This directory contains all script relating to narrative matching. You can always obtain the arguments of a script using python script.py --help.
compute_alignment_performance.py computes the alignment performance of all configuration for a media pair and a similarity computation method (structural, textual of combined), while compute_all.sh use the previous script to compute performance for all media pairs and similarity computation methods. The results are saved as pickled dataframes under out/matching/plot.
plot_gold_alignment.py can be used to plot the reference alignment for a specific media pair.
All the python scripts starting with format_ are used to format results obtained with the compute_alignment_performance.py script.
plot_alignment.py can be used to visualize a specific alignment configuration (see --help for the numerous options).
alignment_commons.py, smith_waterman.py and graph_utils.py form a library of many useful functions for performing narrative matching, used by the other scripts.