Chat with Azure OpenAI models using your own data facing issue data Unrecognized request argument supplied: dataSources

Saurabh Upadhya 25

contain data source  information  i tired to add data source information json is give  error
{
    "error": {
        "message": "Unrecognized request argument supplied: dataSources",
        "type": "invalid_request_error",
        "param": null,
        "code": null
    }
}
i passed data

    "dataSources": [
        {
            "type": "AzureCognitiveSearch",
            "parameters": {
                "endpoint": "https://xyz.search.windows.net",
                "key": "*********YWsWRK8lMu8mtYSAzSeBmhdOT",
                "indexName": "azureblob-index"
            }
        }
    ]
please guide me what endpoint and   key should i  pass ? 
thanks in advance

VasaviLankipalle-MSFT 17,121 Reputation points

2023-07-11T00:21:37.5666667+00:00

Hello @Saurabh Upadhya , Thanks for using Microsoft Q&A platform.

Regarding the endpoint and key details, please navigate to the Azure OpenAI resource which you have created in the Azure portal and copy these values:

Here is the detailed information related to the key and endpoint values: https://learn.microsoft.com/en-us/azure/cognitive-services/openai/use-your-data-quickstart?tabs=powershell&pivots=rest-api#retrieve-required-variables

I hope this helps. Try these and let me know.
Saurabh Upadhya 25 Reputation points

2023-07-11T07:33:28.2633333+00:00

thanks for quick response

i tried access without adding data source its working fine but when i tired with add data source for AzureCognitiveSearch with valid api key and endpoints its giving me error (i want try AzureCognitiveSearch + openai with own data) kindly please me provide which API should i call or what valid parameter should i pass to achieve this

Polisetty, Balaji 0

Even i am facing same issue when calling rest api with dataSources .

{
  "error": {
    "message": "Unrecognized request argument supplied: dataSources",
    "type": "invalid_request_error",
    "param": null,
    "code": null
  }
}

Polisetty, Balaji 0 Reputation points

2023-07-12T00:48:03.1333333+00:00

It would be good if we can have an example with datasource
Robin Egberts 30 Reputation points

2023-07-12T08:49:51.1233333+00:00

It's because the URL has an additional statement: soo instead of deployments/NAME/chat/completions... you would have /deployments/NAME/extensions/chat/completions... Otherwise you get the unrecognized argument supplied error
Polisetty, Balaji 0 Reputation points

2023-07-12T16:09:14.28+00:00

after adding extensions i am getting {"error":{"code":"404","message": "Resource not found"}}
Anonymous

2023-07-12T17:58:59.4433333+00:00

I am with the same issue. After adding extensions I am getting data: { error: { code: '404', message: 'Resource not found' } }
Anonymous

2023-07-12T18:38:42.1466667+00:00

Hey guys, I was able to search using the parameter in the URL 'extensions' and the api version 2023-06-01-preview. Hope this helps.
Polisetty, Balaji 0 Reputation points

2023-07-12T19:04:46.1633333+00:00

thank you @Vitor Pedrosa its working for me.
Atte Mäkinen 5 Reputation points

2023-09-22T13:13:44.7633333+00:00
Hello - I struggled with this one for a while - but finally I got it work.
Here is example - hopefully it helps you guys.
URI: https://xxxxxxazureopenai.openai.azure.com/openai/deployments/TestXXXX/extensions/chat/completions?api-version=2023-06-01-preview

Content-Type: application/json
api-key: xxxxxxxxx

Body (I used Power Automate):

{ "dataSources": [ { "type": "AzureCognitiveSearch", "parameters": { "endpoint": "https://azureopenai-xxxxx-xxxx.search.windows.net", "key": "xxxxxxxx", "indexName": "@{triggerBody()['text_1']}" } } ], "messages": [ { "role": "user", "content": "@{triggerBody()['text']}" } ] }

1 answer

Alexa Kiss 0 Microsoft Employee

Hi,
I got it work based on this document
https://learn.microsoft.com/en-us/azure/ai-services/openai/use-your-data-quickstart?tabs=command-line&pivots=rest-api

deployment_id = 'deployment-gpt-3-5'

request.url = f"{openai.api_base}/openai/deployments/{deployment_id}/extensions/chat/completions?api-version={openai.api_version}"

#use the above URL in the setup_byod function
def setup_byod(deployment_id: str) -> None: ....



def completion(deployment_id: str, prompt_input: str):
      
	completion = openai.ChatCompletion.create(             
		deployment_id=deployment_id,         
		dataSources=[              
			{   "type": "AzureCognitiveSearch",	
				"parameters": {                     
					"endpoint": search_endpoint,                     
					"key": search_key,                     
					"indexName": search_index_name,
                 	}             
			}         
		],        
		messages=[{"role": "user", "content": prompt_input}]
	)     
	print(completion['choices'][0]['message']['content'])

Share via

Chat with Azure OpenAI models using your own data facing issue data Unrecognized request argument supplied: dataSources

1 answer

Your answer