Readme

ranking_evaluator.py

@@ -14,7 +14,7 @@ _CITATION = """\
 """
 
 _DESCRIPTION = """\
-A metric to evaluate ranking tasks using the TREC evaluation tool. It compares predicted rankings of items (e.g., documents) with their true relevance scores. The metric takes two inputs: references (true relevance scores) and predictions (predicted scores), both as lists of lists, where each (i, j) is the truth or the predicted score of the document j in the query i. In a nutshell: simplifies the usage of TREC to compute ranking metrics given scores per sample.
+A metric to evaluate ranking tasks using the TREC evaluation tool. It compares predicted rankings of items (e.g., documents) with their true relevance scores. The metric takes two inputs: references (true relevance scores) and predictions (predicted scores), both as lists of lists, where each (i, j) is the truth or the predicted score of the document j in the collection for the query i. In a nutshell: simplifies the usage of TREC to compute ranking metrics given scores per sample.
 """
 
 _KWARGS_DESCRIPTION = """ Computes MAP, P@K, RR, and NDCG using the TREC evaluation tool.
@@ -27,7 +27,7 @@ Returns:
     Dict: the set of TREC's metrics scores
 
 Example:
-    # (i, j) means the truth/predicted score of the document j in the query i
+    # (i, j) means the truth/predicted score of the document j in the collection for the query i
     references = [[5, 0, 3, 0, 0, 2, 1],
                     [5, 0, 3, 0, 0, 2, 1],
                     [5, 0, 3, 0, 0, 2, 1],