Added CLI support search_replace by changing serialization and deserialization of option.

src/calibre/ebooks/conversion/plumber.py

@@ -628,7 +628,17 @@ OptionRecommendation(name='sr3_replace',
 
 OptionRecommendation(name='search_replace',
     recommended_value='[]', level=OptionRecommendation.LOW,
-    help=_('Modify the document text and structure using user defined patterns.')),
+    help=_('Modify the document text and structure using user defined patterns.'
+           'This option accepts parameters in two forms:\n'
+           '1.file:<path to search/replace definitions file>\n'
+           'The file should contain alternating lines or search/replace strings:\n'
+           '      <search>\n'
+           '      <replace>\n'
+           '      <search>\n'
+           '      <replace>\n'
+           'Files saved through the user interface dialog can be used with this option.\n'
+           '2.json:<json encoded list containg [search, replace] touples:\n'
+           '      json:[["search","replace"],["search","replace"]]\n')),
 ]
         # }}}
 

src/calibre/ebooks/conversion/preprocess.py

@@ -5,7 +5,7 @@ __license__   = 'GPL v3'
 __copyright__ = '2009, Kovid Goyal <kovid@kovidgoyal.net>'
 __docformat__ = 'restructuredtext en'
 
-import functools, re, json
+import functools, re, search_replace_option
 
 from calibre import entity_to_unicode, as_unicode
 
@@ -515,8 +515,8 @@ class HTMLPreProcessor(object):
         if not getattr(self.extra_opts, 'keep_ligatures', False):
             html = _ligpat.sub(lambda m:LIGATURES[m.group()], html)
 
-        search_replace = json.loads(getattr(self.extra_opts, 'search_replace', '[]'))
+        # Function for processing search and replace
-        for search_pattern, replace_txt in search_replace:
+        def do_search_replace(search_pattern, replace_txt):
             if search_pattern:
                 try:
                     search_re = re.compile(search_pattern)
@@ -528,6 +528,17 @@ class HTMLPreProcessor(object):
                     self.log.error('Failed to parse %r regexp because %s' %
                             (search, as_unicode(e)))
 
+        #search / replace using the sr?_search / sr?_replace options
+        for search, replace in [['sr3_search', 'sr3_replace'], ['sr2_search', 'sr2_replace'], ['sr1_search', 'sr1_replace']]:
+            search_pattern = getattr(self.extra_opts, search, '')
+            replace_txt = getattr(self.extra_opts, replace, '')
+            do_search_replace(search_pattern, replace_txt)
+
+        # multi-search / replace using the search_replace option
+        search_replace = search_replace_option.decode(getattr(self.extra_opts, 'search_replace', '[]'))
+        for search_pattern, replace_txt in search_replace:
+            do_search_replace(search_pattern, replace_txt)
+            
         end_rules = []
         # delete soft hyphens - moved here so it's executed after header/footer removal
         if is_pdftohtml:

src/calibre/ebooks/conversion/search_replace_option.py

@@ -0,0 +1,50 @@
+#!/usr/bin/env python
+# vim:fileencoding=UTF-8:ts=4:sw=4:sta:et:sts=4:ai
+
+__license__   = 'GPL v3'
+__copyright__ = '2012, Eli Algranti <idea00@hotmail.com>'
+__docformat__ = 'restructuredtext en'
+
+import json
+from itertools import izip
+
+def encodeJson(definition):
+    '''
+    Encode a search/replace definition using json.
+    '''
+    return 'json:' + json.dumps(definition)
+
+def encodeFile(definition, filename):
+    '''
+    Encode a search/replace definition into a file
+    '''
+    with open(filename, 'w') as f:
+        for search,replace in definition:
+            f.write(search + '\n')
+            f.write(replace + '\n')
+
+    return 'file:'+filename
+
+
+def decode(definition):
+    '''
+    Decodes a search/replace definition
+    '''
+    if definition.startswith('json:'):
+        return json.loads(definition[len('json:'):])
+    elif definition.startswith('file:'):
+        with open(definition[len('file:'):], 'r') as f:
+            ans = []
+            for search, replace in izip(f, f):
+                ans.append([search.rstrip('\n\r'), replace.rstrip('\n\r')])
+        return ans
+    raise Exception('Invalid definition')
+
+            
+            
+
+
+
+
+    
+

src/calibre/gui2/convert/search_and_replace.py

@@ -1,13 +1,14 @@
 # -*- coding: utf-8 -*-
 
 __license__ = 'GPL 3'
-__copyright__ = '2011, John Schember <john@nachtimwald.com>'
+__copyright__ = '2011, John Schember <john@nachtimwald.com>, 2012 Eli Algranti <idea00@hotmail.com>'
 __docformat__ = 'restructuredtext en'
 
-import re, json
+import re
+from calibre.ebooks.conversion import search_replace_option
 
 from PyQt4.QtCore import SIGNAL, Qt
-from PyQt4.QtGui import QTableWidget, QTableWidgetItem, QFileDialog
+from PyQt4.QtGui import QTableWidget, QTableWidgetItem, QFileDialog, QMessageBox
 from calibre.gui2.convert.search_and_replace_ui import Ui_Form
 from calibre.gui2.convert import Widget
 from calibre.gui2 import error_dialog
@@ -41,7 +42,7 @@ class SearchAndReplaceWidget(Widget, Ui_Form):
         self.search_replace.setColumnCount(2)
         self.search_replace.setColumnWidth(0, 300)
         self.search_replace.setColumnWidth(1, 300)
-        self.search_replace.setHorizontalHeaderLabels(['Search Expression', 'Replacement'])
+        self.search_replace.setHorizontalHeaderLabels([_('Search Regular Expression'), _('Replacement Text')])
 
