[GUI Plugin] WordDumb

[xxyzz]

Originally Posted by Sodirow

Hello. I come in need of help.

WordDumb works unless I check the "Run spaCy with GPU” option. I followed the instructions at https://xxyzz.github.io/WordDumb/installation.html. The only thing I'm not sure if I did correctly is setting the PATH, because when I do I only get as a response.

I think you misunderstand the install document... Setting PATH is only for macOS users that choose not to install Python from Homebrew. And the error itself looks like a pip error, you could try to update pip or disable the "run spacy with gpu" option.

Command to update pip: `python3 -m pip install -U --user pip`

or `py -m pip install -U --user pip` on Windows.

[Sodirow]

Originally Posted by xxyzz

I think you misunderstand the install document... Setting PATH is only for macOS users that choose not to install Python from Homebrew. And the error itself looks like a pip error, you could try to update pip or disable the "run spacy with gpu" option.

Command to update pip: `python3 -m pip install -U --user pip`

or `py -m pip install -U --user pip` on Windows.

Thank you, xxyzz. But I ran that Windows command and I'm still getting the same error. It's fine. I'll have to disable the "Run spaCy with GPU” option.

[xxyzz]

Originally Posted by Sodirow

But I ran that Windows command and I'm still getting the same error.

WordDumb v3.29.4 uses the cupy-wheel 12.1.0.post1 package which fixes the pip install error.

GitHub release: https://github.com/xxyzz/WordDumb/releases/tag/v3.29.4

[Sodirow]

Originally Posted by xxyzz

WordDumb v3.29.4 uses the cupy-wheel 12.1.0.post1 package which fixes the pip install error.

GitHub release: https://github.com/xxyzz/WordDumb/releases/tag/v3.29.4

It works, now! Thanks!

[xxyzz]

v3.29.6

Fix "spacy_model" key error when saving customize lemmas
Add tooltip to use wiktionary definition checkbox label
Fix send wrong Kindle Word Wise database for non-English KFX books
Fix save customize lemmas error in calibre 6.26

Full Changelog: v3.29.5...v3.29.6

[Smokus]

I have a question regarding X-Ray.
I use the plugin via Calibre, and it works just fine.

After I fill out metadata entirely, I generate X-Ray and WordWise files and send the ebook (KFX format) onto my Kindle via the press on the WordDumb button on Calibre's toolbar at the top.

It works just fine, and X-Ray is available when I'm reading a book.
However I have noticed that some things aren't available in X-Ray even though there is an article on Fandom for them.

An example:
I am reading The Black Company at the moment. X-Ray and WordWise have been generated for the book via the plugin. In the book there are so called "forvalaka" creatures.
There is an entire fandom wiki for the book: https://blackcompany.fandom.com/wiki/Black_Company_Wiki
There is also an article about the "forvalaka" on fandom: https://blackcompany.fandom.com/wiki/Forvalaka

Since I have the following option toggled on in WordDumb settings, I thought that there would be an entry in X-Ray about "forvalaka" due to there being an article about them on fandom:


Is there something I'm missing, or does the plugin not work in that way?
I'd like to point out that my Kindle is always in the airplane mode if it matters.

[xxyzz]

By default X-Ray data are searched from Wikipedia. Fandom is used only if you input the Fandom URL for that book. And if spaCy model can't find a person's name, you could use the "customize X-Ray" feature to add the name manually. For more information, please read the document: https://xxyzz.github.io/WordDumb/usage.html

[Smokus]

Originally Posted by xxyzz

By default X-Ray data are searched from Wikipedia. Fandom is used only if you input the Fandom URL for that book. And if spaCy model can't find a person's name, you could use the "customize X-Ray" feature to add the name manually. For more information, please read the document: https://xxyzz.github.io/WordDumb/usage.html

Sorry, that was a mistake on my part. I have overlooked that part in the Docs.

So if I understand correctly, I will have to manually enter Fandom link for every book before generating X-Ray?

