[API Compatibility]enhance tests for torch._assert -part

paconvert/api_mapping.json

@@ -2768,12 +2768,7 @@
     "Matcher": "ChangePrefixMatcher"
   },
   "torch._assert": {
-    "Matcher": "AssertMatcher",
-    "min_input_args": 2,
-    "args_list": [
-      "condition",
-      "message"
-    ]
+    "Matcher": "ChangePrefixMatcher"
   },
   "torch._foreach_abs": {
     "Matcher": "ForeachMatcher",

tests/test__assert.py

@@ -71,3 +71,85 @@ def test_case_5():
         """
     )
     obj.run(pytorch_code)
+
+
+def test_case_6():
+    """Tensor condition"""
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([True])
+        torch._assert(x, "tensor should be true")
+        """
+    )
+    obj.run(pytorch_code)
+
+
+def test_case_7():
+    """Boolean literal condition"""
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        torch._assert(True, "literal true")
+        """
+    )
+    obj.run(pytorch_code)
+
+
+def test_case_8():
+    """Integer as condition (truthy)"""
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        torch._assert(1, "one is truthy")
+        """
+    )
+    obj.run(pytorch_code)
+
+
+def test_case_9():
+    """Empty string message"""
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        torch._assert(True, "")
+        """
+    )
+    obj.run(pytorch_code)
+
+
+def test_case_10():
+    """Variable as condition"""
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        cond = (2 > 1)
+        msg = "two is greater"
+        torch._assert(cond, msg)
+        """
+    )
+    obj.run(pytorch_code)
+
+
+def test_case_11():
+    """Comparison expression with len()"""
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        data = [1, 2, 3]
+        torch._assert(len(data) == 3, "length mismatch")
+        """
+    )
+    obj.run(pytorch_code)
+
+
+def test_case_12():
+    """Tensor comparison condition"""
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor(5.0)
+        torch._assert(x > 0, "x should be positive")
+        """
+    )
+    obj.run(pytorch_code)