         self.connect(self.sr_add, SIGNAL('clicked()'), self.sr_add_clicked)
         self.connect(self.sr_change, SIGNAL('clicked()'), self.sr_change_clicked)
@@ -82,18 +83,14 @@ class SearchAndReplaceWidget(Widget, Ui_Form):
             self.search_replace.setCurrentCell(row-1, 0)
 
     def sr_load_clicked(self):
-        filename = QFileDialog.getOpenFileName(self, 'Load Calibre Search-Replace definitions file', '.', 'Calibre Search-Replace definitions file (*.csr)')
+        filename = QFileDialog.getOpenFileName(self, _('Load Calibre Search-Replace definitions file'), '.', _('Calibre Search-Replace definitions file (*.csr)'))
         if filename:
-            with open(filename, 'r') as f:
+            self.set_value_handler(self.opt_search_replace, 'file:'+unicode(filename))
-                val = f.read()
-                self.set_value(self.search_replace, val)
             
     def sr_save_clicked(self):
-        filename = QFileDialog.getSaveFileName(self, 'Save Calibre Search-Replace definitions file', '.', 'Calibre Search-Replace definitions file (*.csr)')
+        filename = QFileDialog.getSaveFileName(self, _('Save Calibre Search-Replace definitions file'), '.', _('Calibre Search-Replace definitions file (*.csr)'))
         if filename:
-            with open(filename, 'w') as f:
+            search_replace_option.encodeFile(self.get_definitions(), unicode(filename))
-                val = self.get_value(self.search_replace)
-                f.write(val)
         
     def sr_currentCellChanged(self, row, column, previousRow, previousColumn) :
         if row >= 0:
@@ -122,14 +119,40 @@ class SearchAndReplaceWidget(Widget, Ui_Form):
         self.sr_search.set_doc(doc)
 
     def pre_commit_check(self):
-        for row in xrange(0, self.search_replace.rowCount()):
+
+
+        definitions = self.get_definitions()
+
+        # Verify the search/replace in the edit widgets has been
+        # included to the list of search/replace definitions
+
+        edit_search = self.sr_search.regex
+
+        if edit_search:
+            edit_replace = unicode(self.sr_replace.text())
+            found = False
+            for search, replace in definitions:
+                if search == edit_search and replace == edit_replace:
+                    found = True
+                    break
+            if not found:
+                msgBox = QMessageBox(self)
+                msgBox.setText(_('The search / replace definition being edited has not been added to the list of definitions'))
+                msgBox.setInformativeText(_('Do you wish to continue with the conversion (the definition will not be used)?'))
+                msgBox.setStandardButtons(QMessageBox.Yes | QMessageBox.No)
+                msgBox.setDefaultButton(QMessageBox.No)
+                if msgBox.exec_() != QMessageBox.Yes:
+                    return False
+
+        # Verify all search expressions are valid        
+        for search, replace in definitions:
             try:
-                pat = unicode(self.search_replace.item(row,0).text())
+                re.compile(search)
-                re.compile(pat)
             except Exception as err:
                 error_dialog(self, _('Invalid regular expression'),
                              _('Invalid regular expression: %s')%err, show=True)
                 return False
+        
         return True
 
     # Options
@@ -171,14 +194,16 @@ class SearchAndReplaceWidget(Widget, Ui_Form):
     def get_value_handler(self, g):
         if g != self.opt_search_replace:
             return None
-        
+        return search_replace_option.encodeJson(self.get_definitions())
+
+    def get_definitions(self):
         ans = []
         for row in xrange(0, self.search_replace.rowCount()):
             colItems = []
             for col in xrange(0, self.search_replace.columnCount()):
                 colItems.append(unicode(self.search_replace.item(row, col).text()))
             ans.append(colItems)
-        return json.dumps(ans)
+        return ans
 
     def set_value_handler(self, g, val):
         if g != self.opt_search_replace:
@@ -186,7 +211,7 @@ class SearchAndReplaceWidget(Widget, Ui_Form):
             return True
 
         try:
-            rowItems = json.loads(val)
+            rowItems = search_replace_option.decode(val)
             if not isinstance(rowItems, list):
                 rowItems = []
         except:

src/calibre/gui2/convert/search_and_replace.ui

@@ -32,7 +32,7 @@
       </sizepolicy>
      </property>
      <property name="title">
-      <string>First expression</string>
+      <string>Search/Replace Definition Edit</string>
      </property>
      <layout class="QGridLayout" name="gridLayout_2">
       <property name="sizeConstraint">
@@ -147,7 +147,7 @@
    <item row="0" column="0">
     <widget class="QLabel" name="label">
      <property name="text">
-      <string>&lt;p&gt;Search and replace uses &lt;i&gt;regular expressions&lt;/i&gt;. See the &lt;a href=&quot;http://manual.calibre-ebook.com/regexp.html&quot;&gt;regular expressions tutorial&lt;/a&gt; to get started with regular expressions. Also clicking the wizard buttons below will allow you to test your regular expression against the current input document.</string>
+      <string>&lt;p&gt;Search and replace uses &lt;i&gt;regular expressions&lt;/i&gt;. See the &lt;a href=&quot;http://manual.calibre-ebook.com/regexp.html&quot;&gt;regular expressions tutorial&lt;/a&gt; to get started with regular expressions. Also clicking the wizard button below will allow you to test your regular expression against the current input document.</string>
      </property>
      <property name="wordWrap">
       <bool>true</bool>