[GUI Plugin] WordDumb

[BeckyEbook]

I confirm, that in your files exists soft-hyphens – & #173;
These are not visible marks, but are easy to remove.
Just move the cursor over a known word (eg protected) and copy the invisible character that is between "pro" and "tected".
Replace this character to "empty/nothing" and you will solve your problem.

[xxyzz]

Originally Posted by jhowell

Perhaps the book uses soft hyphens and those are being treated as word terminators instead of being ignored.

Indeed, soft hyphens are all over the place. Thanks!

Originally Posted by AlaVander

This problem occurs in both AZW3 and KFX

Update to v3.7.5, this version ignores soft hyphens in the book.

[AlaVander]

Originally Posted by BeckyEbook

I confirm, that in your files exists soft-hyphens – & #173;
These are not visible marks, but are easy to remove.
Just move the cursor over a known word (eg protected) and copy the invisible character that is between "pro" and "tected".
Replace this character to "empty/nothing" and you will solve your problem.

Thank you. I did ignore soft hyphens because they don't appear as HTML entity symbols in Sigil. Why do you see them correctly? It should be noted that I have reserved soft hyphen as HTML entity symbol, but only when I input a new soft hyphen, it will be displayed as HTML entity symbol.

[AlaVander]

Originally Posted by xxyzz

Update to v3.7.5, this version ignores soft hyphens in the book.

Great update. I think soft hyphens are necessary for Mobi and azw3.

[BeckyEbook]

Originally Posted by AlaVander

Thank you. I did ignore soft hyphens because they don't appear as HTML entity symbols in Sigil. Why do you see them correctly? It should be noted that I have reserved soft hyphen as HTML entity symbol, but only when I input a new soft hyphen, it will be displayed as HTML entity symbol.

I used the ShadyCharacters plugin to display the softhyphens.

[DiapDealer]

You can also add the soft hyphen entity to your Preserve Entities list in Sigil's preferences. That way they'll always appear as entities vs invisible characters.

Nice to see my old Shady Characters plugin is still getting a little love, though.

[xxyzz]

v3.7.6: fix decode error and detect joint MOBI

Joint MOBI: I can't find a way to update ASIN to both versions, use https://github.com/kevinhendricks/KindleUnpack to split the book(-s option).

ASIN and Wi-Fi: for anyone still has interest in this and also has a rooted Kindle, check out my post at https://www.mobileread.com/forums/sh...d.php?t=339833. I tried to add the fat file system read-only attribute to the file but Kindle ignores it.

[Doitsu]

Originally Posted by xxyzz

Joint MOBI: I can't find a way to update ASIN to both versions, use https://github.com/kevinhendricks/KindleUnpack to split the book(-s option).

KevinH wrote a custom script for this: dualmetafix_mmap.py

[xxyzz]

Originally Posted by Doitsu

KevinH wrote a custom script for this: dualmetafix_mmap.py

Thanks for the information. I agree with Kovid, it's not worth the effort. And the most import thing is I don't understand the script(and calibre and KindleUnpack's code), not no mention these corner cases...

[smby]

I recieved the error message below. I am using 64 bit Calibre and 64 bit Python. Can you take a look please?

calibre, version 5.21.0 (win32, embedded-python: True)
Dumb error: Failed: Generating Word Wise for An American Sickness

Starting job: Generating Word Wise for An American Sickness
Job: "Generating Word Wise for An American Sickness" failed with error:
Traceback (most recent call last):
File "calibre\gui2\threaded_jobs.py", line 83, in start_work
File "calibre_plugins.worddumb.parse_job", line 32, in do_job
File "calibre_plugins.worddumb.parse_job", line 117, in find_named_entity
File "calibre_plugins.worddumb.x_ray", line 134, in search
File "calibre_plugins.worddumb.x_ray", line 104, in insert_entity
File "calibre_plugins.worddumb.x_ray", line 67, in search_wikipedia
KeyError: 'extract'

Called with args: ((4, 'MOBI', 'BBCFD059EV', 'C:\\Users\\Alan ML\\Calibre Library\\Elisabeth Rosenthal\\An American Sickness (4)\\An American Sickness - Elisabeth Rosenthal.mobi', <calibre.ebooks.metadata.book.base.Metadata object at 0x0000019BB75E1670>), True) {'notifications': <queue.Queue object at 0x0000019BB75E1DF0>, 'abort': <threading.Event object at 0x0000019BB75E1D90>, 'log': <calibre.utils.logging.GUILog object at 0x0000019BB75E1B80>}

[xxyzz]

Originally Posted by smby

KeyError: 'extract'

v3.7.7 fixes this error.

[smby]

Originally Posted by xxyzz

v3.7.7 fixes this error.

Thank you!

[steakhutzeee]

Hi! Found this plugin just now.

I have some books already converted to KFX that i'm reading.

What is the correct procedure? Using WordDumb will add Word Wise and X Ray to all formats?

So i just have to connect the Kindle and press WorDumb, correct? I actually have EPUB and KFX versions of all books.

So I just tried with a couple of books. Word wise seems to be working fine. X Ray only partially. When selecting a character name I do not get a description of it but instead I have only the times it is mentioned in the book.

Also I have a lot of terms mixed with the characters and vice versa.

[xxyzz]

Originally Posted by steakhutzeee

When selecting a character name I do not get a description of it but instead I have only the times it is mentioned in the book.

That sounds like a bug. Can you upload a screenshot and the book if it's in the public domain?

Originally Posted by steakhutzeee

Also I have a lot of terms mixed with the characters and vice versa.

X-ray tokens are generated by NLTK which uses machine learning to classify named entities, it can't match the quality of sold books.

[steakhutzeee]

Originally Posted by xxyzz

That sounds like a bug. Can you upload a screenshot and the book if it's in the public domain?



X-ray tokens are generated by NLTK which uses machine learning to classify named entities, it can't match the quality of sold books.

Hi, the book is not in the public domain, what can I do?