To run the code, start run.sh (the script contains/explains all necessary parameters).
Four csv files will be generated: one for the pseudo-qrels visualization, two for the unique contributions of each run and each group respectively to the depth-k pool and a last one for the overlap in retrieved relevant documents between runs.
Add the login/password to the TREC website in download.pl
The first CSV file contains 1 row per TREC run with its effectiveness with respect to the true relevance judgments and with respect to the pseudo-qrel judgments. Used here: http://www.st.ewi.tudelft.nl/~hauff/visualization/trecVis.html
The second CSV file contains 1 row per TREC run with the number of unique document contributions to the assessment pool (relevant as well as non-relevant).