Kobo driver: Automatically unkepubify books when exporting from the device

Also, allow adding the virtual books since nowadays some of them are DRM
free. DRMed books will still give an error when attempting to add them
to the library.

src/calibre/devices/kobo/driver.py

@@ -26,11 +26,12 @@ from calibre.devices.kobo.books import Book, ImageWrapper, KTCollectionsBookList
 from calibre.devices.mime import mime_type_ext
 from calibre.devices.usbms.books import BookList, CollectionsBookList
 from calibre.devices.usbms.driver import USBMS
+from calibre.ebooks import DRMError
 from calibre.ebooks.metadata import authors_to_string
 from calibre.ebooks.metadata.book.base import Metadata
 from calibre.ebooks.metadata.utils import normalize_languages
 from calibre.prints import debug_print
-from calibre.ptempfile import PersistentTemporaryFile, better_mktemp
+from calibre.ptempfile import PersistentTemporaryFile, TemporaryDirectory, better_mktemp
 from calibre.utils.config_base import prefs
 from calibre.utils.date import parse_date
 from polyglot.builtins import iteritems, itervalues, string_or_bytes
@@ -115,8 +116,7 @@ class KOBO(USBMS):
     SUPPORTS_SUB_DIRS = True
     SUPPORTS_ANNOTATIONS = True
 
-    # "kepubs" do not have an extension. The name looks like a GUID. Using an empty string seems to work.
+    VIRTUAL_BOOK_EXTENSIONS = frozenset(('kobo',))
-    VIRTUAL_BOOK_EXTENSIONS = frozenset(('kobo', ''))
 
     EXTRA_CUSTOMIZATION_MESSAGE = [
         _('The Kobo supports several collections including ')+ 'Read, Closed, Im_Reading. ' + _(
@@ -773,7 +773,7 @@ class KOBO(USBMS):
             # Supported database version
             return True
 
-    def get_file(self, path, *args, **kwargs):
+    def get_file(self, path, outfile, end_session=True):
         tpath = self.munge_path(path)
         extension = os.path.splitext(tpath)[1]
         if extension == '.kobo':
@@ -783,8 +783,20 @@ class KOBO(USBMS):
                         'instead they are rows in the sqlite database. '
                     'Currently they cannot be exported or viewed.'),
                     UserFeedback.WARN)
+        if tpath.lower().endswith(KEPUB_EXT + EPUB_EXT):
+            with TemporaryDirectory() as tdir:
+                outpath = os.path.join(tdir, 'file.epub')
+                from calibre.ebooks.oeb.polish.kepubify import unkepubify_path
+                try:
+                    unkepubify_path(path, outpath, allow_overwrite=True)
+                except DRMError:
+                    pass
+                else:
+                    with open(outpath, 'rb') as src:
+                        shutil.copyfile(src, outfile)
+                    return
 
-        return USBMS.get_file(self, path, *args, **kwargs)
+        return USBMS.get_file(self, path, outfile, end_session=end_session)
 
     @classmethod
     def book_from_path(cls, prefix, lpath, title, authors, mime, date, ContentType, ImageID):
@@ -1125,17 +1137,23 @@ class KOBO(USBMS):
         with no file extension.  I just hope that decision causes
         them as much grief as it does me :-)
 
-        This has to make a temporary copy of the book files with a
+        This has to make a temporary copy of the book files with an
         epub extension to allow calibre's normal processing to
         deal with the file appropriately
         '''
         for idx, path in enumerate(paths):
-            if path.find('kepub') >= 0:
+            parts = path.replace(os.sep, '/').split('/')
-                with closing(open(path, 'rb')) as r:
+            if path.lower().endswith(KEPUB_EXT + EPUB_EXT) or ('kepub' in parts and '.' not in parts[-1]):
-                    tf = PersistentTemporaryFile(suffix='.epub')
+                with PersistentTemporaryFile(suffix=EPUB_EXT) as dest:
-                    shutil.copyfileobj(r, tf)
+                    pass
-                    # tf.write(r.read())
+                from calibre.ebooks.oeb.polish.kepubify import unkepubify_path
-                    paths[idx] = tf.name
+                try:
+                    unkepubify_path(path, dest.name, allow_overwrite=True)
+                except DRMError as e:
+                    import traceback
+                    paths[idx] = (path, e, traceback.format_exc())
+                else:
+                    paths[idx] = dest.name
         return paths
 
     @classmethod
@@ -2332,7 +2350,6 @@ class KOBOTOUCH(KOBO):
         return result
 
     def _kepubify(self, path, name, mi, extra_css) -> None:
-        from calibre.ebooks.oeb.polish.errors import DRMError
         from calibre.ebooks.oeb.polish.kepubify import kepubify_path, make_options
         debug_print(f'Starting conversion of {mi.title} ({name}) to kepub')
         opts = make_options(

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

@@ -28,6 +28,7 @@ from calibre.ebooks.oeb.base import OEB_DOCS, OEB_STYLES, XHTML, XPath, escape_c
 from calibre.ebooks.oeb.parse_utils import barename, merge_multiple_html_heads_and_bodies
 from calibre.ebooks.oeb.polish.container import Container, EpubContainer, get_container
 from calibre.ebooks.oeb.polish.cover import find_cover_image, find_cover_image3, find_cover_page
+from calibre.ebooks.oeb.polish.errors import DRMError
 from calibre.ebooks.oeb.polish.parsing import parse
 from calibre.ebooks.oeb.polish.tts import lang_for_elem
 from calibre.ebooks.oeb.polish.utils import extract, insert_self_closing
@@ -516,8 +517,24 @@ def kepubify_path(path, outpath='', max_workers=0, allow_overwrite=False, opts:
     return outpath
 
 
+def check_for_kobo_drm(container: Container) -> None:
+    # sadly rights.xml is not definitive as various dedrm tools leave it behind
+    has_rights_xml = container.has_name_and_is_not_empty('rights.xml')
+    if not has_rights_xml:
+        return
+    for name, is_linear in container.spine_names:
+        mt = container.mime_map[name]
+        if mt in OEB_DOCS:
+            with container.open(name, 'rb') as f:
+                raw = f.read(8192)
+                if b'<?xml' not in raw and b'<html' not in raw and KOBO_SPAN_CLASS.encode() not in raw:
+                    raise DRMError()
+            break
+
+
 def unkepubify_path(path, outpath='', max_workers=0, allow_overwrite=False):
     container = get_container(path, tweak_mode=True, ebook_cls=EpubContainer)
+    check_for_kobo_drm(container)
     unkepubify_container(container, max_workers)
     base, ext = os.path.splitext(path)
     outpath = outpath or base + '.epub'