Kindle driver: When generating page numbers automatically, add an additional method to detect page boundaries, using the presence of <mbp:pagebreak> tags in the source of the book. You can use this setting by cright clicking on the Kindle icon in calibre when the kindle is connected and choosing customize this device. Fixes #1321761 [Code Submission: Real Page numbers modification](https://bugs.launchpad.net/calibre/+bug/1321761)

Merge branch 'master' of https://github.com/user-none/calibre

src/calibre/devices/kindle/apnx.py

@@ -8,6 +8,7 @@ __docformat__ = 'restructuredtext en'
 Generates and writes an APNX page mapping file.
 '''
 
+import re
 import struct
 
 from calibre.ebooks.mobi.reader.mobi6 import MobiReader
@@ -22,15 +23,15 @@ class APNXBuilder(object):
     Create an APNX file using a pseudo page mapping.
     '''
 
-    def write_apnx(self, mobi_file_path, apnx_path, accurate=True, page_count=0):
+    def write_apnx(self, mobi_file_path, apnx_path, method=None, page_count=0):
         '''
         If you want a fixed number of pages (such as from a custom column) then
         pass in a value to page_count, otherwise a count will be estimated
         using either the fast or accurate algorithm.
         '''
         import uuid
-        apnx_meta = { 'guid': str(uuid.uuid4()).replace('-', '')[:8], 'asin':
+        apnx_meta = {'guid': str(uuid.uuid4()).replace('-', '')[:8], 'asin':
-                '', 'cdetype': 'EBOK', 'format': 'MOBI_7', 'acr': '' }
+                '', 'cdetype': 'EBOK', 'format': 'MOBI_7', 'acr': ''}
 
         with open(mobi_file_path, 'rb') as mf:
             ident = PdbHeaderReader(mf).identity()
@@ -60,17 +61,23 @@ class APNXBuilder(object):
         if page_count:
             pages = self.get_pages_exact(mobi_file_path, page_count)
         else:
-            if accurate:
+            try:
-                try:
+                if method == 'accurate':
                     pages = self.get_pages_accurate(mobi_file_path)
-                except:
+                elif method == 'pagebreak':
-                    # Fall back to the fast parser if we can't
+                    pages = self.get_pages_pagebreak_tag(mobi_file_path)
-                    # use the accurate one. Typically this is
+                    if not pages:
-                    # due to the file having DRM.
+                        pages = self.get_pages_accurate(mobi_file_path)
-                    pages = self.get_pages_fast(mobi_file_path)
+                else:
-            else:
+                    raise('no valid accurate method chosen use fast')
+            except:
+                # Fall back to the fast parser if we can't
+                # use the accurate one. Typically this is
+                # due to the file having DRM.
                 pages = self.get_pages_fast(mobi_file_path)
 
+        if not pages:
+            pages = self.get_pages_fast(mobi_file_path)
         if not pages:
             raise Exception(_('Could not generate page mapping.'))
 
@@ -94,9 +101,11 @@ class APNXBuilder(object):
 
         # Updated header if we have a KF8 file...
         if apnx_meta['format'] == 'MOBI_8':
-            content_header = '{"contentGuid":"%(guid)s","asin":"%(asin)s","cdeType":"%(cdetype)s","format":"%(format)s","fileRevisionId":"1","acr":"%(acr)s"}' % apnx_meta
+            content_header = '{"contentGuid":"%(guid)s","asin":"%(asin)s","cdeType":"%(cdetype)s","format":"%(format)s","fileRevisionId":"1","acr":"%(acr)s"}' % apnx_meta  # noqa
         else:
-            # My 5.1.x Touch & 3.4 K3 seem to handle the 'extended' header fine for legacy mobi files, too. But, since they still handle this one too, let's try not to break old devices, and keep using the simple header ;).
+            # My 5.1.x Touch & 3.4 K3 seem to handle the 'extended' header fine for
+            # legacy mobi files, too. But, since they still handle this one too, let's
+            # try not to break old devices, and keep using the simple header ;).
             content_header = '{"contentGuid":"%(guid)s","asin":"%(asin)s","cdeType":"%(cdetype)s","fileRevisionId":"1"}' % apnx_meta
         page_header = '{"asin":"%(asin)s","pageMap":"(1,a,1)"}' % apnx_meta
 
@@ -261,3 +270,23 @@ class APNXBuilder(object):
             pages.append(lines[i])
 
         return pages
+
+    def get_pages_pagebreak_tag(self, mobi_file_path):
+        '''
+        Determine pages based on the presense of
+        <mbp:pagebreak>.
+        '''
+        pages = []
+
+        # Get the MOBI html.
+        mr = MobiReader(mobi_file_path, default_log)
+        if mr.book_header.encryption_type != 0:
+            # DRMed book
+            return self.get_pages_fast(mobi_file_path)
+        mr.extract_text()
+
+        html = mr.mobi_html.lower()
+        for m in re.finditer('<[^>]*pagebreak[^>]*>', html):
+            pages.append(m.end())
+
+        return pages

src/calibre/devices/kindle/driver.py

@@ -311,6 +311,14 @@ class KINDLE2(KINDLE):
               'generator will produce pages that correspond better to a printed book. '
               'However, this method is slower and will slow down sending files '
               'to the Kindle.'),
+        _('Accurate calculation method') +
+            ':::' +
+            _('There are multiple methods to accurately calculate the page numbers. "accurate" which '
+              'is an estimation based on the number of chapters, paragraphs, and visible lines in the book. '
+              'This method is designed to simulate an average paperback book where there are 32 lines per '
+              'page and a maximum of 70 characters per line. \n\n'
+              'The "pagebreak" method uses the presense of <mbp:pagebreak> tags within the book to '
+              'determine pages.'),
         _('Custom column name to retrieve page counts from') +
             ':::' +
             _('If you have a custom column in your library that you use to '
@@ -322,13 +330,15 @@ class KINDLE2(KINDLE):
     EXTRA_CUSTOMIZATION_DEFAULT = [
         True,
         False,
+        'accurate',
         '',
     ]
-    OPT_APNX           = 0
+    OPT_APNX                 = 0
-    OPT_APNX_ACCURATE  = 1
+    OPT_APNX_ACCURATE        = 1
-    OPT_APNX_CUST_COL  = 2
+    OPT_APNX_ACCURATE_METHOD = 2
+    OPT_APNX_CUST_COL        = 3
     # x330 on the PaperWhite
-    THUMBNAIL_HEIGHT = 330
+    THUMBNAIL_HEIGHT         = 330
     # x262 on the Touch. Doesn't choke on x330, though.
 
     def formats_to_scan_for(self):
@@ -441,8 +451,11 @@ class KINDLE2(KINDLE):
         apnx_path = '%s.apnx' % os.path.join(path, filename)
         apnx_builder = APNXBuilder()
         try:
+            method = None
+            if opts.extra_customization[self.OPT_APNX_ACCURATE]:
+            	method = opts.extra_customization[self.OPT_APNX_ACCURATE_METHOD]
             apnx_builder.write_apnx(filepath, apnx_path,
-                                    accurate=opts.extra_customization[self.OPT_APNX_ACCURATE],
+                                    method=method,
                                     page_count=custom_page_count)
         except:
             print 'Failed to generate APNX'