python抓取百度地图数据_Python 百度地图爬虫 - 搜索数据采集

class Hot_City_Coordinates():

"""获取热门城市"""

def __init__(self):

self.headers = {

'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/84.0.4147.105 Safari/537.36',

'Cookie': 'BIDUPSID=708E39B6F293AF8D7B18BD51D6E15C42; PSTM=1596811009; BAIDUID=708E39B6F293AF8D297129216C3D6397:FG=1; BDORZ=B490B5EBF6F3CD402E515D22BCDA1598; MCITY=-%3A; delPer=0; PSINO=5; H_PS_PSSID=32294_1456_32300_32380_32327_31660_32351_32046_32398_32116_32525_32482_22160',

}

self.content_list = [] # 存储市数据

self.content_list_ui = [] # 存储UI界面数据

def initialize(self, path_):

"""

初始化目录

"""

try:

if os.path.exists(path_): # 目录存在，返回为真

return path_

else:

os.mkdir(path_)

return path_

except Exception as e:

print(e)

def hot_city(self,map_pro, content , _path):

"""获取热门城市"""

country_url = 'https://map.baidu.com/?newmap=1&reqflag=pcmap&biz=1&from=webmap&da_par=after_baidu&pcevaname=pc4.1&qt=s&da_src=searchBox.button&wd=' + str(

content) + '&c=1&src=0&wd2=&pn=0&sug=0&l=5&b=(8661331.22,1425243.879999999;14551379.22,7913307.879999999)&from=webmap&biz_forward={%22scaler%22:2,%22styles%22:%22pl%22}&sug_forward=&auth=vQPQf98AKyG0bHBMP8U5%3DyJ31HF0NUV5uxHTNVEERVRtxjhNwzWWvy1uVt1GgvPUDZYOYIZuVtcvY1SGpuEt2gz4yYxGccZcuVtPWv3GuBRtal73hJUvhgMZSguxzBEHLNRTVtcEWe1GD8zv7u%40ZPuzztgwBzvf0wd0vyIUO7AMSFOukoPPB16A2mmiJLL'

print(country_url)

response = requests.get(country_url, headers=self.headers, verify=False)

json_con = response.json()

# 获取所有市的坐标并进行切割拼接 市的坐标可以百度

split_result = Provinces_cities_coordinates.split('\n')

city_dict = {x.split(' ')[1]: x.split(' ')[0] for x in split_result[1:]}

print(city_dict)

mkdir_path = _path + '/' + map_pro + '-' + content

province_path = self.initialize(mkdir_path)

if map_pro == '热门城市':

for x in json_con['content']:

hot_city_code = x['code']

AUTH = json_con['result']['auth']

hot_city_geo = x['geo']

hot_city_name = x['name']

city_ur1 = f'https://map.baidu.com/?newmap=1&reqflag=pcmap&biz=1&from=webmap&da_par=after_baidu&pcevaname=pc4.1&qt=spot&from=webmap&c={str(hot_city_code)}&wd={content}&wd2=&pn=0&nn=0&db=0&sug=0&addr=0&&da_src=pcmappg.poi.page&on_gel=1&src=7&gr=3&l=5&rn=50&tn=B_NORMAL_MAP&auth={AUTH}&ie=utf-8&b=({hot_city_geo})'

print('url', city_ur1)

t = threading.Thread(target=self.get_city_data,

args=(city_ur1, hot_city_name, map_pro, province_path, content))

t.start()

def get_city_data(self, city_ur1, city_name, map_pro, _path, content):

"""获取市的数据"""

city_all = requests.get(city_ur1, headers=self.headers)

city_json = city_all.json()

# 获取数据每个市的数据

try:

for city in city_json['content']:

content_dict = {}

content_dict['middle_name'] = city['name'] # 数据名字

content_dict['middle_address'] = city['addr'] # 数据地址

try:

content_dict['middle_tel'] = city['tel'] # 数据电话

except Exception as e:

content_dict['middle_tel'] = '暂无电话'

self.content_list.append(content_dict)

self.content_list_ui.append(content_dict)

print(city_name, '开始存入')

# 调用保存数据

t = threading.Thread(target=self.get_save, args=(map_pro, city_name, content, _path))

t.start()

self.content_list.clear()

except Exception as e:

print(e)

# print("最后一次打印")

if __name__ == '__main__':

...

# map_pro = input('请输入查询的省份：')

# content = input('请输入查询的数据：')

pcc = Hot_City_Coordinates()

pcc.hot_city('热门城市', '大学', '存放的路径')