def get_results(trec_ress, metric):
results = []
for res in trec_ress:
rs = res.get_results_for_metric(metric).values()
m = np.mean(rs)
ci = misc.confidence_interval(rs, confidence=0.95)
n = res.get_runid()
results.append((n, m, ci))
return results
def get_coverage(trec_runs, trec_qrels, topX=10):
results = []
for r in trec_runs:
#n = r.get_runid()
n = r.run_data.get_value(0, "system")
covs = r.check_qrel_coverage(trec_qrels, topX=topX)
m = np.mean(covs)
ci = misc.confidence_interval(covs, confidence=0.95)
results.append((n, m, ci))
return results