How to use external model as a uploaded model

OpenSearch
1

Versions (relevant - OpenSearch/Dashboard/Server OS/Browser):
OpenSearch: 2.11

Describe the issue:
I am trying to use an external model(deployed on sagemaker). I created a connector like this

{
  "name": "sagemaker_connector_test_10_19_01",
  "description": "The testing connection to sagemaker endpoint",
  "version": "1.0.0",
  "protocol": "aws_sigv4",
  "credential": {
    "access_key": "{{aws_access_key_id}}",
    "secret_key": "{{aws_secret_access_key}}",
    "session_token": "{{aws_session_token}}"
  },
  "parameters": {
    "region": "us-west-2",
    "service_name": "sagemaker"
  },
  "actions": [
    {
      "action_type": "predict",
      "method": "POST",
      "headers": {
        "content-type": "application/json"
      },
      "url": "https://runtime.sagemaker.us-west-2.amazonaws.com/endpoints/huggingface-pytorch-inference-2023-10-17-06-37-51-037/invocations",
      "request_body": "{ \"text_doc\": \"${parameters.text_doc}\" }"
    }
  ]
}

I can make predictions via send request body like this

{ 
    "parameters": {"text_doc": "dog"}
}

How to change the connector in order to send request body like this:
{
“text_doc”: “dog”
}

In the official document, they make request like this, this is no additional “parameters” :

POST /_plugins/_ml/_predict/text_embedding/WWQI44MBbzI2oUKAvNUt
{
  "text_docs":[ "today is sunny"],
  "return_number": true,
  "target_response": ["sentence_embedding"]
}

Configuration:

Relevant Logs or Screenshots:

2

When I send this request body to predict API

{ 
    "text_doc": "cat"
}

The response is

{
    "error": {
        "root_cause": [
            {
                "type": "action_request_validation_exception",
                "reason": "Validation Failed: 1: input data can't be null;"
            }
        ],
        "type": "action_request_validation_exception",
        "reason": "Validation Failed: 1: input data can't be null;"
    },
    "status": 400
}