[GUI Plugin] Kobo Utilities

[isarl]

Originally Posted by decoherent

Hi, I'm having some trouble getting the ToC updated on my Kobo Libra H2O. The plugin worked great to add the series metadata, so thanks for that Something about the ToC update process makes everything fall over, though. When the plugin finishes scanning through the books on my reader, it (correctly) notes that all of them need ToC updates. However, after running a few, I'll get the following error thrown:

calibre, version 6.17.0
ERROR: Unhandled exception: <b>TypeError</b>:tuple indices must be integers or slices, not str

calibre 6.17  embedded-python: True
Windows-10-10.0.19045-SP0 Windows ('64bit', 'WindowsPE')
('Windows', '10', '10.0.19045')
Python 3.10.1
Windows: ('10', '10.0.19045', 'SP0', 'Multiprocessor Free')
Interface language: None
Successfully initialized third party plugins: Kobo Utilities (2, 16, 6)
Traceback (most recent call last):
  File "calibre_plugins.koboutilities.action", line 5390, in update_book_toc_on_device
  File "calibre_plugins.koboutilities.action", line 5818, in update_device_toc_for_books
  File "calibre_plugins.koboutilities.action", line 5833, in update_device_toc_for_book
  File "calibre_plugins.koboutilities.action", line 5864, in getDatabaseChapterId
TypeError: tuple indices must be integers or slices, not str

After this, I have to close calibre and disconnect/reconnect my reader before it'll do anything useful again (it just promptly throws the error otherwise). I've tried using a smaller subset of books, maybe 10 to 50, but sometimes it'll, sometimes it'll crash.

I have no idea what this means, is there a suggested fix?

I encountered problems recently while updating TOCs, but not this specific problem. In my case the issue was that the plugin seemed to be using disk space and I was running out of it when trying to process my entire library at once. I switched to processing a few books at a time (even 1 book at a time, for the largest books) and it ran without issues. Dunno if this will help you, but I hope so. Good luck!

[ownedbycats]

Originally Posted by chaley

No. The code was wrong. I don't see how it could have ever worked as it was unless the semantics of the python database calls changed.

What was wrong with it? I assume theres some specific circumstances that trigger it as I've been using update ToC for months and never come across any problems unless there's entries with duplicate names.

[chaley]

Originally Posted by ownedbycats

What was wrong with it? I've been using update ToC for months and never come across any problems unless there's entries with duplicate names.

This line in this function:

    def getDatabaseChapterId(self, bookId, toc_file, connection):
        cursor = connection.cursor()
        t = ("{0}%{1}%".format(bookId,toc_file),)
        cursor.execute('select ContentID from Content where ContentID like ?', t)
        try:
            result = next(cursor)
            chapterContentId = result['ContentID'] # <==============
        except StopIteration:
            chapterContentId = None

        debug_print('getDatabaseChapterId - chapterContentId=%s' % chapterContentId)
        cursor.close()
        return chapterContentId

The connection was built without a row factory so using the row like a named tuple or a dict can't work. It must be subscripted like a normal tuple.

The fix is

    def getDatabaseChapterId(self, bookId, toc_file, connection):
        cursor = connection.cursor()
        t = ("{0}%{1}%".format(bookId,toc_file),)
        cursor.execute('select ContentID from Content where ContentID like ?', t)
        try:
            result = next(cursor)
            chapterContentId = result[0]  #<===============
        except StopIteration:
            chapterContentId = None

        debug_print('getDatabaseChapterId - chapterContentId=%s' % chapterContentId)
        cursor.close()
        return chapterContentId

[decoherent]

Originally Posted by chaley

The attached beta version of the plugin *might* fix it. I don't use the plugin so I can't easily test it.

Well, I don't know what you fixed, but now it works perfectly, thanks!

Going back to my e-reader after using my tablet for a year or so has been annoying, as now I have to send the books, disconnect & let the device update, and then re-connect to send the series metadata and TOC, but, hey, whatever, at least it works! Thanks again!

[DNSB]

Originally Posted by decoherent

Well, I don't know what you fixed, but now it works perfectly, thanks!

Going back to my e-reader after using my tablet for a year or so has been annoying, as now I have to send the books, disconnect & let the device update, and then re-connect to send the series metadata and TOC, but, hey, whatever, at least it works! Thanks again!

You should not need to reconnect to send the ToC. That should be extracted from the epub metadata when the books is imported. You will still need the second connect to send the series data, subtitles, page and word counts, etc.

Using the NickelSeries addon allows you to avoid the second connection to send series and subtitle but does not help with other metadata.

[ownedbycats]

Oh, NickelSeries also handles subtitle? I should look into it.

[chaley]

Originally Posted by chaley

The attached beta version of the plugin *might* fix it. I don't use the plugin so I can't easily test it.

I have released this version.

[mbovenka]

Originally Posted by ownedbycats

Oh, NickelSeries also handles subtitle? I should look into it.

Yes, it does.

[Karellen]

I receive an error 403 which I reported in this thread... https://www.mobileread.com/forums/sh...21#post4336221

@DNSB believes it is a problem with the Kobo Utilities plugin

Attachment 202384

[ownedbycats]

You can disable it here until somebody can get it fixed.

[DNSB]

Originally Posted by Karellen

I receive an error 403 which I reported in this thread... https://www.mobileread.com/forums/sh...21#post4336221

@DNSB believes it is a problem with the Kobo Utilities plugin

Basically, when I noticed the error, I tried a couple of firmware updates from KoboUtililities dropdown menu. This triggered the error each time. Since Kobo had mentioned making changes which would force using a 4.35+ firmware version to continue using their website, I felt it was a pretty safe bet that this change was triggering the issue with KoboUtilities.

[Karellen]

Originally Posted by ownedbycats

You can disable it here until somebody can get it fixed.

Great. Thanks

and Thanks @DNSB

[ownedbycats]

Line 164 of action.py:

KOBO_FIRMWARE_UPDATE_CHECK_URL = "https://api.kobobooks.com/1.0/UpgradeCheck/Device/{0}/{1}/{2}/{3}"

If I can find an updated URL, I will see if I can fix this.

EDIT: I posted in the Kobo forums about it.

[JSWolf]

What URL does https://pgaskin.net/KoboStuff/kobofirmware.html use to deal with the firmware?

[ownedbycats]

Looks to be this: https://github.com/pgaskin/kfwproxy

Kobo API proxy for kobofirmware to enable cross-origin requests and caching. Each response is cached for 30 minutes to reduce the load on Kobo servers (otherwise there would be around half a million requests per day). In the future, kfwproxy will also support batched requests for better performance (currently, there are 112 requests per page load).

It might work. But I'm not too keen on using it without permission.