Command Line Interface for the Tweak Book feature

src/calibre/debug.py

@@ -54,8 +54,14 @@ Run an embedded python interpreter.
     parser.add_option('-m', '--inspect-mobi', action='store_true',
             default=False,
             help='Inspect the MOBI file(s) at the specified path(s)')
-    parser.add_option('--tweak-kf8', default=None,
+    parser.add_option('--tweak-book', default=None,
-            help='Tweak the KF8 file at the specified path')
+            help='Tweak the book (exports the book as a collection of HTML '
+            'files and metadata, which you can edit using standard HTML '
+            'editing tools, and then rebuilds the file from the edited HTML. '
+            'Makes no additional changes to the HTML, unlike a full calibre '
+            'conversion). Note that this tool will try to open the '
+            'folder containing the HTML files in the editor pointed to by the'
+            ' EDITOR environment variable.')
 
     parser.add_option('--test-build', help='Test binary modules in build',
             action='store_true', default=False)
@@ -242,9 +248,9 @@ def main(args=sys.argv):
             prints('Inspecting:', path)
             inspect_mobi(path)
             print
-    elif opts.tweak_kf8:
+    elif opts.tweak_book:
-        from calibre.ebooks.mobi.tweak import tweak
+        from calibre.ebooks.tweak import tweak
-        tweak(opts.tweak_kf8)
+        tweak(opts.tweak_book)
     elif opts.test_build:
         from calibre.test_build import test
         test()

src/calibre/ebooks/mobi/reader/headers.py

@@ -234,6 +234,22 @@ class MetadataHeader(BookHeader):
         else:
             self.exth = None
 
+    @property
+    def kf8_type(self):
+        if (self.mobi_version == 8 and getattr(self, 'skelidx', NULL_INDEX) !=
+                NULL_INDEX):
+            return u'standalone'
+
+        kf8_header_index = getattr(self.exth, 'kf8_header', None)
+        if kf8_header_index is None:
+            return None
+        try:
+            if self.section_data(kf8_header_index-1) == b'BOUNDARY':
+                return u'joint'
+        except:
+            pass
+        return None
+
     def identity(self):
         self.stream.seek(60)
         ident = self.stream.read(8).upper()

src/calibre/ebooks/mobi/tweak.py

@@ -7,9 +7,9 @@ __license__   = 'GPL v3'
 __copyright__ = '2012, Kovid Goyal <kovid@kovidgoyal.net>'
 __docformat__ = 'restructuredtext en'
 
-import os, glob, sys, shlex, subprocess
+import os, glob
 
-from calibre import CurrentDir, as_unicode, prints
+from calibre import CurrentDir
 from calibre.ebooks.mobi import MobiError
 from calibre.ebooks.mobi.reader.mobi6 import MobiReader
 from calibre.ebooks.mobi.reader.headers import MetadataHeader
@@ -17,15 +17,14 @@ from calibre.utils.logging import default_log
 from calibre.ebooks import DRMError
 from calibre.ebooks.mobi.reader.mobi8 import Mobi8Reader
 from calibre.ebooks.conversion.plumber import Plumber, create_oebbook
-from calibre.ptempfile import TemporaryDirectory
-from calibre.constants import __appname__, iswindows
 from calibre.customize.ui import (plugin_for_input_format,
         plugin_for_output_format)
 
 class BadFormat(ValueError):
     pass
 
-def explode(stream, dest, split_callback=lambda x:True):
+def explode(path, dest, question=lambda x:True):
+    with open(path, 'rb') as stream:
         raw = stream.read(3)
         stream.seek(0)
         if raw == b'TPZ':
@@ -39,19 +38,22 @@ def explode(stream, dest, split_callback=lambda x:True):
         if header.encryption_type != 0:
             raise DRMError(_('This file is locked with DRM. It cannot be tweaked.'))
 
-    stream.seek(0)
+        kf8_type = header.kf8_type
-    mr = MobiReader(stream, default_log, None, None)
 
-    if mr.kf8_type is None:
+        if kf8_type is None:
             raise BadFormat('This MOBI file does not contain a KF8 format book')
 
-    if mr.kf8_type == 'joint':
+        if kf8_type == 'joint':
-        if not split_callback(_('This MOBI file contains both KF8 and '
+            if not question(_('This MOBI file contains both KF8 and '
                 'older Mobi6 data. Tweaking it will remove the Mobi6 data, which '
                 'means the file will not be usable on older Kindles. Are you '
                 'sure?')):
                 return None
 
+
+        stream.seek(0)
+        mr = MobiReader(stream, default_log, None, None)
+
         with CurrentDir(dest):
             mr = Mobi8Reader(mr, default_log)
             opf = os.path.abspath(mr())
@@ -72,52 +74,4 @@ def rebuild(src_dir, dest_path):
     oeb = create_oebbook(default_log, opf, plumber.opts)
     outp.convert(oeb, dest_path, inp, plumber.opts, default_log)
 
