Menangani respons API dan error

Pengantar Amazon Bedrock

Nikhil Rangarajan

Data Scientist

Mengelola rate limit

Terapkan strategi exponential backoff
Optimalkan penggunaan API dengan retry cerdas

def backoff_retry(max_retries=3, 
                  init_delay=1):

  for attempt in range(max_retries):
    try: 
      response = bedrock.invoke_model()

    # Retries when we have an exception
  except Exception as e:
    if "ThrottlingException" in str(e):
    # Exponential backoff
      time.sleep(base_delay * 
                   (2 ** attempt))

Memproses respons batch

Proses banyak permintaan secara efisien
Tangani respons batch dengan aman
Jaga kinerja optimal

def process_batch(prompts, batch=5):
  results = []
  for i in range(0, len(prompts), 
                 batch):
    responses = [invoke_model(p) for p 
                 in prompts[i:i+batch]]

Pascaproses respons

Validasi struktur respons
Muat dan parse body respons
Periksa key yang diharapkan
Pastikan kualitas data konsisten

def postprocess_response(response):

    data=json.loads(response['body']
                   .read())

    if 'content' in data: 
    return data['content'].strip()

return None

Pemulihan error

Terapkan strategi fallback saat model utama bermasalah
Tangani error spesifik model
Pastikan keandalan tingkat produksi

try:
    response = bedrock.invoke_model(modelId="anthropic.claude-3-5-sonnet-v2:0", 
        body=json.dumps(request_body))

except ClientError: 
    response = bedrock.invoke_model(modelId="amazon.nova-lite-v1:0")

Ayo berlatih!

Pengantar Amazon Bedrock