconset>
-     </property>
-     <property name="iconSize">
-      <size>
-       <width>40</width>
-       <height>40</height>
-      </size>
-     </property>
-    </widget>
-   </item>
-  </layout>
- </widget>
- <customwidgets>
-  <customwidget>
-   <class>HistoryLineEdit</class>
-   <extends>QComboBox</extends>
-   <header>calibre/gui2/widgets.h</header>
-  </customwidget>
- </customwidgets>
- <resources>
-  <include location="../../../../resources/images.qrc"/>
- </resources>
- <connections/>
-</ui>

src/calibre/gui2/convert/xpath_wizard.py

@@ -6,10 +6,13 @@ __license__   = 'GPL v3'
 __copyright__ = '2009, Kovid Goyal <kovid@kovidgoyal.net>'
 __docformat__ = 'restructuredtext en'
 
-from qt.core import QDialog, QWidget, Qt, QDialogButtonBox, QVBoxLayout
+from qt.core import (
+    QComboBox, QDialog, QDialogButtonBox, QHBoxLayout, QIcon, QLabel, QSize, Qt,
+    QToolButton, QVBoxLayout, QWidget
+)
 
 from calibre.gui2.convert.xpath_wizard_ui import Ui_Form
-from calibre.gui2.convert.xexp_edit_ui import Ui_Form as Ui_Edit
+from calibre.gui2.widgets import HistoryLineEdit
 from calibre.utils.localization import localize_user_manual_link
 
 
@@ -65,12 +68,37 @@ class Wizard(QDialog):
         return self.widget.xpath
 
 
-class XPathEdit(QWidget, Ui_Edit):
+class XPathEdit(QWidget):
 
-    def __init__(self, parent=None):
+    def __init__(self, parent=None, object_name='', show_msg=True):
         QWidget.__init__(self, parent)
-        self.setupUi(self)
+        self.h = h = QHBoxLayout(self)
-        self.button.clicked.connect(self.wizard)
+        h.setContentsMargins(0, 0, 0, 0)
+        self.l = l = QVBoxLayout()
+        h.addLayout(l)
+        self.button = b = QToolButton(self)
+        b.setIcon(QIcon(I('wizard.png')))
+        b.setToolTip(_('Use a wizard to generate the XPath expression'))
+        b.clicked.connect(self.wizard)
+        h.addWidget(b)
+        self.edit = e = HistoryLineEdit(self)
+        e.setMinimumWidth(350)
+        e.setSizeAdjustPolicy(QComboBox.SizeAdjustPolicy.AdjustToMinimumContentsLengthWithIcon)
+        e.setMinimumContentsLength(30)
+        self.msg = QLabel('')
+        l.addWidget(self.msg)
+        l.addWidget(self.edit)
+        if object_name:
+            self.setObjectName(object_name)
+        if show_msg:
+            b.setIconSize(QSize(40, 40))
+            self.msg.setBuddy(self.edit)
+        else:
+            self.msg.setVisible(False)
+            l.setContentsMargins(0, 0, 0, 0)
+
+    def setPlaceholderText(self, val):
+        self.edit.setPlaceholderText(val)
 
     def wizard(self):
         wiz = Wizard(self)
@@ -89,6 +117,11 @@ class XPathEdit(QWidget, Ui_Edit):
     def text(self):
         return str(self.edit.text())
 
+    @text.setter
+    def text(self, val):
+        self.edit.setText(str(val))
+    value = text
+
     @property
     def xpath(self):
         return self.text

src/calibre/gui2/convert/xpath_wizard.ui

@@ -102,7 +102,11 @@
     </widget>
    </item>
    <item>
-    <widget class="QLineEdit" name="attribute"/>
+    <widget class="QLineEdit" name="attribute">
+     <property name="clearButtonEnabled">
+      <bool>true</bool>
+     </property>
+    </widget>
    </item>
    <item>
     <widget class="QLabel" name="label_3">

src/calibre/gui2/html_transform_rules.py

@@ -15,13 +15,13 @@ from calibre.ebooks.html_transform_rules import (
     validate_rule
 )
 from calibre.gui2 import choose_files, choose_save_file, elided_text, error_dialog
+from calibre.gui2.convert.xpath_wizard import XPathEdit
 from calibre.gui2.tag_mapper import (
     RuleEditDialog as RuleEditDialogBase, RuleItem as RuleItemBase,
     Rules as RulesBase, RulesDialog as RulesDialogBase, SaveLoadMixin
 )
 from calibre.gui2.widgets2 import Dialog
 from calibre.utils.config import JSONConfig
-from calibre.utils.localization import localize_user_manual_link
 
 
 class TagAction(QWidget):
@@ -138,6 +138,44 @@ class ActionsContainer(QScrollArea):
             self.new_action().as_dict = entry
 
