Merge pull request #18049 from BerriAI/litellm_bedrock_ttl_issue

Remove ttl field when routing to bedrock

litellm/llms/bedrock/messages/invoke_transformations/anthropic_claude3_transformation.py

@@ -108,6 +108,27 @@ class AmazonAnthropicClaudeMessagesConfig(
             stream=stream,
         )
 
+    def _remove_ttl_from_cache_control(
+        self, anthropic_messages_request: Dict
+    ) -> None:
+        """
+        Remove `ttl` field from cache_control in messages.
+        Bedrock doesn't support the ttl field in cache_control.
+        
+        Args:
+            anthropic_messages_request: The request dictionary to modify in-place
+        """
+        if "messages" in anthropic_messages_request:
+            for message in anthropic_messages_request["messages"]:
+                if isinstance(message, dict) and "content" in message:
+                    content = message["content"]
+                    if isinstance(content, list):
+                        for item in content:
+                            if isinstance(item, dict) and "cache_control" in item:
+                                cache_control = item["cache_control"]
+                                if isinstance(cache_control, dict) and "ttl" in cache_control:
+                                    cache_control.pop("ttl", None)
+
     def transform_anthropic_messages_request(
         self,
         model: str,
@@ -141,8 +162,11 @@ class AmazonAnthropicClaudeMessagesConfig(
         # 3. `model` is not allowed in request body for bedrock invoke
         if "model" in anthropic_messages_request:
             anthropic_messages_request.pop("model", None)
+
+        # 4. Remove `ttl` field from cache_control in messages (Bedrock doesn't support it)
+        self._remove_ttl_from_cache_control(anthropic_messages_request)
             
-        # 4. AUTO-INJECT beta headers based on features used
+        # 5. AUTO-INJECT beta headers based on features used
         anthropic_model_info = AnthropicModelInfo()
         tools = anthropic_messages_optional_request_params.get("tools")
         messages_typed = cast(List[AllMessageValues], messages)

tests/test_litellm/llms/bedrock/messages/invoke_transformations/test_anthropic_claude3_transformation.py

@@ -79,3 +79,102 @@ def test_chunk_parser_usage_transformation():
     assert "usage" in parsed
     assert parsed["usage"]["input_tokens"] == 10
     assert parsed["usage"]["output_tokens"] == 5
+
+
+def test_remove_ttl_from_cache_control():
+    """Ensure ttl field is removed from cache_control in messages."""
+
+    cfg = AmazonAnthropicClaudeMessagesConfig()
+
+    # Test case 1: Message with cache_control containing ttl
+    request = {
+        "messages": [
+            {
+                "role": "user",
+                "content": [
+                    {
+                        "type": "text",
+                        "text": "Hello",
+                        "cache_control": {
+                            "type": "ephemeral",
+                            "ttl": "1h"
+                        }
+                    }
+                ]
+            }
+        ]
+    }
+
+    cfg._remove_ttl_from_cache_control(request)
+
+    # Verify ttl is removed but cache_control remains
+    assert "cache_control" in request["messages"][0]["content"][0]
+    assert "ttl" not in request["messages"][0]["content"][0]["cache_control"]
+    assert request["messages"][0]["content"][0]["cache_control"]["type"] == "ephemeral"
+
+    # Test case 2: Message with multiple content items
+    request2 = {
+        "messages": [
+            {
+                "role": "user",
+                "content": [
+                    {
+                        "type": "text",
+                        "text": "Hello",
+                        "cache_control": {
+                            "type": "ephemeral",
+                            "ttl": "1h"
+                        }
+                    },
+                    {
+                        "type": "text",
+                        "text": "World",
+                        "cache_control": {
+                            "type": "ephemeral",
+                            "ttl": "2h"
+                        }
+                    }
+                ]
+            }
+        ]
+    }
+
+    cfg._remove_ttl_from_cache_control(request2)
+
+    # Verify ttl is removed from all items
+    for item in request2["messages"][0]["content"]:
+        if "cache_control" in item:
+            assert "ttl" not in item["cache_control"]
+
+    # Test case 3: Message without ttl (should remain unchanged)
+    request3 = {
+        "messages": [
+            {
+                "role": "user",
+                "content": [
+                    {
+                        "type": "text",
+                        "text": "Hello",
+                        "cache_control": {
+                            "type": "ephemeral"
+                        }
+                    }
+                ]
+            }
+        ]
+    }
+
+    cfg._remove_ttl_from_cache_control(request3)
+
+    # Verify cache_control is unchanged
+    assert request3["messages"][0]["content"][0]["cache_control"]["type"] == "ephemeral"
+
+    # Test case 4: Empty messages (should not raise error)
+    request4 = {"messages": []}
+    cfg._remove_ttl_from_cache_control(request4)
+    assert request4 == {"messages": []}
+
+    # Test case 5: Request without messages key (should not raise error)
+    request5 = {}
+    cfg._remove_ttl_from_cache_control(request5)
+    assert request5 == {}