Size limitation for text, and sometimes prior OCR conversion in clipboard is not replaced with new OCR conversion about textshot

[sjscotti]

I found at least one of my problems. I created an anaconda environment for textshot and did the pip install -r requirements.txt. But in searching on the internet, it seem for anaconda I needed to install the win32api used by win10toast with conda install pywin32 and not pip. Before win10toast wasn't finding win32api. I'm still exploring whether this addresses the other points I brought up, but for now I'll close this issue.

[ianzhao05]

Hi, thanks for raising the issue. If you run it from the command line, it should print something that might be able to assist with troubleshooting. With large bodies of text, there is a 5 second timeout which should be enough for most use cases, but you can play with that value on line 92 of textshot.py to see if increasing it helps. Generally, a large text image should be fine.

That's interesting regarding conda + pip not installing win32api, thanks for testing that out!

[sjscotti]

Thanks for the speedy reply! Another thing that I have done which seems to improve my results is to set the --psm flag in the call on line 92 of textshot.py. So that line now reads...

            pil_img, timeout=5, lang=(sys.argv[1] if len(sys.argv) > 1 else None), config="--psm 6" 

At least that is what is working better for me..

[ianzhao05]

That's a good idea to change the PSM. I might add a command line option to make the config easier to customize.