Compartilhar via


Tutorial: Desenvolver um aplicativo do Databricks com o Streamlit

Este tutorial mostra como criar um aplicativo Databricks usando o Conector SQL do Databricks para Python e Streamlit. Você aprenderá a desenvolver um aplicativo que faça o seguinte:

  • Lê uma tabela do Catálogo do Unity e a exibe em uma interface Streamlit.
  • Edita dados e os grava de volta na tabela.

Etapa 1: Configurar privilégios

Esses exemplos pressupõem que seu aplicativo usa autorização de aplicativo. A entidade de serviço do aplicativo deve ter:

  • Privilégio SELECT na tabela do Unity Catalog
  • Privilégio MODIFY na tabela do Unity Catalog
  • CAN USE privilégio no armazém de dados SQL

Para obter mais informações, consulte privilégios do Catálogo do Unity e objetos protegíveis e ACLs do SQL Warehouse.

Etapa 2: Instalar dependências

Crie um requirements.txt arquivo e inclua os seguintes pacotes:

databricks-sdk
databricks-sql-connector
streamlit
pandas

Etapa 3: Configurar a execução do aplicativo

Crie um app.yaml arquivo para definir como seu aplicativo é iniciado nos Aplicativos do Azure Databricks.

command: ['streamlit', 'run', 'app.py']

Etapa 4: Ler uma tabela do Catálogo do Unity

Este código de exemplo demonstra como ler dados de uma tabela do Catálogo do Unity e exibi-los usando Streamlit. Crie um app.py arquivo que atenda aos seguintes objetivos:

  • Usa a autenticação da entidade de serviço do aplicativo.
  • Solicita ao usuário o caminho HTTP do SQL Warehouse e o nome da tabela do Unity Catalog.
  • Executa uma SELECT * consulta na tabela especificada.
  • Exibe o resultado em Streamlit st.dataframe.

app.py

import pandas as pd
import streamlit as st
from databricks import sql
from databricks.sdk.core import Config
import os

cfg = Config()

# Use app service principal authentication
def get_connection(http_path):
    server_hostname = cfg.host
    if server_hostname.startswith('https://'):
        server_hostname = server_hostname.replace('https://', '')
    elif server_hostname.startswith('http://'):
        server_hostname = server_hostname.replace('http://', '')
    return sql.connect(
        server_hostname=server_hostname,
        http_path=http_path,
        credentials_provider=lambda: cfg.authenticate,
        _use_arrow_native_complex_types=False,
    )

# Read data from a Unity Catalog table and return it as a pandas DataFrame
def read_table(table_name: str, conn) -> pd.DataFrame:
    with conn.cursor() as cursor:
        cursor.execute(f"SELECT * FROM {table_name}")
        return cursor.fetchall_arrow().to_pandas()

# Use Streamlit input fields to accept user input
http_path_input = st.text_input(
    "Enter your Databricks HTTP Path:", placeholder="/sql/1.0/warehouses/xxxxxx"
)
table_name = st.text_input(
    "Specify a Unity Catalog table name:", placeholder="catalog.schema.table"
)

# Display the result in a Streamlit DataFrame
if http_path_input and table_name:
    conn = get_connection(http_path_input)
    df = read_table(table_name, conn)
    st.dataframe(df)
else:
    st.warning("Provide both the warehouse path and a table name to load data.")

Etapa 5: Editar uma tabela do Catálogo do Unity

Este código de exemplo permite que os usuários leiam, editem e escrevam alterações em uma tabela do Catálogo do Unity usando os recursos de edição de dados do Streamlit. Adicione a seguinte funcionalidade ao app.py arquivo:

  • Use INSERT OVERWRITE para gravar os dados atualizados de volta na tabela.

app.py

import pandas as pd
import streamlit as st
from databricks import sql
from databricks.sdk.core import Config
import math

cfg = Config()

# Use app service principal authentication
def get_connection(http_path):
    server_hostname = cfg.host
    if server_hostname.startswith('https://'):
        server_hostname = server_hostname.replace('https://', '')
    elif server_hostname.startswith('http://'):
        server_hostname = server_hostname.replace('http://', '')
    return sql.connect(
        server_hostname=server_hostname,
        http_path=http_path,
        credentials_provider=lambda: cfg.authenticate,
        _use_arrow_native_complex_types=False,
    )

# Read data from a Unity Catalog table and return it as a pandas DataFrame
def read_table(table_name: str, conn) -> pd.DataFrame:
    with conn.cursor() as cursor:
        cursor.execute(f"SELECT * FROM {table_name}")
        return cursor.fetchall_arrow().to_pandas()

# Format values for SQL, handling NaN/None as NULL
def format_value(val):
    if val is None or (isinstance(val, float) and math.isnan(val)):
        return 'NULL'
    else:
        return repr(val)

# Use `INSERT OVERWRITE` to update existing rows and insert new ones
def insert_overwrite_table(table_name: str, df: pd.DataFrame, conn):
    progress = st.empty()
    with conn.cursor() as cursor:
        rows = list(df.itertuples(index=False))
        values = ",".join([f"({','.join(map(format_value, row))})" for row in rows])
        with progress:
            st.info("Calling Databricks SQL...")
        cursor.execute(f"INSERT OVERWRITE {table_name} VALUES {values}")
    progress.empty()
    st.success("Changes saved")

# Use Streamlit input fields to accept user input
http_path_input = st.text_input(
    "Enter your Databricks HTTP Path:", placeholder="/sql/1.0/warehouses/xxxxxx"
)
table_name = st.text_input(
    "Specify a Unity Catalog table name:", placeholder="catalog.schema.table"
)

# Display the result in a Streamlit DataFrame
if http_path_input and table_name:
    conn = get_connection(http_path_input)
    if conn:
        st.success("✅ Connected successfully!")
        original_df = read_table(table_name, conn)
        edited_df = st.data_editor(original_df, num_rows="dynamic", hide_index=True)
        df_diff = pd.concat([original_df, edited_df]).drop_duplicates(keep=False)
        if not df_diff.empty:
            st.warning(f"⚠️ You have {len(df_diff) // 2} unsaved changes")
            if st.button("Save changes"):
                insert_overwrite_table(table_name, edited_df, conn)
                st.rerun()
else:
    st.warning("Provide both the warehouse path and a table name to load data.")

Próximas etapas