TesseraPy provides a minimal wrapper for HP/Google's Tesseract OCR library, making use of CFFI instead of running the tesseract executable like other Python wrappers.

Currently (v0.1.4), it is extremely minimal having a class called tesseract with a single method get_text() which takes a PIL image as an argument and returns the text in the image.

So e.g.

import TesseraPy
from PIL import Image

i = Image.open("test.png")
t = TesseraPy.tesseract()
txt = t.get_text(i)
print(txt)

pip install git+https://github.com/owainkenwayucl/TesseraPy

To install a specific version, e.g. the first release, v0.1.4

pip install git+https://github.com/owainkenwayucl/[email protected]

You can control where TesseraPy looks for the various parts of Tesseract with the following environment variable:

• TESSERACT_LIBRARY - path to the Tesseract library - default: /lib64/libtesseract.so.5.3.4
• TESSERACT_DATA - path to Tesseract data files - default: /usr/share/tesseract/tessdata/
• TESSERACT_ENCODING - encoding - default: utf-8
• TESSERACT_LANGUAGE - language - default: eng

The defaults are those of the development platform - AlmaLinux 10.