-def ask_question(msg):
-    prints(msg, end=' [y/N]: ')
-    sys.stdout.flush()
-
-    if iswindows:
-        import msvcrt
-        ans = msvcrt.getch()
-    else:
-        import tty, termios
-        old_settings = termios.tcgetattr(sys.stdin.fileno())
-        try:
-            tty.setraw(sys.stdin.fileno())
-            ans = sys.stdin.read(1)
-        finally:
-            termios.tcsetattr(sys.stdin.fileno(), termios.TCSADRAIN, old_settings)
-    print()
-    return ans == b'y'
-
-def tweak(mobi_file):
-    with TemporaryDirectory('_tweak_'+os.path.basename(mobi_file)) as tdir:
-        with open(mobi_file, 'rb') as stream:
-            try:
-                opf = explode(stream, tdir, split_callback=ask_question)
-            except BadFormat as e:
-                prints(as_unicode(e), file=sys.stderr)
-                raise SystemExit(1)
-        if opf is None:
-            return
-
-        ed = os.environ.get('EDITOR', None)
-        proceed = False
-        if ed is None:
-            prints('KF8 extracted to', tdir)
-            prints('Make your tweaks and once you are done,', __appname__,
-                    'will rebuild', mobi_file, 'from', tdir)
-            proceed = ask_question('Rebuild ' + mobi_file + '?')
-        else:
-            cmd = shlex.split(ed)
-            try:
-                subprocess.check_call(cmd + [tdir])
-            except:
-                prints(ed, 'failed, aborting...')
-                raise SystemExit(1)
-            proceed = True
-
-        if proceed:
-            rebuild(tdir, mobi_file)
-            prints(mobi_file, 'successfully tweaked')
 

src/calibre/ebooks/tweak.py

@@ -0,0 +1,126 @@
+#!/usr/bin/env python
+# vim:fileencoding=UTF-8:ts=4:sw=4:sta:et:sts=4:ai
+from __future__ import (unicode_literals, division, absolute_import,
+                        print_function)
+
+__license__   = 'GPL v3'
+__copyright__ = '2012, Kovid Goyal <kovid@kovidgoyal.net>'
+__docformat__ = 'restructuredtext en'
+
+import sys, os, shlex, subprocess
+
+from calibre import prints, as_unicode, walk
+from calibre.constants import iswindows, __appname__
+from calibre.ptempfile import TemporaryDirectory
+from calibre.libunzip import extract as zipextract
+from calibre.utils.zipfile import ZipFile, ZIP_DEFLATED, ZIP_STORED
+
+class Error(ValueError):
+    pass
+
+def ask_cli_question(msg):
+    prints(msg, end=' [y/N]: ')
+    sys.stdout.flush()
+
+    if iswindows:
+        import msvcrt
+        ans = msvcrt.getch()
+    else:
+        import tty, termios
+        old_settings = termios.tcgetattr(sys.stdin.fileno())
+        try:
+            tty.setraw(sys.stdin.fileno())
+            try:
+                ans = sys.stdin.read(1)
+            except KeyboardInterrupt:
+                ans = b''
+        finally:
+            termios.tcsetattr(sys.stdin.fileno(), termios.TCSADRAIN, old_settings)
+    print()
+    return ans == b'y'
+
+def mobi_exploder(path, tdir, question=lambda x:True):
+    from calibre.ebooks.mobi.tweak import explode, BadFormat
+    try:
+        return explode(path, tdir, question=question)
+    except BadFormat as e:
+        raise Error(as_unicode(e))
+
+def zip_exploder(path, tdir, question=lambda x:True):
+    zipextract(path, tdir)
+    for f in walk(tdir):
+        if f.lower().endswith('.opf'):
+            return f
+    raise Error('Invalid book: Could not find .opf')
+
+def zip_rebuilder(tdir, path):
+    with ZipFile(path, 'w', compression=ZIP_DEFLATED) as zf:
+        # Write mimetype
+        mt = os.path.join(tdir, 'mimetype')
+        if os.path.exists(mt):
+            zf.write(mt, 'mimetype', compress_type=ZIP_STORED)
+        # Write everything else
+        exclude_files = {'.DS_Store', 'mimetype', 'iTunesMetadata.plist'}
+        for root, dirs, files in os.walk(tdir):
+            for fn in files:
+                if fn in exclude_files:
+                    continue
+                absfn = os.path.join(root, fn)
+                zfn = os.path.relpath(absfn, tdir).replace(os.sep, '/')
+                zf.write(absfn, zfn)
+
+def get_tools(fmt):
+    fmt = fmt.lower()
+
+    if fmt in {'mobi', 'azw', 'azw3'}:
+        from calibre.ebooks.mobi.tweak import rebuild
+        ans = mobi_exploder, rebuild
+    elif fmt in {'epub', 'htmlz'}:
+        ans = zip_exploder, zip_rebuilder
+    else:
+        ans = None, None
+
+    return ans
+
+def tweak(ebook_file):
+    ''' Command line interface to the Tweak Book tool '''
+    fmt = ebook_file.rpartition('.')[-1].lower()
+    exploder, rebuilder = get_tools(fmt)
+    if exploder is None:
+        prints('Cannot tweak %s files. Supported formats are: EPUB, HTMLZ, AZW3, MOBI'
+                , file=sys.stderr)
+        raise SystemExit(1)
+
+    with TemporaryDirectory('_tweak_'+
+            os.path.basename(ebook_file).rpartition('.')[0]) as tdir:
+        try:
+            opf = exploder(ebook_file, tdir, question=ask_cli_question)
+        except Error as e:
+            prints(as_unicode(e), file=sys.stderr)
+            raise SystemExit(1)
+
+        if opf is None:
+            # The question was answered with No
+            return
+
+        ed = os.environ.get('EDITOR', None)
+        proceed = False
+        if ed is None:
+            prints('Book extracted to', tdir)
+            prints('Make your tweaks and once you are done,', __appname__,
+                    'will rebuild', ebook_file, 'from', tdir)
+            print()
+            proceed = ask_cli_question('Rebuild ' + ebook_file + '?')
+        else:
+            cmd = shlex.split(ed)
+            try:
+                subprocess.check_call(cmd + [tdir])
+            except:
+                prints(ed, 'failed, aborting...')
+                raise SystemExit(1)
+            proceed = True
+
+        if proceed:
+            rebuilder(tdir, ebook_file)
+            prints(ebook_file, 'successfully tweaked')
+