missing "name" for ChatDatabricks.with_structured_output(method="json_schema")

[WIll-Xu35]

Hi,

When using ChatDatabricks.with_structured_output(method="json_schema"), there is this error:

{
  "external_model_provider":"openai",
  "external_model_error": {
    "error":{
      "message":"Missing required parameter: 'response_format.json_schema.name'.",
      "type":"invalid_request_error",
      "param":"response_format.json_schema.name",
      "code":"missing_required_parameter"
    }
  }
}

The problem is within this code block:
https://github.com/databricks/databricks-ai-bridge/blob/main/integrations/langchain/src/databricks_langchain/chat_models.py#L973

"response_format" should have another key named "name", per GPT's request.

[camhpj]

There are several issues with this. Adding a name to the input response_format leads to:

openai.BadRequestError: Error code: 400 - {'error_code': 'BAD_REQUEST', 'message': '{\n  "error": {\n    "message": "Invalid schema for response_format \'generic-schema-name\': In context=(), \'additionalProperties\' is required to be supplied and to be false.",\n    "type": "invalid_request_error",\n    "param": "response_format",\n    "code": null\n  }\n}'}
During task with name 'extraction_node' and id '1d9a9708-7fbc-61f2-6d94-f3652e7cc87f'

Adding additionalProperties then leads to:

openai.BadRequestError: Error code: 400 - {'error_code': 'BAD_REQUEST', 'message': '{\n  "error": {\n    "message": "Invalid schema for response_format \'generic-schema-name\': In context=(), \'required\' is required to be supplied and to be an array including every key in properties. Extra required key \'x\' supplied.",\n    "type": "invalid_request_error",\n    "param": "response_format",\n    "code": null\n  }\n}'}

Toy example used for the above:

    from databricks_langchain import ChatDatabricks
    from pydantic import BaseModel, ConfigDict, Field
    class Schema(BaseModel):
        model_config = ConfigDict(strict=True)
        z: int = Field(...)
        x: Dict[str, float] = Field(...)

    model = ChatDatabricks(model="databricks-gpt-5", stream_usage=True)
    schema=Schema.model_json_schema()
    schema["additionalProperties"] = False

    # `name` is added to the `response_format` created under the hood
    sm = model.with_structured_output(schema, include_raw=True, method="json_schema")
    raw = sm.invoke("1+1=z")

[sunishsheth2009]

Thank you for raising this issue, is there a fix that you can help contribute to fix this issue?