Upload server/evaluator.py with huggingface_hub

server/evaluator.py

@@ -0,0 +1,20 @@
+from sklearn.linear_model import LogisticRegression
+from sklearn.model_selection import cross_val_score
+import numpy as np
+
+
+class Evaluator:
+    def __init__(self):
+        self.model = LogisticRegression(max_iter=200, random_state=42)
+
+    def evaluate(self, df):
+        """Returns accuracy score. Handles missing values by dropping rows."""
+        clean = df.dropna()
+        if len(clean) < 20:
+            return 0.0
+        X = clean.drop("label", axis=1).values
+        y = clean["label"].values
+        if len(set(y)) < 2:
+            return 0.0
+        scores = cross_val_score(self.model, X, y, cv=3, scoring="accuracy")
+        return float(np.mean(scores))