Edit Book: The Remove Unused CSS tool now has an option to also merge CSS rules that have identical selectors

src/calibre/ebooks/oeb/polish/css.py

@@ -6,6 +6,7 @@ from __future__ import (unicode_literals, division, absolute_import,
 __license__ = 'GPL v3'
 __copyright__ = '2014, Kovid Goyal <kovid at kovidgoyal.net>'
 
+from collections import defaultdict
 from functools import partial
 
 from cssutils.css import CSSRule, CSSStyleDeclaration
@@ -48,12 +49,35 @@ def get_imported_sheets(name, container, sheets, recursion_level=10, sheet=None)
     ans.discard(name)
     return ans
 
-def remove_unused_css(container, report=None, remove_unused_classes=False):
+
+def merge_declarations(first, second):
+    for prop in second.getProperties():
+        first.setProperty(prop)
+
+
+def merge_identical_selectors(sheet):
+    ' Merge rules that have identical selectors '
+    selector_map = defaultdict(list)
+    for rule in sheet.cssRules.rulesOfType(CSSRule.STYLE_RULE):
+        selector_map[rule.selectorText].append(rule)
+    remove = []
+    for rule_group in selector_map.itervalues():
+        if len(rule_group) > 1:
+            for i in range(1, len(rule_group)):
+                merge_declarations(rule_group[0].style, rule_group[i].style)
+                remove.append(rule_group[i])
+    for rule in remove:
+        sheet.cssRules.remove(rule)
+    return len(remove)
+
+
+def remove_unused_css(container, report=None, remove_unused_classes=False, merge_rules=False):
     '''
     Remove all unused CSS rules from the book. An unused CSS rule is one that does not match any actual content.
 
     :param report: An optional callable that takes a single argument. It is called with information about the operations being performed.
     :param remove_unused_classes: If True, class attributes in the HTML that do not match any CSS rules are also removed.
+    :param merge_rules: If True, rules with identical selectors are merged.
     '''
     report = report or (lambda x:x)
 
@@ -64,6 +88,13 @@ def remove_unused_css(container, report=None, remove_unused_classes=False):
             pass
     sheets = {name:safe_parse(name) for name, mt in container.mime_map.iteritems() if mt in OEB_STYLES}
     sheets = {k:v for k, v in sheets.iteritems() if v is not None}
+    num_merged = 0
+    if merge_rules:
+        for name, sheet in sheets.iteritems():
+            num = merge_identical_selectors(sheet)
+            if num:
+                container.dirty(name)
+                num_merged += num
     import_map = {name:get_imported_sheets(name, container, sheets) for name in sheets}
     if remove_unused_classes:
         class_map = {name:{icu_lower(x) for x in classes_in_rule_list(sheet.cssRules)} for name, sheet in sheets.iteritems()}
@@ -80,6 +111,11 @@ def remove_unused_css(container, report=None, remove_unused_classes=False):
         for style in root.xpath('//*[local-name()="style"]'):
             if style.get('type', 'text/css') == 'text/css' and style.text:
                 sheet = container.parse_css(style.text)
+                if merge_rules:
+                    num = merge_identical_selectors(sheet)
+                    if num:
+                        num_merged += num
+                        container.dirty(name)
                 if remove_unused_classes:
                     used_classes |= {icu_lower(x) for x in classes_in_rule_list(sheet.cssRules)}
                 imports = get_imported_sheets(name, container, sheets, sheet=sheet)
@@ -130,18 +166,24 @@ def remove_unused_css(container, report=None, remove_unused_classes=False):
             [sheet.cssRules.remove(r) for r in unused_rules]
             container.dirty(name)
 
