A simple Elixir wrapper for the Tesseract OCR.
It's assumed that you have tesseract and all desired languages installed. For example, if you wanted to scan English (Tesseract's default language) and Japanese, you could get set up in a Debian-based system with:
$ sudo apt-get install tesseract-ocr tesseract-ocr-jpn
I didn't bother to put this in Hex since it's so simple. lib/tesseract.ex is really all there is.
$ curl https://raw.githubusercontent.com/bchase/tesseract-elixir/master/lib/tesseract.ex >> lib/tesseract.ex
We can scan the text from the image at test/support/reibun.png like so:
Tesseract.scan! "test/support/reibun.png" # defaults to English
# => "my: (flLMSi/v) (n) model sentence: (P):\n\n"
Tesseract.scan! "test/support/reibun.png", lang: :jpn
# => "例文 (れいぶん) (n) m0deー SentenCe; (P);\n\n"
Tesseract.scan! "test/support/reibun.png", lang: [:jpn, :eng]
# => "例文 (れいぶん) (n) model sentence: (P):\n\n"