Memory leak problem about pytrec_eval

[seanmacavaney]

This fix should take care of the problem! Copies of the qid and docno were not being freed properly.

Line #    Mem usage    Increment   Line Contents
================================================
    79   52.750 MiB   52.750 MiB   @profile
    80                             def runme():
    81                                 
    82   55.148 MiB    2.398 MiB       qrel = load_reference('qrels.txt')
    83  103.582 MiB   48.434 MiB       run = load_candidate('run.txt')
    84                             
    85  104.793 MiB    1.211 MiB       evaluator = pytrec_eval.RelevanceEvaluator(qrel, {'map', 'ndcg'})
    86                             
    87  104.793 MiB    0.000 MiB       N = 100
    88  123.664 MiB    0.000 MiB       for i in range(1,N):
    89  123.664 MiB   18.832 MiB           res = evaluator.evaluate(run)
    90  123.664 MiB    0.000 MiB           del res

[Ricocotam]

Great, I quit this tool for this reason, thanks for fixing it. Just a thing on your PR, it's big, author might not accept it directly. May be consider having several smaller PR for each issue ? Since they are not related

Thanks a lot for this work

[seanmacavaney]

@Ricocotam I agree in spirit (and it was originally the plan to have multiple PRs), but there's a lot of interdependence between the changes made (e.g., 3 of the 4 rely on the addition of a python wrapper around the extension). @cvangysel will this hold up the PR, and would it help speed up the process if I split them up? I could, but it would be a pain.