For example if I have 3 different books, the workflow is like this:
1. Enter manually Fandom URL of the 1st book.
2. Generate X-Ray & WordWise & send book to the Kindle
3. Enter manually Fandom URL of the 2nd book
4. Generate X-Ray & WordWise & send book to the Kindle
5. Enter manually Fandom URL of the 3rd book.
6. Generate X-Ray & WordWise & send book to the Kindle

Is there a way to dynamically search for relevant Fandom URL's? So that I can skip the steps 1, 3, 5 from above?

[xxyzz]

Originally Posted by Smokus

Is there a way to dynamically search for relevant Fandom URL's? So that I can skip the steps 1, 3, 5 from above?

You have to set the Fandom URL manually for each book unless some books use the same Fandom site.

[Smokus]

Originally Posted by xxyzz

You have to set the Fandom URL manually for each book unless some books use the same Fandom site.

Alright, thanks.

I managed to get the plugin to obtain data from Fandom, however I have a question.

As Fandom URL I have entered this: https://blackcompany.fandom.com/
I also selected Large spaCy model.

But there were no X-Ray entries for the creatures "forvalaka" even though there is a Fandom article about them: https://blackcompany.fandom.com/wiki/Forvalaka

Why is there no X-Ray for Forvalaka when clearly there is an article about them on Fandom? That's what I'm trying to figure out.

I have read the "Customize X-Ray" section in the Docs, but I don't understand it.

[iyner629]

This plugin is really awesome.
I have been using it smoothly together with Kindle's SendToKindle (STK) service, so I can benefit from Kindle Whispersync as well.

However, since recently, Kindle has been delivering KFX instead of azw3. And it breaks my normal operation.

By normal operation, I mean:
1 - Send my epub to Kindle with STK service.
2 - Deliver the ebook (KFX now) to my Kindle device.
3 - Connect my Kindle to my PC and launch Calibre
4 - In Calibre, switch to my device, select the ebook, right click to Add Book to Library.
5 - In Calibre, switch to my library, select the newly added ebook, Click the StarFish of WordDumb.

This does generate, e.g., LanguageLayer.en.BBG0ACA1HA.kll, and sends it to the corresponding .sdr directory on my kindle under documents/Download/Items01/

However, when opening the KFX on my kindle device, nothing happens.
It seems the kll file is being ignored.

It used to work when Kindle delivered azw3 before.

I can get the kll working by deleting all files inside the sdr directory, but that would lose Whispersync.

Also try to rename LanguageLayer.en.BBG0ACA1HA.kll to LanguageLayer.en.<strings_from_kindle_book>.kll to match the book name, but without luck.

Anyone met similar issue?

[xxyzz]

Originally Posted by Smokus

Why is there no X-Ray for Forvalaka when clearly there is an article about them on Fandom? That's what I'm trying to figure out.

I have read the "Customize X-Ray" section in the Docs, but I don't understand it.

I think spaCy models are trained to recognize normal human names not rare names in novels. Just add that name and description in customize X-Ray window and it'll be included in the X-Ray file.

[xxyzz]

Originally Posted by iyner629

However, when opening the KFX on my kindle device, nothing happens.
It seems the kll file is being ignored.

Should be fixed in https://github.com/xxyzz/WordDumb/co...8436bfe49ef29e

You can download the test plugin from https://github.com/xxyzz/WordDumb/ac...uns/6429259868

[Smokus]

Originally Posted by xxyzz

I think spaCy models are trained to recognize normal human names not rare names in novels. Just add that name and description in customize X-Ray window and it'll be included in the X-Ray file.

Alright thanks, will try it out.

[iyner629]

Hi xxyzz,

Thanks for the quick turnaround.

However, the testbuild does not seem to work either.

I removed worddumb from within Calibre, and deleted related files inside plugin directories as well. Then install the test build.

The generated kll file is still being ignored on my device.

Is there a way for the user to verify that the asin on the device is updated?

For example, 5AD8E5CE9A074215A58E020031557344.mf still contains the following id:
{"content":{"id":"5AD8E5CE9A074215A58E020031557344"