Added PyTDC

my_metric.py

@@ -3,7 +3,7 @@ import datasets
 import moses
 from moses import metrics
 import pandas as pd
-
+from tdc import Evaluator
 
 
 _DESCRIPTION = """
@@ -64,28 +64,38 @@ class my_metric(evaluate.Metric):
             features=datasets.Features(
                 {
                     "list_of_generated_smiles": datasets.Sequence(datasets.Value("string")),
+                    "list_of_test_smiles": datasets.Sequence(datasets.Value("string")),
                 }
                 if self.config_name == "multilabel"
                 else {
                     "list_of_generated_smiles": datasets.Value("string"),
+                    "list_of_test_smiles": datasets.Value("string"),
                 }
             ),
             reference_urls=["https://github.com/molecularsets/moses"],
         )
 
-    def _compute(self, list_of_generated_smiles):
-        # test_set = moses.get_dataset('test')
-        # preprocessed_smiles = [smile for smile in list_of_generated_smiles if moses.utils.canonicalize_smiles(smile)]
-        preprocessed_smiles = list_of_generated_smiles
-        results = metrics.get_all_metrics(preprocessed_smiles)
-
-        return {
-            "results": results
-            }
+    def _compute(self, list_of_generated_smiles, list_of_test_smiles=moses.get_dataset('test')):
+        
+        Results = metrics.get_all_metrics(gen = preprocessed_smiles, test = list_of_test_smiles)
+        
+        evaluator = Evaluator(name = 'Diversity')
+        Diversity = evaluator(list_of_generated_smiles)
+        
+        evaluator = Evaluator(name = 'KL_Divergence')
+        KL_Divergence = evaluator(list_of_generated_smiles, list_of_test_smiles)
+        
+        evaluator = Evaluator(name = 'FCD_Distance')
+        FCD_Distance = evaluator(list_of_generated_smiles, list_of_test_smiles)
+        
+        evaluator = Evaluator(name = 'Novelty')
+        Novelty = evaluator(list_of_generated_smiles, list_of_test_smiles)
+        
+        Results.update({
+            "PyTDC_Diversity": Diversity,
+            "PyTDC_KL_Divergence": KL_Divergence,
+            "PyTDC_FCD_Distance": FCD_Distance,
+            "PyTDC_Novelty": Novelty
+        })
 
-    # def _compute(self, predictions, references, normalize=True, sample_weight=None):
-    #     return {
-    #         "accuracy": float(
-    #             accuracy_score(references, predictions, normalize=normalize, sample_weight=sample_weight)
-    #         )
-    #     }
+        return {"results": Results}

requirements.txt

@@ -2,4 +2,5 @@ git+https://github.com/molecularsets/moses.git
 evaluate
 rdkit
 pandas==1.5.3
-gradio==3.43.1
+gradio==3.43.1
+PyTDC