This notebook provides a simple script to run the ocrevalUAtion tool (see: https://github.com/impactcentre/ocrevalUAtion) automatically through a batch of files. Whereas the outcome provided by the tool is the total number per category for the complete batch, this code provides an outcome on file level. The output is a .csv file, with for every file the 'CER', 'WER' and 'WER (order independent)' scores after comparison with the Ground Truth file.

This code is written for files with a filename in this format: "idenitifier_pagenr_type.extension"

Identifier: a unique code which is used in both the Ground Truth file and the OCR file
Pagenr: the pagenumber of the source that is captured in the file
Type: GT for Ground Truth, OCR for the OCR file to be evaluated
Extension: This can be any extensions that is allowed by the ocrevalUAtion tool. In this example .xml is used.