[Bugfix][Frontend] Fix Gemma4 streaming HTML duplication after tool calls (#38909)

Signed-off-by: yoke233 <yoke2012@gmail.com>
2026-04-08 11:03:54 +08:00
parent 927975ead8
commit d734445fcd
2 changed files with 64 additions and 2 deletions
--- a/tests/tool_parsers/test_gemma4_tool_parser.py
+++ b/tests/tool_parsers/test_gemma4_tool_parser.py
@@ -531,3 +531,63 @@ class TestStreamingExtraction:
        assert "<|" not in args_text, (
            f"Partial delimiter leaked into JSON: {args_text!r}"
        )
+
+    def test_streaming_does_not_duplicate_plain_text_after_tool_call(
+        self, parser, mock_request, monkeypatch
+    ):
+        """Buffered plain text after a tool call must not corrupt current_text."""
+        captured_current_texts: list[str] = []
+        original_extract_streaming = parser._extract_streaming
+
+        def wrapped_extract_streaming(previous_text, current_text, delta_text):
+            captured_current_texts.append(current_text)
+            return original_extract_streaming(previous_text, current_text, delta_text)
+
+        monkeypatch.setattr(parser, "_extract_streaming", wrapped_extract_streaming)
+
+        chunks = [
+            "<|tool_call>",
+            "call:get_weather{",
+            'location:<|"|>Paris<|"|>}',
+            "<tool_call|><",
+            "div>",
+        ]
+
+        results = self._simulate_streaming(parser, mock_request, chunks)
+        content_parts = [
+            delta.content for delta, _ in results if delta is not None and delta.content
+        ]
+        assert "".join(content_parts) == "<div>"
+        assert captured_current_texts[-1].endswith("<tool_call|><div>")
+        assert not captured_current_texts[-1].endswith("<tool_call|><<div>")
+
+    def test_streaming_html_argument_does_not_duplicate_tag_prefixes(
+        self, parser, mock_request
+    ):
+        """HTML content inside tool arguments must not be duplicated."""
+        chunks = [
+            "<|tool_call>",
+            "call:write_file{",
+            'path:<|"|>index.html<|"|>,',
+            'content:<|"|><!DOCTYPE html>\n<',
+            'html lang="zh-CN">\n<',
+            "head>\n    <",
+            'meta charset="UTF-8">\n    <',
+            'meta name="viewport" content="width=device-width">\n',
+            '<|"|>}',
+            "<tool_call|>",
+        ]
+
+        results = self._simulate_streaming(parser, mock_request, chunks)
+        args_text = self._collect_arguments(results)
+        assert args_text
+
+        parsed_args = json.loads(args_text)
+        assert parsed_args["path"] == "index.html"
+        assert (
+            parsed_args["content"] == "<!DOCTYPE html>\n"
+            '<html lang="zh-CN">\n'
+            "<head>\n"
+            '    <meta charset="UTF-8">\n'
+            '    <meta name="viewport" content="width=device-width">\n'
+        )