Edit book: Add a tool to split the tag at the current cursor position,

creating a new tag with the same style and class attributes. To add the tool go to the Toolbars section in the editor preferences.  Fixes #1912958 [Feature: Merge/split paragraphs tool in ebook editor](https://bugs.launchpad.net/calibre/+bug/1912958)

imgsrc/split.svg

@@ -0,0 +1,54 @@
+<?xml version="1.0" encoding="UTF-8" standalone="no"?>
+<svg
+   xmlns:dc="http://purl.org/dc/elements/1.1/"
+   xmlns:cc="http://creativecommons.org/ns#"
+   xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
+   xmlns:svg="http://www.w3.org/2000/svg"
+   xmlns="http://www.w3.org/2000/svg"
+   xmlns:sodipodi="http://sodipodi.sourceforge.net/DTD/sodipodi-0.dtd"
+   xmlns:inkscape="http://www.inkscape.org/namespaces/inkscape"
+   height="300px"
+   width="300px"
+   fill="#000000"
+   version="1.1"
+   x="0px"
+   y="0px"
+   viewBox="0 0 97.8 97.8"
+   enable-background="new 0 0 97.8 97.8"
+   xml:space="preserve"
+   id="svg4"
+   sodipodi:docname="split.svg"
+   inkscape:version="1.0.2 (e86c870879, 2021-01-15)"><metadata
+     id="metadata10"><rdf:RDF><cc:Work
+         rdf:about=""><dc:format>image/svg+xml</dc:format><dc:type
+           rdf:resource="http://purl.org/dc/dcmitype/StillImage" /></cc:Work></rdf:RDF></metadata><defs
+     id="defs8" /><sodipodi:namedview
+     pagecolor="#ffffff"
+     bordercolor="#666666"
+     borderopacity="1"
+     objecttolerance="10"
+     gridtolerance="10"
+     guidetolerance="10"
+     inkscape:pageopacity="0"
+     inkscape:pageshadow="2"
+     inkscape:window-width="2560"
+     inkscape:window-height="1400"
+     id="namedview6"
+     showgrid="false"
+     inkscape:zoom="3.42"
+     inkscape:cx="69.444444"
+     inkscape:cy="150"
+     inkscape:window-x="0"
+     inkscape:window-y="40"
+     inkscape:window-maximized="0"
+     inkscape:current-layer="svg4" /><g
+     id="g852"><path
+       d="m 19.3408,34.258128 c -0.162508,-4.4e-4 -0.314692,0.04583 -0.448514,0.148541 L 0.52457424,48.433157 c -0.13860367,0.105527 -0.25984757,0.258831 -0.29900965,0.424402 -0.0130483,0.05519 0,0.111894 0,0.169761 0,0.231446 0.11420026,0.453466 0.29900965,0.594163 L 18.913644,63.669191 c 0.13453,0.102691 0.265721,0.14854 0.427156,0.14854 0.159311,0 0.33605,-0.04795 0.469872,-0.14854 0.269053,-0.201203 0.379495,-0.579601 0.256293,-0.891244 l -3.694899,-9.273185 h 21.656811 c 0.41421,0 0.747523,-0.331166 0.747523,-0.742704 v -7.448256 c 0,-0.411537 -0.333313,-0.763924 -0.747523,-0.763924 H 16.372066 l 3.694899,-9.294406 c 0.123909,-0.310939 0.01277,-0.647607 -0.256293,-0.848803 -0.13453,-0.10095 -0.307363,-0.148101 -0.469872,-0.148541 z"
+       id="path839"
+       style="fill:#2caf45;fill-opacity:1;stroke-width:1.22124" /><path
+       d="m 78.459201,34.130807 c -0.162511,4.41e-4 -0.335342,0.04758 -0.469872,0.148542 -0.269061,0.201195 -0.380203,0.537863 -0.256294,0.848803 l 3.6949,9.294406 H 59.771122 c -0.41421,0 -0.747524,0.352387 -0.747524,0.763924 v 7.448256 c 0,0.411538 0.333314,0.742702 0.747524,0.742702 h 21.656813 l -3.6949,9.273186 c -0.123198,0.311643 -0.01278,0.690042 0.256294,0.891245 0.13382,0.100597 0.31056,0.14854 0.469872,0.14854 0.161438,0 0.292625,-0.04586 0.427156,-0.14854 L 97.275426,49.494163 c 0.184798,-0.140698 0.299009,-0.362717 0.299009,-0.594164 0,-0.05786 0.01305,-0.114572 0,-0.16976 -0.03916,-0.165573 -0.160413,-0.31888 -0.299009,-0.424402 L 78.907715,34.279349 c -0.13382,-0.102716 -0.286003,-0.148981 -0.448514,-0.148542 z"
+       id="path837"
+       style="fill:#2caf45;fill-opacity:1;stroke-width:1.22124" /><path
+       d="m 43.902273,11.748602 c -0.781685,0 -1.409615,1.11798 -1.409615,2.509715 v 69.283364 c 0,1.391734 0.627222,2.510931 1.409615,2.509716 h 9.546938 c 0.781686,0 1.430973,-1.117982 1.430973,-2.509716 V 14.258317 c 0,-1.391735 -0.649287,-2.509715 -1.430973,-2.509715 z"
+       id="path2"
+       style="fill:#2271d5;fill-opacity:1;stroke-width:1.63481" /></g></svg>

