Installing on linux issues

[wwardaww]

Hi, we cannot open the gui

Traceback (most recent call last):
  File "/media/arda/E64E3CD94E3CA3E91/Projeler/MomoTranslator/pyqt5_momotranslator.py", line 45, in <module>
    from PyQt6.QtCore import QPointF, QRectF, QSettings, QSize, QTranslator, Qt, pyqtSignal
ImportError: /home/arda/.local/lib/python3.10/site-packages/PyQt6/QtCore.abi3.so: undefined symbol: _ZN14QReadWriteLock16destroyRecursiveEP21QReadWriteLockPrivate, version Qt_6_PRIVATE_API

[alicewish]

What device and OS are you using?

[wwardaww]

Ubuntu 22.04 and desktop

[wwardaww]

and also im using python 3.10

[wwardaww]

hi i found it!! i install pyqt6==6.2.0 and then i un comment nltk but now i have another problem:

Traceback (most recent call last):
  File "/media/arda/E64E3CD94E3CA3E91/Projeler/MomoTranslator/pyqt5_momotranslator.py", line 9315, in <module>
    msyh_font20 = ImageFont.truetype(msyh_font_ttc, 20)
  File "/home/arda/.local/lib/python3.10/site-packages/PIL/ImageFont.py", line 797, in truetype
    return freetype(font)
  File "/home/arda/.local/lib/python3.10/site-packages/PIL/ImageFont.py", line 794, in freetype
    return FreeTypeFont(font, size, index, encoding, layout_engine)
  File "/home/arda/.local/lib/python3.10/site-packages/PIL/ImageFont.py", line 230, in __init__
    load_from_bytes(font)
  File "/home/arda/.local/lib/python3.10/site-packages/PIL/ImageFont.py", line 210, in load_from_bytes
    self.font_bytes = f.read()
AttributeError: 'NoneType' object has no attribute 'read'

[wwardaww]

Hi i fix the all issue but interface is chiniese is there a way to change it or should i translate it? i willopen a pr for fixes it's looks like when program cant found the font it's fails

[alicewish]

comment lang_code = 'zh_CN' uncomment # lang_code = 'en_US'

[alicewish]

Install font 微软雅黑 or change global_font_name and font_dic accordingly in your MomoTranslator_XXX_config.yml file

[wwardaww]

now i can access gui bu i cant process this image

console output

trace:

gui:

check_pts image:

[alicewish]

There is one bubble so there should only be one textblock but the check_pts image shows two textblocks, one with blue filling and one with white filling. Give me the whole name of the scan file, I'll use it to find the file and get a proper WLB_ext_px value.

[wwardaww]

what is scan file? iamge name is : RCO003_1663976825.jpg

[alicewish]

You can get the newest code and try again. Remember to download comictextdetector.pt.onnx if you hadn't.

Also I tried the image myself:

[alicewish]

If you are famaliar with Python then change config file to do_mode: 'do_dev_folder' The error message might be clearer without the GUI.

[wwardaww]

thx for the assist i will try tomorrow

[wwardaww]

Progress update

• i change pyqt6 version to pyqt6==6.5.0 in requirements.txt
• I install tsseract via apt-get
• I chage multimode, median, mean to statistics.multimode, statistics.median, statistics.mean since python thorws median module does not exist
• I installed pandoc via Github page:

AND ITS WORKING !!!

• when i try to open do_mode: 'do_dev_folder' it's gices me a new error:

[alicewish]

In the config file, when using do_mode: 'do_dev_folder', change folder_name to your actual folder name, also make sure there is a corresponding folder under ComicProcess folder with that name.

[alicewish]

Search for comictextdetector.pt.onnx in the code and you'll get the path. Or you can change the path in the code.

[alicewish]

In the config file, change target_lang: 'zh-CN' to your desired language, and make sure it is valid. See more at https://github.com/nidhaloff/deep-translator

The ChatGPT web translator is written for Mac automation and does not work on any other system.

[alicewish]

I forgot to put in from statistics import median, multimode, mean can be from statistics or from cv2 or from numpy, not much difference. Code updated on that.

[wwardaww]

I forgot to put in from statistics import median, multimode, mean can be from statistics or from cv2 or from numpy, not much difference. Code updated on that.

mean on cv2 expecting diffrent format (some numpy) so it's also must import from statistics