xiami_exporter.py

import requests
import re
import math
import traceback
from bs4 import BeautifulSoup

UID = '' # Put your UID here
member_auth = '' # Put your "member_auth" cookie here

ItemsPerPage = 25
XiamiUrl = 'http://www.xiami.com/space/lib-song/u/%s/page/%d'
UserAgent = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_3) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/43.0.2357.130 Safari/537.36'

def getProfilePages():
    ItemCountHint = '\u7b2c\\d+\u9875, \u5171(\\d+)\u6761'
    print('getting', XiamiUrl % (UID, 1))
    s = requests.Session()
    s.headers['user-agent'] = UserAgent
    s.cookies['member_auth'] = member_auth
    r = s.get(XiamiUrl % (UID, 1), allow_redirects=False)
    if r.status_code != 200:
        raise Exception('error: server returned code %d' % r.status_code)

    c = re.findall(ItemCountHint, r.text, re.U | re.S)
    if not c:
        raise Exception('error: invalid page', XiamiUrl % (UID, 1))
    count = int(math.ceil(int(c[0]) / ItemsPerPage))

    yield r.text
    for i in range(2, count + 1):
        print('getting', XiamiUrl % (UID, i))
        r = s.get(XiamiUrl % (UID, i))
        if r.status_code != 200:
            raise Exception('error: server returned code %d' % r.status_code)
        if not re.findall(ItemCountHint, r.text, re.U | re.S):
            raise Exception('error: invalid page', XiamiUrl % (UID, i))
        yield r.text

def newTrackTag(soup, name, artist):
    tag_file = soup.new_tag('File')
    tag_type = soup.new_tag('MediaFileType')
    tag_type.string = '0'
    tag_name = soup.new_tag('FileName')
    tag_name.string = '%s - %s.mp3' % (artist, name)
    tag_path = soup.new_tag('FilePath')
    tag_path.string = '.'
    tag_file.append(tag_type)
    tag_file.append(tag_name)
    tag_file.append(tag_path)
    return tag_file

def main():
    output = BeautifulSoup('<?xml version="1.0" encoding="utf-8"?>')
    root = output.new_tag('List', ListName='Imported')

    try:
        for page in getProfilePages():
            soup = BeautifulSoup(page)
            tracks = soup.find_all('td', class_='song_name')
            for t in tracks:
                track_name = t.a.text
                track_artist = t.find(class_='artist_name').text
                root.append(newTrackTag(output, track_name, track_artist))
    except:
        traceback.print_exc()
    else:
        output.append(root)
        open('output.kgl', 'w').write(output.prettify())

if __name__ == '__main__':
    main()