 
+class GenericEdit(QLineEdit):
+
+    def __init__(self, parent=None):
+        super().__init__(parent)
+        self.setClearButtonEnabled(True)
+
+    @property
+    def value(self):
+        return self.text()
+
+    @value.setter
+    def value(self, val):
+        self.setText(str(val))
+
+
+class CSSEdit(QWidget):
+
+    def __init__(self, parent=None):
+        super().__init__(parent)
+        l = QHBoxLayout(self)
+        l.setContentsMargins(0, 0, 0, 0)
+        self.edit = le = GenericEdit(self)
+        l.addWidget(le)
+        l.addSpacing(5)
+        self.la = la = QLabel(_('<a href="{}">CSS selector help</a>').format('https://developer.mozilla.org/en-US/docs/Learn/CSS/Building_blocks/Selectors'))
+        la.setOpenExternalLinks(True)
+        l.addWidget(la)
+        self.setPlaceholderText = self.edit.setPlaceholderText
+
+    @property
+    def value(self):
+        return self.edit.value
+
+    @value.setter
+    def value(self, val):
+        self.edit.value = val
+
+
 class RuleEdit(QWidget):  # {{{
 
     MSG = _('Create the rule to transform HTML tags below')
@@ -152,7 +190,7 @@ class RuleEdit(QWidget):  # {{{
         l.addWidget(la)
         l.addLayout(h)
         english_sentence = '{preamble} {match_type}'
-        sentence = _('{preamble} {match_type} {query}')
+        sentence = _('{preamble} {match_type}')
         if set(sentence.split()) != set(english_sentence.split()):
             sentence = english_sentence
         parts = sentence.split()
@@ -168,15 +206,10 @@ class RuleEdit(QWidget):  # {{{
             if clause is not parts[-1]:
                 h.addWidget(QLabel('\xa0'))
         h.addStretch(1)
-        self.hl = h = QHBoxLayout()
+        self.generic_query = gq = GenericEdit(self)
-        l.addLayout(h)
+        self.css_query = cq = CSSEdit(self)
-        self.query = q = QLineEdit(self)
+        self.xpath_query = xq = XPathEdit(self, object_name='html_transform_rules_xpath', show_msg=False)
-        q.setClearButtonEnabled(True)
+        l.addWidget(gq), l.addWidget(cq), l.addWidget(xq)
-        h.addWidget(q)
-        h.addSpacing(20)
-        self.query_help_label = la = QLabel(self)
-        la.setOpenExternalLinks(True)
-        h.addWidget(la)
 
         self.thenl = QLabel(_('Then:'))
         l.addWidget(self.thenl)
@@ -193,28 +226,29 @@ class RuleEdit(QWidget):  # {{{
         a.setWidth(a.width() + 125)
         return a
 
+    @property
+    def current_query_widget(self):
+        return {'css': self.css_query, 'xpath': self.xpath_query}.get(self.match_type.currentData(), self.generic_query)
+
     def update_state(self):
         r = self.rule
         mt = r['match_type']
-        self.query.setVisible(mt != '*')
+        self.generic_query.setVisible(False), self.css_query.setVisible(False), self.xpath_query.setVisible(False)
-        self.query.setPlaceholderText(MATCH_TYPE_MAP[mt].placeholder)
+        self.current_query_widget.setVisible(True)
-        self.query_help_label.setVisible(mt in ('css', 'xpath'))
+        self.current_query_widget.setPlaceholderText(MATCH_TYPE_MAP[mt].placeholder)
-        if self.query_help_label.isVisible():
-            if mt == 'css':
-                url = 'https://developer.mozilla.org/en-US/docs/Learn/CSS/Building_blocks/Selectors'
-                text = _('CSS selector help')
-            else:
-                url = localize_user_manual_link('https://manual.calibre-ebook.com/xpath.html')
-                text = _('XPath selector help')
-            self.query_help_label.setText(f'<a href="{url}">{text}</a>')
 
     @property
     def rule(self):
-        return {
+        try:
-            'match_type': self.match_type.currentData(),
+            return {
-            'query': self.query.text().strip(),
+                'match_type': self.match_type.currentData(),
-            'actions': self.actions.as_list,
+                'query': self.current_query_widget.value,
-        }
+                'actions': self.actions.as_list,
+            }
+        except Exception:
+            import traceback
+            traceback.print_exc()
+            raise
 
     @rule.setter
     def rule(self, rule):
@@ -222,7 +256,7 @@ class RuleEdit(QWidget):  # {{{
             c = getattr(self, name)
             c.setCurrentIndex(max(0, c.findData(str(rule.get(name, '')))))
         sc('match_type')
-        self.query.setText(str(rule.get('query', '')).strip())
+        self.current_query_widget.value = str(rule.get('query', '')).strip()
         self.actions.as_list = rule.get('actions') or []
         self.update_state()
 
@@ -419,7 +453,7 @@ if __name__ == '__main__':
     app = Application([])
     d = RulesDialog()
     d.rules = [
-        {'match_type':'*', 'query':'', 'actions':[{'type': 'remove'}]},
+        {'match_type':'xpath', 'query':'//h:h2', 'actions':[{'type': 'remove'}]},
     ]
     d.exec_()
     from pprint import pprint