[脚本的魅力] 分享一点代码

v2 都是大牛，所以只分享一点代码，原文描述部分基本都删了，感兴趣的可以查看原文掘金、Github。

一、js 写爬虫

首先：

var ids = [1, 2, 3]; // 项目 id

var mrs = []; // 数据结果

function mr(idx) {
  if (idx < ids.length) {
    fetch(
      "https://xxx.com/gitlab/projects/" +
        ids[idx] +
        "/getMergeRequests?order_by=updated_at&page=1&per_page=100&state=all",
      { credentials: "same-origin" }
    )
      .then((res) => res.json())
      .then((r) => {
        if (r.status === "success") {
          mrs.push(...r.result.merge_requests);
        } else {
          console.log(ids[idx] + " failed");
        }
        mr(idx + 1); // 为了避免并行发起过多请求，所以这里采用回调递归调用
      });
  } else {
    console.log("done");
  }
}

mr(0); // 开始抓取

然后导入 excel ，使用 Power Query M 语言转换数据：

let responseJson = Json.Document(File.Contents("D:\mrs.json")),
headers =
 let
 allHeaders = List.Combine(List.Transform(responseJson, Record.FieldNames)),
 uniqueHeaders = List.Distinct(allHeaders)
 in
 uniqueHeaders,
 testTable = Table.FromRecords(responseJson, headers, MissingField.UseNull)
 in
 testTable

二、统计上百个 Java 代码仓库中单元测试的数量

find . -name '*Test.java' | xargs grep -i '@Test[^a-z]' | awk -F / '{count[$3]++;} END {for(i in count) {print i,count[i]}}' | clip.exe

三、python 扫描 redis

import redis
r = redis.Redis()
t = r.scan()
while t[0]:
    t = r.scan(t[0])

四、rabbitmq 消息队列转发

import pika

connection = pika.BlockingConnection(pika.ConnectionParameters(host='ip', port=5672, virtual_host='/', credentials=pika.PlainCredentials('account','password')))

channel = connection.channel()

def backcall(ch, method, properties, body):
    # 转发
    channel.basic_publish(exchange='exchange', routing_key='routing_key', body=body)

channel.basic_consume('原队列',backcall, True)

channel.start_consuming()
connection.close()

五、python 生成 mysql 数据字典

import mysql.connector
import importlib
import sys


def generate(database_name):
    """
    生成数据库字典表
    """
    importlib.reload(sys)

    # 使用前修改配置
    conn = mysql.connector.connect(
        host='localhost',
        port='3306',
        user='',
        password='',
        use_pure=True
    )

    cursor = conn.cursor()

    cursor.execute(
        "SELECT TABLE_NAME, TABLE_COMMENT FROM information_schema.TABLES WHERE table_type='BASE TABLE' AND TABLE_SCHEMA='%s'" % database_name
    )

    tables = cursor.fetchall()

    markdown_table_header = """\n\n\n### %s (%s) \n| 序号 | 字段名称 | 数据类型 | 是否为空 | 字段说明 |\n| :--: |----| ---- | ---- | ---- |\n"""
    markdown_table_row = """| %s | %s | %s | %s | %s |"""

    f = open('dict/'+database_name + '.md', 'w', encoding="utf-8")

    for table in tables:

        cursor.execute(
            "SELECT ORDINAL_POSITION, COLUMN_NAME, COLUMN_TYPE, IS_NULLABLE, COLUMN_COMMENT "
            "FROM information_schema.COLUMNS WHERE TABLE_SCHEMA='%s' AND TABLE_NAME='%s'" % (
                database_name, table[0]
            )
        )

        tmp_table = cursor.fetchall()
        p = markdown_table_header % (table[0], remove_newline(table[1]))
        for col in tmp_table:
            colf = list(col)
            colf[2]=col[2].decode() # mysql 高级版本需要解码，代码有点丑，临时性的，能用就行
            colf[4]=col[4].decode()
            p += (remove_newline(markdown_table_row % tuple(colf)) + "\n")
        print(p)
        f.writelines(p)

    f.close()
    cursor.close()
    conn.close()


def remove_newline(text):
    """
    去除文本中的换行符号
    """
    return text.replace("\r", "").replace("\n", "")


if __name__ == '__main__':
    conn = mysql.connector.connect(
        host='localhost',
        port='3306',
        user='',
        password='',
        use_pure=True
    )

    cursor = conn.cursor()

    cursor.execute("SHOW DATABASES");

    dbs = cursor.fetchall()

    for db in dbs:
        generate(db[0])

    cursor.close()
    conn.close()