fix(anthropic): prevent crash with cache_control and empty message content (#34025)

libs/partners/anthropic/langchain_anthropic/chat_models.py

@@ -1008,27 +1008,27 @@ class ChatAnthropic(BaseChatModel):
 
         system, formatted_messages = _format_messages(messages)
 
-        # If cache_control is provided in kwargs, add it to last message
+        # If cache_control is provided in kwargs, add it to the last message with
-        # and content block.
+        # content (Anthropic requires cache_control to be nested within a message
-        if "cache_control" in kwargs and formatted_messages:
+        # block).
-            if isinstance(formatted_messages[-1]["content"], list):
+        cache_control = kwargs.pop("cache_control", None)
-                formatted_messages[-1]["content"][-1]["cache_control"] = kwargs.pop(
+        if cache_control and formatted_messages:
-                    "cache_control"
+            for formatted_message in reversed(formatted_messages):
-                )
+                content = formatted_message.get("content")
-            elif isinstance(formatted_messages[-1]["content"], str):
+                if isinstance(content, list) and content:
-                formatted_messages[-1]["content"] = [
+                    content[-1]["cache_control"] = cache_control
-                    {
+                    break
-                        "type": "text",
+                if isinstance(content, str):
-                        "text": formatted_messages[-1]["content"],
+                    formatted_message["content"] = [
-                        "cache_control": kwargs.pop("cache_control"),
+                        {
-                    }
+                            "type": "text",
-                ]
+                            "text": content,
-            else:
+                            "cache_control": cache_control,
-                pass
+                        }
-
+                    ]
-        # If cache_control remains in kwargs, it would be passed as a top-level param
+                    break
-        # to the API, but Anthropic expects it nested within a message
+            # If we didn't find a message with content we silently drop the control.
-        _ = kwargs.pop("cache_control", None)
+            # Anthropic would reject a payload with empty content blocks.
 
         payload = {
             "model": self.model,

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

@@ -1505,6 +1505,7 @@ def test_cache_control_kwarg() -> None:
             ],
         },
     ]
+    assert isinstance(messages[-1].content, str)  # test no mutation
 
     messages = [
         HumanMessage("foo"),
@@ -1528,6 +1529,23 @@ def test_cache_control_kwarg() -> None:
             ],
         },
     ]
+    assert "cache_control" not in messages[-1].content[-1]  # test no mutation
+
+
+def test_cache_control_kwarg_skips_empty_messages() -> None:
+    llm = ChatAnthropic(model=MODEL_NAME)
+
+    messages = [HumanMessage("foo"), AIMessage(content=[])]
+    payload = llm._get_request_payload(messages, cache_control={"type": "ephemeral"})
+    assert payload["messages"] == [
+        {
+            "role": "user",
+            "content": [
+                {"type": "text", "text": "foo", "cache_control": {"type": "ephemeral"}}
+            ],
+        },
+        {"role": "assistant", "content": []},
+    ]
 
 
 def test_context_management_in_payload() -> None: