getpoi - Biendata

ACM SIGKDD • $36,500 • 6114 Team • 6704 participants

KDD CUP of Fresh Air

2018-03-15 - Launch

2018-05-07 - Team Merger Deadline

2018-06-01 - Close

Home Competitions

tkn

getpoi

posted in KDD CUP of Fresh Air

2018-04-29 13:33

```python
# encoding=utf-8
import time
import requests

page_size = 20 
page_num = 0
ak = ''
sleep_time = 0.1

data_input_path = './result/'
data_output_path = './35poi-2001/'

def generate_url(query='', bounds='', page_size=page_size, page_num=page_num, ak=ak):
    if bounds != '':  # 如果按坐标查询
        return ('http://api.map.baidu.com/place/v2/search?query='
                + query + '&bounds=' + bounds + '&page_size='
                + str(page_size) + '&page_num=' + str(page_num) + '&output=json&ak=' + str(ak))
    return ''

def download_page(url):

session = requests.session()
    session.trust_env = False
    return session.get(url, headers={
        'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_2)'
                      ' AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.80 Safari/537.36'
    })

from math import*
def distance(lat1,lng1,lat2,lng2):
    radlat1 = radians(lat1)
    radlat2 = radians(lat2)
    a = radlat1 - radlat2
    b = radians(lng1) - radians(lng2)
    s = 2 * asin(sqrt(pow(sin(a / 2), 2) + cos(radlat1) * cos(radlat2) * pow(sin(b / 2), 2)))
    earth_radius = 6378137
    s = s * earth_radius
    if s < 0:
        return -s
    else:
        return s

def parse_content(content, bounds, query,latitude,longitude,name):

global page_num

poi_count = content.json()['total']

print(bounds+' '+query+' '+str(poi_count))
    if poi_count == 0:
        return ''

with open(data_output_path+'35poi_'+name+'_'+str(query)+'.txt', 'a', encoding='utf-8') as f:
        for i in content.json()['results']:
            if distance(latitude, longitude, i['location']['lng'], i['location']['lat'] < 2000):
                f.write(str(i['name'])+','+str(i['location']['lng'])+','+str(i['location']['lat'])+'\n')

if int((poi_count + page_size - 1) / page_size) > (page_num + 1):
        page_num += 1
        return generate_url(query=query, page_num=page_num, bounds=bounds)

return ''

def download_poi(query='', bounds='', latitude=0, longitude=0, name=''):

if query == '' and bounds == '':
        print('没有要下载的内容')
        return

print('开始下载 ' + query + bounds)

url = generate_url(query=query, bounds=bounds)
    while (url):
        print(url)
        content = download_page(url)
        url = parse_content(content, bounds, query,latitude,longitude,name)
    page_num = 0

time.sleep(sleep_time)

location = ''
name = []
latitude = []
longitude = []

with open(data_input_path+'BD_location.txt','r', encoding='utf-8') as f:
    location = f.read()
    location = location.split('\n')
for i in range(0,len(location)):
    name.append(location[i].split(' ')[0])
    longitude.append(float(location[i].split(' ')[1]))
    latitude.append(float(location[i].split(' ')[2]))

with open(data_input_path + 'query.txt', 'r', encoding='utf-8') as f:
    query = f.read()

k=0
for loc in location:
    bounds = ''
    for i in range(0, 4):
        for j in range(0, 4):
            lat = latitude[k] + (j-2) * 0.009
            lon = longitude[k] + (i-2) * 0.012
            bounds += str(lat) + ',' + str(lon) + ',' + str(lat + 0.009) + ',' + str(lon + 0.012) + '\n'
    for i in query.split('\n'):
        with open(data_output_path + '35poi_' + name[k] + '_' + str(i) + '.txt', 'a', encoding='utf-8') as f:
            pass
        for j in bounds.split('\n'):
            print(j)
            print(i)

download_poi(i, j, latitude[k], longitude[k], name[k])
    k+=1
```

23 comments

hortune

2018-04-30 06:41

tnlin

2018-05-02 12:57

hortune

2018-05-03 08:40