Added oracles

my_metric.py

@@ -8,16 +8,18 @@ from tdc import Evaluator
 
 _DESCRIPTION = """
 Moses and PyTDC metrics
+Comprehensive suite of metrics designed to assess the performance of molecular generation models, for understanding how well a model can produce novel, chemically valid molecules that are relevant to specific research objectives.
+
 """
 
 
 _KWARGS_DESCRIPTION = """
 Args:
-    generated_smiles (`list` of `string`): Predicted labels.
-    train_smiles (`list` of `string`): test.
+    generated_smiles (`list` of `string`): A collection of SMILES (Simplified Molecular Input Line Entry System) strings generated by the model, ideally encompassing more than 30,000 samples.
+    train_smiles (`list` of `string`): The dataset of SMILES strings used to train the model, serving as a reference to evaluate the novelty and diversity of the generated molecules.
 
 Returns:
-    All moses metrics
+    Dectionary item containing various metrics to evaluate model performance
 """
 
 
@@ -70,7 +72,7 @@ class my_metric(evaluate.Metric):
                     "train_smiles": datasets.Value("string"),
                 }
             ),
-            reference_urls=["https://github.com/molecularsets/moses"],
+            reference_urls=["https://github.com/molecularsets/moses", "https://tdcommons.ai/functions/oracles/"],
         )
 
     def _compute(self, generated_smiles, train_smiles):
@@ -91,13 +93,52 @@ class my_metric(evaluate.Metric):
         
         evaluator = Evaluator(name = 'Validity')
         Validity = evaluator(generated_smiles)
+
+        oracle = Oracle(name = 'QED')
+        QED = oracle(generated_smiles)
+
+        oracle = Oracle(name = 'SA')
+        SA = oracle(generated_smiles)
+
+        oracle = Oracle(name = 'MPO')
+        MPO = oracle(generated_smiles)
+        MPO = {key: sum(values)/len(values) for key, values in MPO.items()}
+
+        oracle_list = [
+        'QED', 'SA', 'MPO', '3pbl_docking', 'GSK3B', 'JNK3',
+        'DRD2', 'LogP', 'Rediscovery', 'Similarity',
+        'Median', 'Isomers', 'Valsartan_SMARTS', 'Hop'
+    ]
+    
+    # Iterate through each oracle and compute its score
+    for oracle_name in oracle_list:
+        oracle = Oracle(name=oracle_name)
+        if oracle_name in ['Rediscovery', 'MPO', 'Similarity', 'Median', 'Isomers', 'Hop']:
+            # Assuming these oracles return a dictionary where values are lists of scores
+            score = oracle(generated_smiles)
+            if isinstance(score, dict):
+                # Convert lists of scores to average score for these specific metrics
+                score = {key: sum(values)/len(values) for key, values in score.items()}
+        else:
+            # Assuming other oracles return a list of scores
+            score = oracle(generated_smiles)
+            if isinstance(score, list):
+                # Convert list of scores to average score
+                score = sum(score) / len(score)
+        
+        Results.update({f"PyTDC_{oracle_name}": score})
+    
         
         Results.update({
             "PyTDC_Diversity": Diversity,
             "PyTDC_KL_Divergence": KL_Divergence,
             "PyTDC_FCD_Distance": FCD_Distance,
             "PyTDC_Novelty": Novelty,
-            "PyTDC_Validity": Validity
+            "PyTDC_Validity": Validity,
+
+            "PyTDC_QED": sum(QED)/len(QED),
+            "PyTDC_SA": sum(SA)/len(SA),
+            "PyTDC_MPO": MPO
         })
 
         return {"results": Results}