From 12531734b59742d6b881aa3e4acb7e042b3c4a8e Mon Sep 17 00:00:00 2001
From: cl <2020334843@qq.com>
Date: Thu, 12 Mar 2026 10:23:42 +0800
Subject: [PATCH] Use ChangePrefixMatcher for torch.diagflat

---
 paconvert/api_mapping.json | 11 +----------
 tests/test_diagflat.py     | 25 +++++++++++++++++++++----
 2 files changed, 22 insertions(+), 14 deletions(-)

diff --git a/paconvert/api_mapping.json b/paconvert/api_mapping.json
index 75f94adda..f41725b35 100644
--- a/paconvert/api_mapping.json
+++ b/paconvert/api_mapping.json
@@ -4248,16 +4248,7 @@
     "Matcher": "ChangePrefixMatcher"
   },
   "torch.diagflat": {
-    "Matcher": "GenericMatcher",
-    "paddle_api": "paddle.diagflat",
-    "min_input_args": 1,
-    "args_list": [
-      "input",
-      "offset"
-    ],
-    "kwargs_change": {
-      "input": "x"
-    }
+    "Matcher": "ChangePrefixMatcher"
   },
   "torch.diagonal": {
     "Matcher": "ChangePrefixMatcher"
diff --git a/tests/test_diagflat.py b/tests/test_diagflat.py
index 8678d599b..008cb9de9 100644
--- a/tests/test_diagflat.py
+++ b/tests/test_diagflat.py
@@ -60,13 +60,19 @@ def test_case_4():
     pytorch_code = textwrap.dedent(
         """
         import torch
-        x = torch.tensor([[-0.4264, 0.0255,-0.1064],
-                        [ 0.8795,-0.2429, 0.1374],
-                        [ 0.1029,-0.6482,-1.6300]])
+        x = torch.tensor([1, 2, 3])
         result = torch.diagflat(input=x, offset=-3)
         """
     )
-    obj.run(pytorch_code, ["result"])
+    expect_paddle_code = textwrap.dedent(
+        """
+        import paddle
+
+        x = paddle.tensor([1, 2, 3])
+        result = paddle.diagflat(input=x, offset=-3)
+        """
+    )
+    obj.run(pytorch_code, expect_paddle_code=expect_paddle_code)
 
 
 # generated by validate_unittest autofix, based on test_case_4
@@ -81,3 +87,14 @@ def test_case_5():
         """
     )
     obj.run(pytorch_code, ["result"])
+
+
+def test_case_6():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor(7)
+        result = torch.diagflat(input=x)
+        """
+    )
+    obj.run(pytorch_code, ["result"])