Upload evaluation.py

src/evaluation.py

@@ -0,0 +1,28 @@
+from src.model import compare_candidates
+
+def evaluate_model(openai, data):
+    correct_predictions = 0
+    count = 0
+
+    for index, row in data.iterrows():
+        count += 1
+        candidateA = {
+            'resume': row['candidateAResume'],
+            'transcript': row['candidateATranscript']
+        }
+        candidateB = {
+            'resume': row['candidateBResume'],
+            'transcript': row['candidateBTranscript']
+        }
+        role = row['role']
+        
+        prediction = compare_candidates(openai, candidateA, candidateB, role)
+        
+        if prediction:
+            if (prediction == 'Candidate A' and row['winnerId'] == row['candidateAId']) or \
+               (prediction == 'Candidate B' and row['winnerId'] == row['candidateBId']):
+                correct_predictions += 1
+        print('current_accuracy:', correct_predictions / count)
+
+    accuracy = correct_predictions / len(data)
+    return accuracy