Handle Groq tool_use_failed errors by telling model to retry as usual

Author: DouweM

pydantic_ai_slim/pydantic_ai/models/groq.py

@@ -7,6 +7,7 @@
 from datetime import datetime
 from typing import Any, Literal, cast, overload
 
+from pydantic import BaseModel, Json, ValidationError
 from typing_extensions import assert_never
 
 from .. import ModelHTTPError, UnexpectedModelBehavior, _utils, usage
@@ -48,7 +49,7 @@
 )
 
 try:
-    from groq import NOT_GIVEN, APIStatusError, AsyncGroq, AsyncStream
+    from groq import NOT_GIVEN, APIError, APIStatusError, AsyncGroq, AsyncStream
     from groq.types import chat
     from groq.types.chat.chat_completion_content_part_image_param import ImageURL
 except ImportError as _import_error:
@@ -169,9 +170,22 @@ async def request(
         model_request_parameters: ModelRequestParameters,
     ) -> ModelResponse:
         check_allow_model_requests()
-        response = await self._completions_create(
-            messages, False, cast(GroqModelSettings, model_settings or {}), model_request_parameters
-        )
+        try:
+            response = await self._completions_create(
+                messages, False, cast(GroqModelSettings, model_settings or {}), model_request_parameters
+            )
+        except ModelHTTPError as e:
+            if isinstance(e.body, dict):  # pragma: no branch
+                try:
+                    error = _GroqToolUseFailedError.model_validate(e.body)  # pyright: ignore[reportUnknownMemberType]
+                    tool_call_part = ToolCallPart(
+                        tool_name=error.error.failed_generation.name,
+                        args=error.error.failed_generation.arguments,
+                    )
+                    return ModelResponse(parts=[tool_call_part])
+                except ValidationError:
+                    pass
+            raise
         model_response = self._process_response(response)
         return model_response
 
@@ -449,36 +463,50 @@ class GroqStreamedResponse(StreamedResponse):
     _provider_name: str
 
     async def _get_event_iterator(self) -> AsyncIterator[ModelResponseStreamEvent]:
-        async for chunk in self._response:
-            self._usage += _map_usage(chunk)
-
-            try:
-                choice = chunk.choices[0]
-            except IndexError:
-                continue
-
-            # Handle the text part of the response
-            content = choice.delta.content
-            if content is not None:
-                maybe_event = self._parts_manager.handle_text_delta(
-                    vendor_part_id='content',
-                    content=content,
-                    thinking_tags=self._model_profile.thinking_tags,
-                    ignore_leading_whitespace=self._model_profile.ignore_streamed_leading_whitespace,
-                )
-                if maybe_event is not None:  # pragma: no branch
-                    yield maybe_event
-
-            # Handle the tool calls
-            for dtc in choice.delta.tool_calls or []:
-                maybe_event = self._parts_manager.handle_tool_call_delta(
-                    vendor_part_id=dtc.index,
-                    tool_name=dtc.function and dtc.function.name,
-                    args=dtc.function and dtc.function.arguments,
-                    tool_call_id=dtc.id,
-                )
-                if maybe_event is not None:
-                    yield maybe_event
+        try:
+            async for chunk in self._response:
+                self._usage += _map_usage(chunk)
+
+                try:
+                    choice = chunk.choices[0]
+                except IndexError:
+                    continue
+
+                # Handle the text part of the response
+                content = choice.delta.content
+                if content is not None:
+                    maybe_event = self._parts_manager.handle_text_delta(
+                        vendor_part_id='content',
+                        content=content,
+                        thinking_tags=self._model_profile.thinking_tags,
+                        ignore_leading_whitespace=self._model_profile.ignore_streamed_leading_whitespace,
+                    )
+                    if maybe_event is not None:  # pragma: no branch
+                        yield maybe_event
+
+                # Handle the tool calls
+                for dtc in choice.delta.tool_calls or []:
+                    maybe_event = self._parts_manager.handle_tool_call_delta(
+                        vendor_part_id=dtc.index,
+                        tool_name=dtc.function and dtc.function.name,
+                        args=dtc.function and dtc.function.arguments,
+                        tool_call_id=dtc.id,
+                    )
+                    if maybe_event is not None:
+                        yield maybe_event
+        except APIError as e:
+            if isinstance(e.body, dict):  # pragma: no branch
+                try:
+                    error = _GroqToolUseFailedInnerError.model_validate(e.body)  # pyright: ignore[reportUnknownMemberType]
+                    yield self._parts_manager.handle_tool_call_part(
+                        vendor_part_id='tool_use_failed',
+                        tool_name=error.failed_generation.name,
+                        args=error.failed_generation.arguments,
+                    )
+                    return
+                except ValidationError as e:
+                    pass
+            raise
 
     @property
     def model_name(self) -> GroqModelName:
@@ -510,3 +538,28 @@ def _map_usage(completion: chat.ChatCompletionChunk | chat.ChatCompletion) -> us
         input_tokens=response_usage.prompt_tokens,
         output_tokens=response_usage.completion_tokens,
     )
+
+
+class _GroqToolUseFailedGeneration(BaseModel):
+    name: str
+    arguments: dict[str, Any]
+
+
+class _GroqToolUseFailedInnerError(BaseModel):
+    message: str
+    type: Literal['invalid_request_error']
+    code: Literal['tool_use_failed']
+    failed_generation: Json[_GroqToolUseFailedGeneration]
+
+
+class _GroqToolUseFailedError(BaseModel):
+    # {
+    #     'error': {
+    #         'message': "Tool call validation failed: tool call validation failed: parameters for tool get_something_by_name did not match schema: errors: [missing properties: 'name', additionalProperties 'foo' not allowed]",
+    #         'type': 'invalid_request_error',
+    #         'code': 'tool_use_failed',
+    #         'failed_generation': '{"name": "get_something_by_name", "arguments": {\n  "foo": "bar"\n}}',
+    #     }
+    # }
+
+    error: _GroqToolUseFailedInnerError

tests/models/cassettes/test_groq/test_tool_regular_error.yaml

@@ -0,0 +1,49 @@
+interactions:
+- request:
+    headers:
+      accept:
+      - application/json
+      accept-encoding:
+      - gzip, deflate
+      connection:
+      - keep-alive
+      content-length:
+      - '92'
+      content-type:
+      - application/json
+      host:
+      - api.groq.com
+    method: POST
+    parsed_body:
+      messages:
+      - content: hello
+        role: user
+      model: non-existent
+      n: 1
+      stream: false
+    uri: https://api.groq.com/openai/v1/chat/completions
+  response:
+    headers:
+      alt-svc:
+      - h3=":443"; ma=86400
+      cache-control:
+      - private, max-age=0, no-store, no-cache, must-revalidate
+      connection:
+      - keep-alive
+      content-length:
+      - '153'
+      content-type:
+      - application/json
+      transfer-encoding:
+      - chunked
+      vary:
+      - Origin
+    parsed_body:
+      error:
+        code: model_not_found
+        message: The model `non-existent` does not exist or you do not have access to it.
+        type: invalid_request_error
+    status:
+      code: 404
+      message: Not Found
+version: 1