Add new etric for calculating execution accuracy for text-to-sql llms.

execution_accuracy.py

@@ -17,18 +17,24 @@ import evaluate
 import datasets
 
 
+
 # TODO: Add BibTeX citation
 _CITATION = """\
-@InProceedings{huggingface:module,
-title = {A great new module},
-authors={huggingface, Inc.},
-year={2020}
+@misc{li2023llm,
+      title={Can LLM Already Serve as A Database Interface? A BIg Bench for Large-Scale Database Grounded Text-to-SQLs}, 
+      author={Jinyang Li and Binyuan Hui and Ge Qu and Jiaxi Yang and Binhua Li and Bowen Li and Bailin Wang and Bowen Qin and Rongyu Cao and Ruiying Geng and Nan Huo and Xuanhe Zhou and Chenhao Ma and Guoliang Li and Kevin C. C. Chang and Fei Huang and Reynold Cheng and Yongbin Li},
+      year={2023},
+      eprint={2305.03111},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL}
 }
 """
 
 # 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 new module is designed to calculate the EX, which is defined as the proportion of examples in the evaluation set for
+which the executed results of both the predicted and ground-truth SQLs are identical, relative to the
+overall number of SQLs.
 """
 
 
@@ -86,10 +92,32 @@ class ExecutionAccuracy(evaluate.Metric):
         # TODO: Download external resources if needed
         pass
 
-    def _compute(self, predictions, references):
+    def _compute(self, predictions, references, execute_func, filter_func=None):
         """Returns the scores"""
         # TODO: Compute the different scores of the module
-        accuracy = sum(i == j for i, j in zip(predictions, references)) / len(predictions)
+        
+        if len(predictions) != len(references):
+            raise ValueError("Predictions and references must have the same number of elements.")
+        
+        # Run filter_func on predictions and references if needed
+        filtered_predictions = []
+        filtered_references = []
+        divider = 0
+        if filter_func is not None:
+            for prediction, reference in zip(predictions, references):
+                # Only keep if both prediction and reference pass the filter
+                pred_bool = filter_func(prediction)
+                ref_bool = filter_func(reference)
+                if pred_bool and ref_bool:
+                    filtered_predictions.append(prediction)
+                    filtered_references.append(reference)
+                    divider += 1
+                # If only the reference passes the filter, count it
+                elif pred_bool != ref_bool:
+                    divider += 1
+                
+        accuracy = sum(execute_func(i) == execute_func(j) for i, j in zip(filtered_predictions, filtered_references)) / divider
+        
         return {
             "accuracy": accuracy,
         }