[GUI Plugin] WordDumb

[killabunny]

Originally Posted by xxyzz

This plugin uses Python launcher on Windows(the `py` command), it should be installed by default.

Users shouldn't set the Python binary path if Python is installed with default install methods listed in the document: https://xxyzz.github.io/WordDumb/installation.html

Maybe the Python path you're using is not a valid path? You could try to remove that path from the plugin settings.

Thank you. That progressed things somewhat. I installed the latest Python (3.12.4) now and left everything as it is. However, I am getting stuck on this step now:

"OSError: [WinError 127] The specified procedure could not be found. Error loading "C:\Users\dtale\AppData\Roaming\calibre\plugins\wo rddumb-libs-py3.12\torch\lib\cufftw64_11.dll" or one of its dependencies."

https://i.imgur.com/3181Pg7.jpeg

These are my WordDumb preferences:

https://i.imgur.com/cyn5Kj7.jpeg

[xxyzz]

You may need to downgrade cuda, pytrorch only supports cuda 12.1 and 11.8: https://pytorch.org/get-started/locally/

Or you could disable the "use GPU" option.

[killabunny]

Originally Posted by xxyzz

You may need to downgrade cuda, pytrorch only supports cuda 12.1 and 11.8: https://pytorch.org/get-started/locally/

Or you could disable the "use GPU" option.

Thank you so much. That did it. So grateful. BTW, is the quality the same WITH and WITHOUT using Cuda? If so, what's the purpose of having Cuda's option in there?

[xxyzz]

spaCy provides transformer models claim better NER performance so I added them, but now I'm kind regret doing this because most bug reports are CUDA dependencies issues.

[PoP]

Originally Posted by xxyzz

spaCy provides transformer models claim better NER performance so I added them, ...

FWIW, questionable.

[enhaohaotan]

I just mada an issue in github and you replied almost immediately, thank you very much! It did not solve my problem but from you reply I think it's not a bug so I change to continue my question here.

As you can see from the pictures in the attachments, all the danish words are not recognized correctly

The danish word drengen (the boy) is recognized to inquest,
The danish word stolte (proud) is recognized to bubbly,
The danish word nummber (number) is recognized to amplify,
All the words are wrong like this.

I did the following steps to add wordwise to the book:
1. book is local in calibre in azw3 format
2. Clicked Customize Kindle Word Wise in Customize WordDumb Preference
3. Set Book Language to Danish and clicked OK
4. Clicked Save in the popped out vocabulay list
5. Clicked OK in the Worddumb Preference
6. Clicked the star icon with kindle connected
7. Waited for jobs to finish
8. Unpluged kindle from computer and checked the book, word wise is not correct
9. Tried to change word wise language to Chinese and nothing changed but the word wise language was acutually changed to Chinese and were still wrong.

Thank you for you kind help! I'm learning danish and this function is super helpful!

[xxyzz]

You have to change the word wise language to Chinese in your Kindle's settings

If Kindle's default Chinese word wise database on Kindle is not replaced by Wiktionary database, connect kindle to calibre run the plugin again.

[enhaohaotan]

Originally Posted by xxyzz

You have to change the word wise language to Chinese in your Kindle's settings

If Kindle's default Chinese word wise database on Kindle is not replaced by Wiktionary database, connect kindle to calibre run the plugin again.

-----------------------------------------------------------

It's not working, the explaination was actually set to chinese. I tried a lot of times with the kindle plugged-in, with language set to Chinese, and clicked the icon many times

And I also tried english-english word wise with or withour wikitonary on the same book, I got exactly the same wrong explanation. Do you think it is because the book source? It is downloaded from z-library not from amazon shop

[xxyzz]

Check the word wise db file in Kindle's "/system/kll/" folder, file "kll.en.zh.klld" should be replaced by the plugin to Wiktionary data. Maybe you're using China Amazon account? Kindle uses a different file if login with China account.

[enhaohaotan]

Thank you!!!!! It was because of the account for this device is a chinese account, I changed it to a german account and it works! But still it marks some simple words but missed some difficult ones, is there any way to solve this?

I'll donate your work! It really helped a lot! Thank you!

[xxyzz]

Because I can't find a good data source of word difficulty for most languages, the default difficulty values are converted from the word frequency and all words are enabled. You can customize(and disable) the words difficulty value manually in the plugin settings window or edit the word wise sqlite database file.

I almost forget about the China account issue since Amazon exited China a few years ago, if you still have the old file, you can compare the files to see the words censored for Chinese users.

[explay]

Hi, I've been trying for several hours to figure it out by myself but maybe I'm too dumb or I'm doing something wrong. I can't figure out how to make wordwise work in russian books. I've tried all the combinations - trying different definition and book languages, trying with setting book language metadata to english or russian, tried it with different book formats and different books. The wordwise just simply doesn't work. Most of the times it doesn't show the "wordwise" in the bottom right of the kindle screen and even when it does the definitions aren't visible anywhere in the book. The x-ray is generated for russian and both x-ray and wordwise work perfectly for english books. I have tried both english and chinese setting for wordwise and both don't work.

[xxyzz]

I downloaded a Russian book from gutenberg.org and converted it to AZW3 because Kindle seems doesn't support MOBI now and the ru-en Word Wise works fine.

I don't know what went wrong for you, the description is too vague. I can only suggest you read the documents: https://xxyzz.github.io/WordDumb/installation.html and https://xxyzz.github.io/WordDumb/usage.html

Please follow the steps listed in documents and don't change the book language metadata.

[AliceK9]

So, my issue has been this, I've been trying on sending a pdf I converted to kfx through WordDumb to my kindle to access it's full features but so far nothing has been working for in terms of sending, it keeps giving me the same error while trying to convert it

[nnkhanh]

Originally Posted by xxyzz

calibre preference -> plugins -> search worddumb -> click 'customize plugin' -> disable x-ray -> click ok

Just one to come here and say thank you for the reply that you had 3 years ago. First my WordDumb couldn't find Python and I couldn't find anywhere to fix it. Come across your comment and it worked. Thank you!