Adding _compute

mcq_eval.py

@@ -16,6 +16,9 @@
 import evaluate
 import datasets
 
+bleu = evaluate.load('bleu')
+#bert_score = evaluate.load('bertscore')
+
 
 # TODO: Add BibTeX citation
 _CITATION = """\
@@ -28,34 +31,30 @@ year={2020}
 
 # TODO: Add description of the module here
 _DESCRIPTION = """\
-This new module is designed to solve this great ML task and is crafted with a lot of care.
+This metric is designed to evaluate MCQ generations tasks.
 """
 
 
 # TODO: Add description of the arguments of the module here
 _KWARGS_DESCRIPTION = """
-Calculates how good are predictions given some references, using certain scores
+Calculates Accuracy and Blue-1 between generations and gold answers in a MCQ context.
 Args:
-    predictions: list of predictions to score. Each predictions
-        should be a string with tokens separated by spaces.
-    references: list of reference for each prediction. Each
-        reference should be a string with tokens separated by spaces.
+    generations: list of predictions to score. Each predictions
+        should be a string generated by a LM model.
+    golds: list of reference for each prediction. Each
+        reference should be a string only containing one letter (eg. A, B, C...).
 Returns:
-    accuracy: description of the first score,
-    another_score: description of the second score,
+    accuracy: ratio of good answers,
+    bleu-1: calculated by the module evaluate
 Examples:
-    Examples should be written in doctest format, and should illustrate how
-    to use the function.
+    Here is an exemple on how to use the metric:
 
-    >>> my_new_module = evaluate.load("my_new_module")
-    >>> results = my_new_module.compute(references=[0, 1], predictions=[0, 1])
+    >>> metric = evaluate.load("rfr2003/MQC_eval")
+    >>> results = metric.compute(references=["A", "B"], predictions=["A", "D"])
     >>> print(results)
-    {'accuracy': 1.0}
+    {'accuracy': 0.5, 'bleu-1': 0.5}
 """
 
-# TODO: Define external resources urls if needed
-BAD_WORDS_URL = "http://url/to/external/resource/bad_words.txt"
-
 
 @evaluate.utils.file_utils.add_start_docstrings(_DESCRIPTION, _KWARGS_DESCRIPTION)
 class MCQ_eval(evaluate.Metric):
@@ -71,14 +70,14 @@ class MCQ_eval(evaluate.Metric):
             inputs_description=_KWARGS_DESCRIPTION,
             # This defines the format of each prediction and reference
             features=datasets.Features({
-                'predictions': datasets.Value('int64'),
-                'references': datasets.Value('int64'),
+                'generations': datasets.Value('string'),
+                'golds': datasets.Value('string'),
             }),
             # Homepage of the module for documentation
-            homepage="http://module.homepage",
+            #homepage="http://module.homepage",
             # Additional links to the codebase or references
-            codebase_urls=["http://github.com/path/to/codebase/of/new_module"],
-            reference_urls=["http://path.to.reference.url/new_module"]
+            #codebase_urls=["http://github.com/path/to/codebase/of/new_module"],
+            #reference_urls=["http://path.to.reference.url/new_module"]
         )
 
     def _download_and_prepare(self, dl_manager):
@@ -86,10 +85,32 @@ class MCQ_eval(evaluate.Metric):
         # TODO: Download external resources if needed
         pass
 
-    def _compute(self, predictions, references):
+    def _compute(self, generations, golds):
         """Returns the scores"""
-        # TODO: Compute the different scores of the module
-        accuracy = sum(i == j for i, j in zip(predictions, references)) / len(predictions)
-        return {
-            "accuracy": accuracy,
-        }
+        assert len(generations) == len(golds)
+    
+        correct, total = 0, 0
+    
+        predictions, references = [], []
+    
+        for gen, gold in zip(generations, golds):
+            gen = gen.strip().upper()
+            gold = gold.upper()
+            if len(gen) > 1:
+                gen = gen[0]
+            if gen == gold:
+                correct += 1
+    
+            total += 1
+    
+            predictions.append(gen)
+            references.append(gold)
+            
+        metrics = {}
+        #metrics = {f"bert_score_{k}":np.mean(v).item() for k,v in bert_score.compute(predictions=predictions, references=references, lang="en").items() if k in ['recall', 'precision', 'f1']}
+        metrics.update({
+            'accuracy': correct/total,  
+            'bleu-1': bleu.compute(predictions=predictions, references=references, max_order=1)['bleu']
+        })
+    
+        return metrics

tests.py

@@ -1,17 +1,12 @@
 test_cases = [
     {
-        "predictions": [0, 0],
-        "references": [1, 1],
-        "result": {"metric_score": 0}
+        "predictions": ["A", "B"],
+        "references": ["A", "B"],
+        "result": {"accuracy": 1, 'bleu-1': 1}
     },
     {
-        "predictions": [1, 1],
-        "references": [1, 1],
-        "result": {"metric_score": 1}
-    },
-    {
-        "predictions": [1, 0],
-        "references": [1, 1],
-        "result": {"metric_score": 0.5}
+        "predictions": ["A", "B"],
+        "references": ["A", "C"],
+        "result": {"accuracy": 0.5, 'bleu-1': 0.5}
     }
 ]