src/calibre/gui2/tweak_book/editor/smarts/html.py

@@ -5,23 +5,27 @@
 __license__ = 'GPL v3'
 __copyright__ = '2014, Kovid Goyal <kovid at kovidgoyal.net>'
 
-import sys, re
+import re
-from operator import itemgetter
+import sys
-from itertools import chain
+from contextlib import contextmanager
-
 from css_parser import parseStyle
-from PyQt5.Qt import QTextEdit, Qt, QTextCursor
+from itertools import chain
+from operator import itemgetter
+from PyQt5.Qt import Qt, QTextCursor, QTextEdit
 
 from calibre import prepare_string_for_xml, xml_entity_to_unicode
-from calibre.ebooks.oeb.polish.container import OEB_DOCS
 from calibre.ebooks.oeb.base import css_text
+from calibre.ebooks.oeb.polish.container import OEB_DOCS
 from calibre.gui2 import error_dialog
-from calibre.gui2.tweak_book.editor.syntax.html import ATTR_NAME, ATTR_END, ATTR_START, ATTR_VALUE
+from calibre.gui2.tweak_book import current_container, tprefs
-from calibre.gui2.tweak_book import tprefs, current_container
 from calibre.gui2.tweak_book.editor.smarts import NullSmarts
 from calibre.gui2.tweak_book.editor.smarts.utils import (
-    no_modifiers, get_leading_whitespace_on_block, get_text_before_cursor,
+    expand_tabs, get_leading_whitespace_on_block, get_text_after_cursor,
-    get_text_after_cursor, smart_home, smart_backspace, smart_tab, expand_tabs)
+    get_text_before_cursor, no_modifiers, smart_backspace, smart_home, smart_tab
+)
+from calibre.gui2.tweak_book.editor.syntax.html import (
+    ATTR_END, ATTR_NAME, ATTR_START, ATTR_VALUE
+)
 from calibre.utils.icu import utf16_length
 from polyglot.builtins import unicode_type
 
@@ -214,24 +218,52 @@ def find_closing_tag(tag, max_tags=sys.maxsize):
 def select_tag(cursor, tag):
     cursor.setPosition(tag.start_block.position() + tag.start_offset)
     cursor.setPosition(tag.end_block.position() + tag.end_offset + 1, QTextCursor.MoveMode.KeepAnchor)
-    return unicode_type(cursor.selectedText()).replace(PARAGRAPH_SEPARATOR, '\n').rstrip('\0')
+    return cursor.selectedText().replace(PARAGRAPH_SEPARATOR, '\n').rstrip('\0')
+
+
+@contextmanager
+def edit_block(cursor):
+    cursor.beginEditBlock()
+    try:
+        yield
+    finally:
+        cursor.endEditBlock()
 
