This package provide a custom Heroku buildpack providing the Tesseract OCR binary and all the required libraries to Heroku apps. Training data for English and Finnish language is provided.
The first step consists in allowing your Heroku app to use multiple buildpacks. Heroku natively supports multiple buildpacks per app.
-
setup your app as
heroku buildpacks:set heroku/LANG heroku buildpacks:add https://github.com/chanzuckerberg/heroku-buildpack-tesseract
where
LANG
is the language used by your app (e.g.,ruby
,python
, ornodejs
). A complete list of Heroku buildpacks can be found here. -
you can use the
tesseract
binary in your Heroku app! -
deploy :)
A minimal functioning Heroku app using this buildpack can be found here. The app is coded in Python and provides a REST method that accept an image and return the Tesseract OCR output as a JSON object. The REST functionality is implemented through the Flask web microframework.
The fork before this fork solves the issue of the missing libraries libtiff4
and libjpeg62
.
This fork includes the english training data so it doesn't require a remote download which can end up being a dead link.
MIT License.
Original work Copyright (c) 2013 Marco Azimonti
Modified work Copyright (c) 2015 Matteo Maggioni
Modified work Copyright (c) 2016 Yun-Fang Juan