[GUI Plugin] WordDumb

[steakhutzeee]

Hi, I actually use kfx output plugin to convert my epub and send them to my kindle.

So using Worddumb I should first convert to kfx with kfx output and then use worddumb?

Inviato dal mio Mi 9T utilizzando Tapatalk

[TheWizardOf]

Does wordwise only work with English books?

[xxyzz]

Originally Posted by steakhutzeee

So using Worddumb I should first convert to kfx with kfx output and then use worddumb?

It also supports AZW3 and MOBI books.

[xxyzz]

Originally Posted by TheWizardOf

Does wordwise only work with English books?

That's right, Word Wise is for helping people read English books. But X-Ray supports more languages.

[xxyzz]

v3.11.1

- save disambiguation Wikipedia pages to cache
- install spaCy's dependencies for NLTK
- update dependencies
- update book ASIN inside ThreadedJob
- tell 32bit calibre Windows users to install 64bit calibre in error window

Some dependencies for X-Ray don't provide 32bit wheel file and I don't know how to build them. So 32bit calibre is not supported.

[xxyzz]

I write a script to parse Kindle English-Japanese dictionary to create a Japanese klld file. I can't upload the klld file but you can find the code at: https://github.com/xxyzz/WordDumb/tree/master/klld

I want to use xpath to extract definitions but didn't figure out how to write the correct expression.

[PoP]

^
Very nice!

[xxyzz]

v3.11.2

- check processed named entity's length and content
- update dependencies

I updated the klld code last week to support Oxford dictionaries. I only tested four Oxford dictionaries but other Oxford dictionaries could also be used as long as they have the same xml structure.

[xxyzz]

v3.12.0

- use spaCy's PhraseMatcher to find word wise phrases
- match lemma contains "'"
- use `continue` instead of `return` to check rest named entities when a token contains 'chapter'

v3.12.1
insert end column for word wise phrases

[cat0705]

Hello, recently I've just switched from Windows to Linux so I'm still pretty new with the systems. Somehow the plugins is not working even though I had python3-pip installed.

Code:

calibre, version 5.26.0 (linux, embedded-python: True)
Weak: subprocess.run() failed

Starting job: Generating Word Wise and X-Ray for _
Job: "Generating Word Wise and X-Ray for _" 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 24, in do_job
install_libs(model, create_ww)
File "calibre_plugins.worddumb.unzip", line 50, in install_libs
pip_install(pkg, value['version'], value['compiled'])
File "calibre_plugins.worddumb.unzip", line 104, in pip_install
subprocess.run(args, check=True, capture_output=True)
File "subprocess.py", line 512, in run
subprocess.CalledProcessError: Command '['python3', '-m', 'pip', 'install', '-t', PosixPath('/home/cat0705/.var/app/com.calibre_ebook.calibre/config/calibre/plugins/worddumb-libs/blis_0.7.4_3.8'), '--no-deps', '--python-version', '3.8', 'blis==0.7.4']' returned non-zero exit status 1.

Called with args: ((136, 'AZW3', 'B071HDXV91', '/home/cat0705/Calibre Library/_.azw3', <calibre.ebooks.metadata.book.base.Metadata object at 0x7f919ce4d3d0>, False, {'spacy': 'en_core_web_', 'wiki': 'en'}), True, True) {'notifications': <queue.Queue object at 0x7f919ce46970>, 'abort': <threading.Event object at 0x7f919fcf0430>, 'log': <calibre.utils.logging.GUILog object at 0x7f919cee0e50>}
/usr/bin/python3: No module named pip

I think it must be due to the way Linux dealing with pip package. Any ideas on how can I solve this? Thank you!

[xxyzz]

Originally Posted by cat0705

Somehow the plugins is not working even though I had python3-pip installed.

The error message says `/usr/bin/python3: No module named pip`. Which distro are you using? I'm pretty sure Ubuntu's python3 will have pip module once `python3-pip` package is installed.

[cat0705]

Originally Posted by xxyzz

The error message says `/usr/bin/python3: No module named pip`. Which distro are you using? I'm pretty sure Ubuntu's python3 will have pip module once `python3-pip` package is installed.

I'm using Zorin OS, which is based on Ubuntu. The pip is installed and located at `./.local/lib/python3.8/site-packages`.

I wonder if this is due to the “User Scheme” of Debian/Ubuntu. When I install a package through pip in terminal I will have to include 'pip install --user', otherwise it will return the message "Defaulting to user installation because normal site-packages is not writeable".

Sorry if this is confused, I still got a lot to learn with linux

[xxyzz]

Originally Posted by cat0705

I'm using Zorin OS, which is based on Ubuntu. The pip is installed and located at `./.local/lib/python3.8/site-packages`.

Don't download calibre from flathub, it's running in a sandbox and maybe that causes the error.

Run the command

sudo -v && wget -nv -O- https://download.calibre-ebook.com/linux-installer.sh | sudo sh /dev/stdin

to install calibre(see https://calibre-ebook.com/download_linux). All Linux users should use this command except Arch Linux users.

[cat0705]

Originally Posted by xxyzz

Don't download calibre from flathub, it's running in a sandbox and maybe that causes the error.

Run the command

sudo -v && wget -nv -O- https://download.calibre-ebook.com/linux-installer.sh | sudo sh /dev/stdin

to install calibre(see https://calibre-ebook.com/download_linux). All Linux users should use this command except Arch Linux users.

I reinstall calibre and it works perfectly now. Did not think flathub is the problem. Many thanks!

[xxyzz]

v3.13.0

- ignore "hapter*" while creating X-Ray
- add spaCy model link and tooltip to config label
- tell Debian based Linux users to install pip module and uninstall Flatpak calibre in error window
- use Flashtext to search lemmas and phrases, remove NLTK

Flashtext is way faster than regex, now word wise can be created instantly and don't need to install dependencies. So mac users no longer need to install python.