Repository files navigation

Sumi is a simple OCR application with support for corrections.

• Go
• GTK+ 3.16 or later
• Tesseract 3.04.00 or later
• Trained data for your language
• One of the following:
  • X11
    • maim +slop
    • scrot
    • GraphicsMagick
    • ImageMagick
  • Windows
    • boxcutter

If none of the screenshot utilities above are available on your system, you canuse theSUMI_SCREENCAPTURE environment variable to provide your own. Theutility is expected to select a part of the screen and write to the file pathspecified in the last argument, e.g. forscrot the validSUMI_SCREENCAPTUREvalue would bescrot -s.

1. Download and install all dependencies from the list above
2. go get github.com/tsudoko/sumi

To use a language other than Japanese (or more than one language at once), passthe ISO 639-3 code of the desired language in a-l flag, i.e.sumi -l eng.Please note though thatsumi was designed to work specifically with Japanese,therefore it might give worse results when used with other languages.

Sumi prints scanned text tostdout. It's possible to send it to otherprograms automatically, examples below.

X11, requiresxclip.

./sumi | while read -r a; do echo "$a" | xclip -i -sel clip; done

Windows, requires ash-compatible shell andiconv. You have to replace$cpwith your locale's codepage, for Japanese it'scp932.

./sumi.exe | while read -r a; do echo "$a" | iconv -t $cp | clip; done

Withep:

./sumi | xargs -n1 ep

Withmyougiden:

./sumi | xargs -n1 myougiden

• tsuntsun
• Capture2Text
• KanjiTomo