Migrate python/js script to c#?

[nk-alex]

I'm using OpenAI API to perform data extraction from images of driver licenses uploaded by users.

Here is a sample code for python api:

		client = OpenAI(api_key=self.api_key, organization="", project="")

        file_obj = io.BytesIO(byte_array)
        file_obj.name = "driverLicense" + file_extension

        if file_extension == ".pdf":
            content_type = "input_file"

        else:
            content_type = "input_image"

        file = client.files.create(file=file_obj, purpose="user_data")

        response = client.responses.create(
            model="gpt-4o-mini", 
            input=[
                {
                    "role": "system",
                    "content": [
                        {
                            "type": "input_text",
                            "text": openai_inference_prompt,
                        }
                    ]
                },
                {
                    "role": "user",
                    "content": [
                        {
                            "type": content_type,
                            "file_id": file.id,
                        }
                    ]
                }
            ],
            max_output_tokens = 1000,
            temperature=1,
            top_p=1,
            text={ "format": { "type": "json_object" } }
        )

        return response.output_text

On javascript, the script is very similar to this as well. But now I need to migrate this script to c# in order to be able to run it from my .NET MAUI application. But I'm really struggling since c# API is COMPLETELY different than python and js (which I don't quite understand since this is an official repo). Official docs also don't show the code for c#.

The samples on this repo Readme page brought me to AssistantClient which seems to be unusable without adding #pragma warning disable OPENAI001 as won't complie due to OpenAI.Assistants.AssistantClient is for evaluation purposes only and is subject to change or removel in future updates. Suppress this diagnostic to proceed.

So, which is the correct way to achieve the same as in the javascript/python script?

[joseharriaga]

Thank you for reaching out, @nk-alex ! The code above appears to be using the Responses API, but you mention trying to use the Assistants API in your C# code. This might be the reason for why your solution might be looking so different.

You can use the Responses API starting with version 2.2.0-beta.4 of the NuGet package. You will find everything related to it under the OpenAI.Responses namespace.