Apache Zeppelin

# Download e instalação
wget https://archive.apache.org/dist/zeppelin/zeppelin-0.11.0/zeppelin-0.11.0-bin-all.tgz
tar -xzf zeppelin-0.11.0-bin-all.tgz
cd zeppelin-0.11.0-bin-all

# Configuração básica
cp conf/zeppelin-site.xml.template conf/zeppelin-site.xml
cp conf/zeppelin-env.sh.template conf/zeppelin-env.sh

# Configurar Spark (zeppelin-env.sh)
export SPARK_HOME=/path/to/spark
export PYSPARK_PYTHON=python3
export PYSPARK_DRIVER_PYTHON=python3

# Iniciar Zeppelin
./bin/zeppelin-daemon.sh start

# Acesso via navegador: http://localhost:8080

# Docker Compose
version: '3.8'
services:
  zeppelin:
    image: apache/zeppelin:0.11.0
    container_name: zeppelin
    ports:
      - "8080:8080"
    volumes:
      - ./notebooks:/opt/zeppelin/notebook
      - ./logs:/opt/zeppelin/logs
      - ./conf:/opt/zeppelin/conf
    environment:
      - ZEPPELIN_ADDR=0.0.0.0
      - ZEPPELIN_PORT=8080
      - SPARK_MASTER=local[*]
    networks:
      - zeppelin-network

networks:
  zeppelin-network:
    driver: bridge

# Executar
docker-compose up -d

// Célula 1: Configuração Spark
%spark.conf
spark.sql.adaptive.enabled true
spark.sql.adaptive.coalescePartitions.enabled true

// Célula 2: Carregar dados
%spark.sql
CREATE OR REPLACE TEMPORARY VIEW sales USING parquet
OPTIONS (path "/data/sales.parquet")

// Célula 3: Análise de vendas por mês
%spark.sql
SELECT 
    DATE_FORMAT(order_date, 'yyyy-MM') as month,
    SUM(amount) as total_sales,
    COUNT(*) as order_count,
    AVG(amount) as avg_order_value
FROM sales 
WHERE order_date >= '2024-01-01'
GROUP BY DATE_FORMAT(order_date, 'yyyy-MM')
ORDER BY month

// Célula 4: Python para análise avançada
%spark.pyspark
from pyspark.sql.functions import *
from pyspark.ml.feature import VectorAssembler
from pyspark.ml.regression import LinearRegression

# Carregar dados
df = spark.table("sales")

# Preparar features
assembler = VectorAssembler(
    inputCols=["quantity", "discount"],
    outputCol="features"
)

# Treinar modelo
lr = LinearRegression(featuresCol="features", labelCol="amount")
model = lr.fit(assembler.transform(df))

print(f"RMSE: {model.summary.rootMeanSquaredError}")
print(f"R²: {model.summary.r2}")

// Célula com visualização customizada
%spark.sql
SELECT 
    category,
    SUM(amount) as revenue,
    COUNT(*) as orders
FROM sales 
GROUP BY category
ORDER BY revenue DESC

// Configuração de visualização (via UI)
{
  "type": "multiBarChart",
  "settings": {
    "xAxis": "category",
    "yAxis": ["revenue", "orders"],
    "group": [],
    "scatter": {}
  }
}

// Visualização com D3.js customizada
%angular

# REST API para executar notebooks
# Executar parágrafo
curl -X POST \
  http://localhost:8080/api/notebook/2A94M5J1Z/paragraph/20150210-015259_1403135953 \
  -H 'Content-Type: application/json'

# Obter resultado
curl -X GET \
  http://localhost:8080/api/notebook/2A94M5J1Z/paragraph/20150210-015259_1403135953

# Criar notebook via API
curl -X POST \
  http://localhost:8080/api/notebook \
  -H 'Content-Type: application/json' \
  -d '{
    "name": "API Notebook",
    "paragraphs": [
      {
        "title": "Spark SQL Example",
        "text": "%spark.sql\nSELECT * FROM sales LIMIT 10"
      }
    ]
  }'

# Agendar execução
curl -X POST \
  http://localhost:8080/api/notebook/2A94M5J1Z/cron \
  -H 'Content-Type: application/json' \
  -d '{
    "cron": "0 9 * * MON",
    "releaseResource": true
  }'