streamlit 实现 flink SQL运行界面

实现效果

streamlit

flink-playground.py 文件如下：

py 复制代码

import streamlit as st
import io
import contextlib
import sys
import os
import uuid
import subprocess
from jinja2 import Template

st.set_page_config(layout="wide")

# 设置页面标题
st.title("Flink SQL")

# 初始化会话状态
if 'user_id' not in st.session_state:
    st.session_state.user_id = str(uuid.uuid4())


# 创建一个输入框用于配置 JobManager 地址
st.session_state.jobmanager_address = st.text_input("JobManager 地址", value="10.50.108.7:48085")
# 创建一个文本框用于输入配置项
default_config = """
execution.checkpointing.interval=10s
execution.runtime-mode=batch
sql-client.execution.result-mode=table
sql-client.execution.max-table-result.rows=10000
pipeline.auto-watermark-interval=200
pipeline.max-parallelism=10
table.exec.state.ttl=1000
restart-strategy.type=fixed-delay
table.optimizer.join-reorder-enabled=true
table.exec.spill-compression.enabled=true
table.exec.spill-compression.block-size=128kb
""".strip()
st.session_state.config_input = st.text_area("输入配置项 (格式: key=value)", height=300, value=default_config)

# 创建一个大的文本框用于输入代码
st.session_state.sql_input = st.text_area("输入你的 Flink SQL 代码", height=500)

# 创建一个按钮
if st.button("执行 Flink Job"):
    try:
        # 读取模板文件
        with open("/work/template/pyflink-job.py.template", "r") as template_file:
            template_content = template_file.read()
        
        # 解析配置项
        config_dict = {}
        for line in st.session_state.config_input.splitlines():
            if '=' in line:
                key, value = line.split('=', 1)
                config_dict[key.strip()] = value.strip()
        
        # 使用 jinja2 模板引擎渲染模板
        template = Template(template_content)
        job_content = template.render(sqls=st.session_state.sql_input, config_items=config_dict)
        st.text("完整pyflink任务代码")
        st.code(job_content, language='python')
        
        # 将替换后的内容保存到临时文件
        file_name = f"flink_job_{st.session_state.user_id}.py"
        with open(file_name, "w") as job_file:
            job_file.write(job_content)
        
        # 使用 subprocess 执行 flink run 命令，并传递 JobManager 地址
        command = f"flink run -m {st.session_state.jobmanager_address} -py {file_name}"
        result = subprocess.run(command, shell=True, capture_output=True, text=True)
        
        # 获取捕获的输出
        captured_output = result.stdout
        
        # 显示输出结果
        st.text_area("执行结果", value=captured_output, height=200)
    except Exception as e:
        # 如果代码执行出错，打印错误信息
        st.error(f"代码执行出错: {e}")
    finally:
        # 删除临时文件
        if file_name and os.path.exists(file_name):
            os.remove(file_name)

运行：

nohup streamlit run /work/flink-playground.py --server.port 9999 2>&1  > .streamlit.log &

模板文件

模板文件根据用户输入动态更新任务配置和SQL

text 复制代码

import re
from pyflink.table import EnvironmentSettings, TableEnvironment

def remove_comments(sql):
    # 使用正则表达式删除单行注释和多行注释
    sql = re.sub(r'--.*$', '', sql, flags=re.MULTILINE)  # 删除单行注释
    sql = re.sub(r'/\*.*?\*/', '', sql, flags=re.DOTALL)  # 删除多行注释
    return sql

def execute_sql_file(table_env, sql_statements):
    sql_statements = sql_statements.split(';')
    for sql in sql_statements:
        # 删除注释
        sql = remove_comments(sql)
        sql = sql.strip()
        if sql:
            print(f"Executing SQL: {sql}")
            result = table_env.execute_sql(sql)
            # if result:
            #     result.print()

def main():
    # 创建 TableEnvironment
    env_settings = EnvironmentSettings.new_instance().in_batch_mode().build()
    table_env = TableEnvironment.create(env_settings)
    table_config = table_env.get_config()
    {% for key, value in config_items.items() %}
    table_config.get_configuration().set_string("{{ key }}", "{{ value }}")
    {% endfor %}
    sqls = """
    {{ sqls }}
    """

    # 读取 SQL 文件并执行
    execute_sql_file(table_env, sqls)

if __name__ == "__main__":
    main()