 
 def rename_tag(cursor, opening_tag, closing_tag, new_name, insert=False):
-    cursor.beginEditBlock()
+    with edit_block(cursor):
-    text = select_tag(cursor, closing_tag)
+        text = select_tag(cursor, closing_tag)
-    if insert:
+        if insert:
-        text = '</%s>%s' % (new_name, text)
+            text = '</%s>%s' % (new_name, text)
-    else:
+        else:
-        text = re.sub(r'^<\s*/\s*[a-zA-Z0-9]+', '</%s' % new_name, text)
+            text = re.sub(r'^<\s*/\s*[a-zA-Z0-9]+', '</%s' % new_name, text)
-    cursor.insertText(text)
+        cursor.insertText(text)
-    text = select_tag(cursor, opening_tag)
+        text = select_tag(cursor, opening_tag)
-    if insert:
+        if insert:
-        text += '<%s>' % new_name
+            text += '<%s>' % new_name
-    else:
+        else:
-        text = re.sub(r'^<\s*[a-zA-Z0-9]+', '<%s' % new_name, text)
+            text = re.sub(r'^<\s*[a-zA-Z0-9]+', '<%s' % new_name, text)
-    cursor.insertText(text)
+        cursor.insertText(text)
-    cursor.endEditBlock()
+
+
+def split_tag(cursor, opening_tag, closing_tag):
+    pos = cursor.position()
+    with edit_block(cursor):
+        open_text = select_tag(cursor, opening_tag)
+        open_text = re.sub(r'''\bid\s*=\s*['"].*?['"]''', '', open_text)
+        open_text = re.sub(r'\s+', ' ', open_text)
+        tag_name = re.search(r'<\s*(\S+)', open_text).group(1).lower()
+        is_block = tag_name in BLOCK_TAG_NAMES
+        prefix = ''
+        if is_block:
+            cursor.setPosition(cursor.anchor())
+            cursor.movePosition(QTextCursor.MoveOperation.StartOfLine, QTextCursor.MoveMode.KeepAnchor)
+            x = cursor.selectedText()
+            if x and not x.strip():
+                prefix = PARAGRAPH_SEPARATOR + x
+        close_text = select_tag(cursor, closing_tag)
+        cursor.setPosition(pos)
+        cursor.insertText(f'{close_text}{prefix}{open_text}')
 
 
 def ensure_not_within_tag_definition(cursor, forward=True):
@@ -397,9 +429,28 @@ class Smarts(NullSmarts):
                     ' before trying to rename tags.') % tag.name, show=True)
             rename_tag(c, tag, closing_tag, new_name, insert=tag.name in {'body', 'td', 'th', 'li'})
         else:
-            return error_dialog(editor, _('No found'), _(
+            return error_dialog(editor, _('No tag found'), _(
                 'No suitable block level tag was found to rename'), show=True)
 
+    def split_tag(self, editor):
+        editor.highlighter.join()
+        c = editor.textCursor()
+        block, offset = c.block(), c.positionInBlock()
+        tag, closing = find_tag_definition(block, offset)
+        if tag is not None:
+            return error_dialog(editor, _('Cursor inside tag'), _(
+                'Cannot split as the cursor is inside the tag definition'), show=True)
+        tag = find_closest_containing_tag(block, offset)
+        if tag is None:
+            return error_dialog(editor, _('No tag found'), _(
+                'No suitable tag was found to split'), show=True)
+        closing_tag = find_closing_tag(tag)
+        if closing_tag is None:
+            return error_dialog(editor, _('Invalid HTML'), _(
+                'There is an unclosed %s tag. You should run the Fix HTML tool'
+                ' before trying to split tags.') % tag.name, show=True)
+        split_tag(c, tag, closing_tag)
+
     def get_smart_selection(self, editor, update=True):
         editor.highlighter.join()
         cursor = editor.textCursor()

src/calibre/gui2/tweak_book/editor/text.py

@@ -900,6 +900,10 @@ version="1.1" width="100%%" height="100%%" viewBox="0 0 {w} {h}" preserveAspectR
         if hasattr(self.smarts, 'remove_tag'):
             self.smarts.remove_tag(self)
 
+    def split_tag(self):
+        if hasattr(self.smarts, 'split_tag'):
+            self.smarts.split_tag(self)
+
     def keyPressEvent(self, ev):
         if ev.key() == Qt.Key.Key_X and ev.modifiers() == Qt.KeyboardModifier.AltModifier:
             if self.replace_possible_unicode_sequence():

src/calibre/gui2/tweak_book/editor/widget.py

@@ -120,6 +120,9 @@ def register_text_editor_actions(_reg, palette):
     ac = reg('trash.png', _('Remove &tag'), ('remove_tag',), 'remove-tag', ('Ctrl+>'), _('Remove tag'), syntaxes=('html', 'xml'))
     ac.setToolTip(_('<h3>Remove tag</h3>Remove the currently highlighted tag'))
 
+    ac = reg('split.png', _('&Split tag'), ('split_tag',), 'split-tag', ('Ctrl+Alt+>'), _('Split current tag'), syntaxes=('html', 'xml'))
+    ac.setToolTip(_('<h3>Split tag</h3>Split the current tag at the cursor position'))
+
     editor_toolbar_actions['html']['fix-html-current'] = actions['fix-html-current']
     for s in ('xml', 'html', 'css'):
         editor_toolbar_actions[s]['pretty-current'] = actions['pretty-current']