📄 Templates

Códigos prontos para usar nos seus projetos. Copie ou baixe.

🔐 SSO
🤖 Bedrock
🔍 Trino
🐘 PostgreSQL
📦 S3
⚡ DynamoDB

Login SSO — Credenciais temporárias

Autentica via IAM Identity Center e retorna credenciais AWS temporárias para usar em qualquer serviço.

import json, time, webbrowser, boto3

SSO_START_URL = "https://d-90663e488b.awsapps.com/start"
SSO_REGION = "us-east-1"
ACCOUNT_ID = "248189947068"
ROLE_NAME = "BedrockInvoke"
# ROLE_NAME = "BedrockFullAccess"  # ML Engineers only

sso_oidc = boto3.client("sso-oidc", region_name=SSO_REGION)
client = sso_oidc.register_client(clientName="bedrock-login", clientType="public")
auth = sso_oidc.start_device_authorization(
    clientId=client["clientId"],
    clientSecret=client["clientSecret"],
    startUrl=SSO_START_URL,
)

print(f"Abra: {auth['verificationUriComplete']}")
webbrowser.open(auth["verificationUriComplete"])

token = None
while not token:
    time.sleep(auth["interval"])
    try:
        token = sso_oidc.create_token(
            clientId=client["clientId"],
            clientSecret=client["clientSecret"],
            grantType="urn:ietf:params:oauth:grant-type:device_code",
            deviceCode=auth["deviceCode"],
        )
    except sso_oidc.exceptions.AuthorizationPendingException:
        pass

sso = boto3.client("sso", region_name=SSO_REGION)
creds = sso.get_role_credentials(
    roleName=ROLE_NAME, accountId=ACCOUNT_ID,
    accessToken=token["accessToken"],
)["roleCredentials"]

print(f"export AWS_ACCESS_KEY_ID={creds['accessKeyId']}")
print(f"export AWS_SECRET_ACCESS_KEY={creds['secretAccessKey']}")
print(f"export AWS_SESSION_TOKEN={creds['sessionToken']}")

Bedrock Converse API

Chama um modelo de IA generativa. Troque os parâmetros:

ENV: AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_SESSION_TOKEN

      

Query no Trino

Conecta ao Trino e executa uma query SQL no Data Lake.

ENV: TRINO_HOST, TRINO_PORT, TRINO_USER, TRINO_PASSWORD
import os
import trino
from trino.dbapi import connect

conn = connect(
    host=os.environ["TRINO_HOST"],
    port=int(os.environ.get("TRINO_PORT", "443")),
    user=os.environ["TRINO_USER"],
    http_scheme="https",
    auth=trino.auth.BasicAuthentication(
        os.environ["TRINO_USER"],
        os.environ["TRINO_PASSWORD"],
    ),
    catalog="hive",
    schema="default",
)

cursor = conn.cursor()
cursor.execute("SELECT * FROM minha_tabela LIMIT 10")
rows = cursor.fetchall()

for row in rows:
    print(row)

Query no PostgreSQL

Conecta ao PostgreSQL e executa uma query.

ENV: PG_HOST, PG_PORT, PG_USER, PG_PASSWORD, PG_DATABASE
import os
import psycopg2

conn = psycopg2.connect(
    host=os.environ["PG_HOST"],
    port=os.environ.get("PG_PORT", "5432"),
    user=os.environ["PG_USER"],
    password=os.environ["PG_PASSWORD"],
    dbname=os.environ["PG_DATABASE"],
)

cursor = conn.cursor()
cursor.execute("SELECT * FROM minha_tabela LIMIT 10")
rows = cursor.fetchall()

for row in rows:
    print(row)

conn.close()

Ler arquivo do S3

Baixa e lê um arquivo CSV/JSON de um bucket S3.

ENV: AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_SESSION_TOKEN, S3_BUCKET
import os
import boto3
import pandas as pd
from io import StringIO

s3 = boto3.client("s3")
bucket = os.environ["S3_BUCKET"]
key = "caminho/para/arquivo.csv"

obj = s3.get_object(Bucket=bucket, Key=key)
df = pd.read_csv(StringIO(obj["Body"].read().decode("utf-8")))

print(df.head())

CRUD no DynamoDB

Lê e escreve itens em uma tabela DynamoDB.

ENV: AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_SESSION_TOKEN, DYNAMO_TABLE
import os
import boto3

dynamodb = boto3.resource("dynamodb")
table = dynamodb.Table(os.environ["DYNAMO_TABLE"])

# Escrever
table.put_item(Item={"id": "001", "nome": "João", "score": 95})

# Ler por chave
resp = table.get_item(Key={"id": "001"})
print(resp["Item"])

# Listar todos
items = table.scan()["Items"]
for item in items:
    print(item)