anthropic[patch]: tool output parser fix (#23647)

libs/partners/anthropic/langchain_anthropic/output_parsers.py

@@ -1,6 +1,6 @@
-from typing import Any, List, Optional, Type
+from typing import Any, List, Optional, Type, Union, cast
 
-from langchain_core.messages import ToolCall
+from langchain_core.messages import AIMessage, ToolCall
 from langchain_core.output_parsers import BaseGenerationOutputParser
 from langchain_core.outputs import ChatGeneration, Generation
 from langchain_core.pydantic_v1 import BaseModel
@@ -31,19 +31,18 @@ class ToolsOutputParser(BaseGenerationOutputParser):
         """
         if not result or not isinstance(result[0], ChatGeneration):
             return None if self.first_tool_only else []
-        message = result[0].message
+        message = cast(AIMessage, result[0].message)
-        if isinstance(message.content, str):
+        tool_calls: List = [
-            tool_calls: List = []
+            dict(tc) for tc in _extract_tool_calls_from_message(message)
-        else:
+        ]
-            content: List = message.content
+        if isinstance(message.content, list):
-            _tool_calls = [dict(tc) for tc in extract_tool_calls(content)]
             # Map tool call id to index
             id_to_index = {
                 block["id"]: i
-                for i, block in enumerate(content)
+                for i, block in enumerate(message.content)
-                if block["type"] == "tool_use"
+                if isinstance(block, dict) and block["type"] == "tool_use"
             }
-            tool_calls = [{**tc, "index": id_to_index[tc["id"]]} for tc in _tool_calls]
+            tool_calls = [{**tc, "index": id_to_index[tc["id"]]} for tc in tool_calls]
         if self.pydantic_schemas:
             tool_calls = [self._pydantic_parse(tc) for tc in tool_calls]
         elif self.args_only:
@@ -63,13 +62,25 @@ class ToolsOutputParser(BaseGenerationOutputParser):
         return cls_(**tool_call["args"])
 
 
-def extract_tool_calls(content: List[dict]) -> List[ToolCall]:
+def _extract_tool_calls_from_message(message: AIMessage) -> List[ToolCall]:
     """Extract tool calls from a list of content blocks."""
-    tool_calls = []
+    if message.tool_calls:
-    for block in content:
+        return message.tool_calls
-        if block["type"] != "tool_use":
+    return extract_tool_calls(message.content)
-            continue
+
-        tool_calls.append(
+
-            ToolCall(name=block["name"], args=block["input"], id=block["id"])
+def extract_tool_calls(content: Union[str, List[Union[str, dict]]]) -> List[ToolCall]:
-        )
+    """Extract tool calls from a list of content blocks."""
-    return tool_calls
+    if isinstance(content, list):
+        tool_calls = []
+        for block in content:
+            if isinstance(block, str):
+                continue
+            if block["type"] != "tool_use":
+                continue
+            tool_calls.append(
+                ToolCall(name=block["name"], args=block["input"], id=block["id"])
+            )
+        return tool_calls
+    else:
+        return []

libs/partners/anthropic/tests/unit_tests/test_output_parsers.py

@@ -70,3 +70,19 @@ def test_tools_output_parser_pydantic() -> None:
     expected = [_Foo1(bar=0), _Foo2(baz="a")]
     actual = output_parser.parse_result(_RESULT)
     assert expected == actual
+
+
+def test_tools_output_parser_empty_content() -> None:
+    class ChartType(BaseModel):
+        chart_type: Literal["pie", "line", "bar"]
+
+    output_parser = ToolsOutputParser(
+        first_tool_only=True, pydantic_schemas=[ChartType]
+    )
+    message = AIMessage(
+        "",
+        tool_calls=[{"name": "ChartType", "args": {"chart_type": "pie"}, "id": "foo"}],
+    )
+    actual = output_parser.invoke(message)
+    expected = ChartType(chart_type="pie")
+    assert expected == actual