调用阿里云接口的python脚本

# coding=utf-8
import ssl, sys, re
import json
import time
from hdfs import Client
# Hadoop 集群端口号
Hdfs_client = Client("http://10.255.254.91:50070")
# 阿里云秘钥
ali_code = 'be6b33399ef64216ac409d7d8cf965a6'


# python3
def get_ali(phone, code):

    from urllib import request as u_req
    host = 'https://api04.aliyun.venuscn.com'
    path = '/mobile'
    method = 'GET'
    # 阿里云 账号秘钥
    appcode = str(code)
    querys = 'mobile=' + str(phone)
    bodys = {}
    url = host + path + '?' + querys

    request = u_req.Request(url)
    request.add_header('Authorization', 'APPCODE ' + appcode)
    ctx = ssl.create_default_context()
    ctx.check_hostname = False
    ctx.verify_mode = ssl.CERT_NONE
    response = u_req.urlopen(request, context=ctx)
    content = response.read().decode()
    if content:
        print(content)
        return content


# python2
def get_ali_for_py2(phone, code):

    import urllib2
    host = 'https://api04.aliyun.venuscn.com'
    path = '/mobile'
    method = 'GET'
    # 阿里云 账号秘钥
    appcode = str(code)
    querys = 'mobile=' + str(phone)
    bodys = {}
    url = host + path + '?' + querys

    request = urllib2.Request(url)
    request.add_header('Authorization', 'APPCODE ' + appcode)
    ctx = ssl.create_default_context()
    ctx.check_hostname = False
    ctx.verify_mode = ssl.CERT_NONE
    response = urllib2.urlopen(request, context=ctx)
    content = response.read()
    if content:
        print(content)
        return content


# 根据路径返回路径下的所有文件path
def make_file_list(path, file_list):
    # 判断python版本
    if sys.version_info >= (3, 0):
        file_info = Hdfs_client.walk(path).__next__()
    else:
        file_info = Hdfs_client.walk(path).next()
    for i in file_info[2]:
        file_list.append(path + '/' + i)
    for e in file_info[1]:
        make_file_list(path + '/' + e, file_list)


def get_phone():
    paths = []
    make_file_list(path='/user/hive/warehouse/thematic.db/client_phone_info_temp', file_list=paths)
    string = ''
    for path in paths:
        with Hdfs_client.read(path) as f:
            ff = f.read().decode()
            # print(ff)
        for i in ff.split('\n')[:-1]:
            print(i)
            if sys.version_info >= (3, 0):
                res = json.loads(get_ali(i.split(',')[1], ali_code))
            else:

                res = json.loads(get_ali_for_py2(i.split(',')[1], ali_code))

            if res['ret'] != 200:
                string += (i.split(',')[0]) + ','
                string += (i.split(',')[1]) + ','
                string += '未查询到相关信息' + ','
                string += '\n'
                continue

            string += (i.split(',')[0]) + ','
            string += (i.split(',')[1]) + ','
            string += res['data']['types'] + ','
            string += res['data']['prov'] + ','
            string += res['data']['city'] + ','
            string += res['data']['isp'] + ','
            string += res['data']['area_code'] + ','
            string += res['data']['city_code'] + ','
            string += res['data']['zip_code'] + ','
            string += res['data']['lng'] + ','
            string += res['data']['lat'] + ','
            string += str(res['data']['num'])
            string += '\n'
    return string


# get_phone()
def upload_to_hdfs(file_path):
    data = get_phone()
    Hdfs_client.write(file_path, data=data, encoding='utf-8')


if __name__ == '__main__':

    rq = time.strftime('%Y%m%d%H', time.localtime(time.time()))
    file_path = '/phone_message/phone_' + str(rq)
    upload_to_hdfs(file_path)