Condividi tramite

Usare OpenAI di Azure in Fabric con l'API REST (anteprima)

Importante

Questa funzionalità si trova in anteprima.

Questo documento mostra esempi di come usare OpenAI di Azure in Fabric usando l'API REST.

Inizializzazione

from synapse.ml.mlflow import get_mlflow_env_config
from trident_token_library_wrapper import PyTridentTokenLibrary

mlflow_env_configs = get_mlflow_env_config()
mwc_token = PyTridentTokenLibrary.get_mwc_token(mlflow_env_configs.workspace_id, mlflow_env_configs.artifact_id, 2)

auth_headers = {
    "Authorization" : "MwcToken {}".format(mwc_token)
}

Conversazione

GPT-4o e GPT-4o-mini sono modelli linguistici ottimizzati per le interfacce conversazionali.

import requests

def print_chat_result(messages, response_code, response):
    print("==========================================================================================")
    print("| OpenAI Input    |")
    for msg in messages:
        if msg["role"] == "system":
            print("[System] ", msg["content"])
        elif msg["role"] == "user":
            print("Q: ", msg["content"])
        else:
            print("A: ", msg["content"])
    print("------------------------------------------------------------------------------------------")
    print("| Response Status |", response_code)
    print("------------------------------------------------------------------------------------------")
    print("| OpenAI Output   |")
    if response.status_code == 200:
        print(response.json()["choices"][0]["message"]["content"])
    else:
        print(response.content)
    print("==========================================================================================")


deployment_name = "gpt-4o" # deployment_id could be one of {gpt-4o or gpt-4o-mini}
openai_url = mlflow_env_configs.workload_endpoint + f"cognitive/openai/openai/deployments/{deployment_name}/chat/completions?api-version=2025-04-01-preview"
payload = {
    "messages": [
        {"role": "system", "content": "You are an AI assistant that helps people find information."},
        {"role": "user", "content": "Does Azure OpenAI support customer managed keys?"}
    ]
}

response = requests.post(openai_url, headers=auth_headers, json=payload)
print_chat_result(payload["messages"], response.status_code, response)

Risultato

==========================================================================================
| OpenAI Input    |
[System]  You are an AI assistant that helps people find information.
Q:  Does Azure OpenAI support customer managed keys?
------------------------------------------------------------------------------------------
| Response Status | 200
------------------------------------------------------------------------------------------
| OpenAI Output   |
As of my last training cut-off in October 2023, Azure OpenAI Service did not natively support customer-managed keys (CMK) for encryption of data at rest. Data within Azure OpenAI is typically encrypted using Microsoft-managed keys. 

However, you should verify this information on the official Azure documentation or by contacting Microsoft support, as cloud service features and capabilities are frequently updated.
==========================================================================================

Incorporamenti

Un incorporamento è un formato speciale di rappresentazione dei dati che può essere facilmente usato da modelli e algoritmi di apprendimento automatico. Contiene un significato semantico ricco di informazioni di un testo, rappresentato da un vettore di numeri a virgola mobile. La distanza tra due incorporamenti nello spazio vettoriale è correlata alla somiglianza semantica tra due input originali. Ad esempio, se due testi sono simili, anche le loro rappresentazioni vettoriali dovrebbero essere simili.

Per accedere all'endpoint di incorporamenti OpenAI di Azure in Fabric, puoi inviare una richiesta API usando il seguente formato:

POST <url_prefix>/openai/deployments/<deployment_name>/embeddings?api-version=2024-02-01

deployment_name potrebbe essere text-embedding-ada-002.

import requests

def print_embedding_result(prompts, response_code, response):
    print("==========================================================================================")
    print("| OpenAI Input    |\n\t" + "\n\t".join(prompts))
    print("------------------------------------------------------------------------------------------")
    print("| Response Status |", response_code)
    print("------------------------------------------------------------------------------------------")
    print("| OpenAI Output   |")
    if response_code == 200:
        for data in response.json()['data']:
            print("\t[" + ", ".join([f"{n:.8f}" for n in data["embedding"][:10]]) + ", ... ]")
    else:
        print(response.content)
    print("==========================================================================================")

deployment_name = "text-embedding-ada-002"
openai_url = mlflow_env_configs.workload_endpoint + f"cognitive/openai/openai/deployments/{deployment_name}/embeddings?api-version=2025-04-01-preview"
payload = {
    "input": [
        "empty prompt, need to fill in the content before the request",
        "Once upon a time"
    ]
}

response = requests.post(openai_url, headers=auth_headers, json=payload)
print_embedding_result(payload["input"], response.status_code, response)

Risultato:

==========================================================================================
| OpenAI Input    |
	empty prompt, need to fill in the content before the request
	Once upon a time
------------------------------------------------------------------------------------------
| Response Status | 200
------------------------------------------------------------------------------------------
| OpenAI Output   |
	[-0.00258819, -0.00449802, -0.01700411, 0.00405622, -0.03064079, 0.01899395, -0.01295485, -0.01426286, -0.03512142, -0.01831212, ... ]
	[0.02129045, -0.02013996, -0.00462094, -0.01146069, -0.01123944, 0.00199124, 0.00228992, -0.01370478, 0.00855917, -0.01470356, ... ]
==========================================================================================

Contenuto correlato