-    if num_of_removed_rules > 0:
+    num_changes = num_of_removed_rules + num_merged + num_of_removed_classes
-        report(ngettext('Removed %d unused CSS style rule', 'Removed %d unused CSS style rules',
+    if num_changes > 0:
-                        num_of_removed_rules) % num_of_removed_rules)
+        if num_of_removed_rules > 0:
-    else:
+            report(ngettext('Removed %d unused CSS style rule', 'Removed %d unused CSS style rules',
-        report(_('No unused CSS style rules found'))
+                            num_of_removed_rules) % num_of_removed_rules)
-    if remove_unused_classes:
         if num_of_removed_classes > 0:
             report(ngettext('Removed %d unused class from the HTML', 'Removed %d unused classes from the HTML',
                    num_of_removed_classes) % num_of_removed_classes)
-        else:
+        if num_merged > 0:
-            report(_('No unused class attributes found'))
+            report(ngettext('Merged %d CSS style rule', 'Merged %d CSS style rules',
-    return num_of_removed_rules + num_of_removed_classes > 0
+                            num_merged) % num_merged)
+    if num_of_removed_rules == 0:
+        report(_('No unused CSS style rules found'))
+    if remove_unused_classes and num_of_removed_classes == 0:
+        report(_('No unused class attributes found'))
+    if merge_rules and num_merged == 0:
+        report(_('No style rules that could be merged found'))
+    return num_changes > 0
 
 def filter_declaration(style, properties=()):
     changed = False

src/calibre/ebooks/oeb/polish/main.py

@@ -37,6 +37,7 @@ ALL_OPTS = {
 
 CUSTOMIZATION = {
     'remove_unused_classes': False,
+    'merge_identical_selectors': False,
 }
 
 SUPPORTED = {'EPUB', 'AZW3'}
@@ -201,7 +202,8 @@ def polish_one(ebook, opts, report, customization=None):
 
     if opts.remove_unused_css:
         rt(_('Removing unused CSS rules'))
-        if remove_unused_css(ebook, report, remove_unused_classes=customization['remove_unused_classes']):
+        if remove_unused_css(
+                ebook, report, remove_unused_classes=customization['remove_unused_classes'], merge_rules=customization['merge_identical_selectors']):
             changed = True
         report('')
 

src/calibre/gui2/tweak_book/__init__.py

@@ -49,6 +49,7 @@ d['spell_check_case_sensitive_sort'] = False
 d['inline_spell_check'] = True
 d['custom_themes'] = {}
 d['remove_unused_classes'] = False
+d['merge_identical_selectors'] = False
 d['global_book_toolbar'] = [
 'new-file', 'open-book',  'save-book', None, 'global-undo', 'global-redo', 'create-checkpoint', None, 'donate', 'user-manual']
 d['global_tools_toolbar'] = [

src/calibre/gui2/tweak_book/polish.py

@@ -29,14 +29,25 @@ def customize_remove_unused_css(name, parent, ans):
     d.l = l = QVBoxLayout()
     d.setLayout(d.l)
     d.setWindowTitle(_('Remove unused CSS'))
-    d.la = la = QLabel(_(
+    def label(text):
-        'This will remove all CSS rules that do not match any actual content. You'
+        la = QLabel(text)
-        ' can also have it automatically remove any class attributes from the HTML'
+        la.setWordWrap(True), l.addWidget(la), la.setMinimumWidth(450)
-        ' that do not match any CSS rules, by using the check box below:'))
+        l.addWidget(la)
-    la.setWordWrap(True), l.addWidget(la)
+        return la
+
+    d.la = label(_(
+        'This will remove all CSS rules that do not match any actual content.'
+        ' There are a couple of additional cleanups you can enable, below:'))
     d.c = c = QCheckBox(_('Remove unused &class attributes'))
     c.setChecked(tprefs['remove_unused_classes'])
     l.addWidget(c)
+    d.la2 = label('<span style="font-size:small; font-style: italic">' + _(
+        'Remove all class attributes from the HTML that do not match any existing CSS rules'))
+    d.m = m = QCheckBox(_('Merge identical CSS rules'))
+    m.setChecked(tprefs['merge_identical_selectors'])
+    l.addWidget(m)
+    d.la3 = label('<span style="font-size:small; font-style: italic">' + _(
+        'Merge CSS rules in the same stylesheet that have identical selectors.'))
     d.bb = QDialogButtonBox(QDialogButtonBox.Ok | QDialogButtonBox.Cancel)
     d.l.addWidget(d.bb)
     d.bb.rejected.connect(d.reject)
@@ -44,6 +55,7 @@ def customize_remove_unused_css(name, parent, ans):
     if d.exec_() != d.Accepted:
         raise Abort()
     ans['remove_unused_classes'] = tprefs['remove_unused_classes'] = c.isChecked()
+    ans['merge_identical_selectors'] = tprefs['merge_identical_selectors'] = m.isChecked()
 
 def get_customization(action, name, parent):
     ans = CUSTOMIZATION.copy()