OCR

[sejvlond]

For restaurants such as http://www.stravovanihrbek.websnadno.cz/jidelnicek.html would be nice to have some ocr to parse lunch menu :-)

[tomaskadlec]

I've made a quick research and it looks possible at least. Generally two approaches are possible:

• build a service providing OCR
• use existing service via its public API

The first one makes better sense in our case. The application was originally written to test our containerized infrastructure and deployment process (several diagrams will be published later this week). Adding complexity is somehow desirable.

[tomaskadlec]

Custom service using Tesseract, integration with PHP - http://www.sitepoint.com/ocr-in-php-read-text-from-images-with-tesseract/

[tomaskadlec]

A PHP wrapper for Tesseract https://github.com/thiagoalessio/tesseract-ocr-for-php

[tomaskadlec]

Example OCR services that can be used for the second approach:

• https://www.newocr.com/api/ (200 scans limit / ever?)