Strengthen local model evaluation checks

SovereignAI-Starter/tests/test_model_eval.py

@@ -0,0 +1,22 @@
+from __future__ import annotations
+
+import unittest
+
+from scripts.run_model_eval import run_check
+
+
+class ModelEvalCheckTests(unittest.TestCase):
+    def test_bullet_count_accepts_numbered_three_point_answer(self) -> None:
+        answer = "1. أول نقطة\n2. ثاني نقطة\n3. ثالث نقطة"
+
+        self.assertTrue(run_check({"type": "bullet_count", "value": 3}, answer))
+        self.assertFalse(run_check({"type": "bullet_count", "value": 2}, answer))
+
+    def test_single_sentence_rejects_multiple_sentences_and_lines(self) -> None:
+        self.assertTrue(run_check({"type": "single_sentence"}, "إجابة واحدة."))
+        self.assertFalse(run_check({"type": "single_sentence"}, "جملة أولى. جملة ثانية."))
+        self.assertFalse(run_check({"type": "single_sentence"}, "سطر أول\nسطر ثان."))
+
+
+if __name__ == "__main__":
+    unittest.main()