If you came here from the paper

• "Comparing Free Reference Extraction Pipelines" by T. Backes, A. Iurshina, M. A. Shahid, P. Mayr

the code you will be interested is in

• code/M_evaluate_references.py

Of particular interest should be the functions

• assign() for the alignment of values of the same attribute
• similar_enough() for deciding matching of aligned values

The gold data is in

• resources/gold_references_...