Python爬取国外天气预报网站的方法

#encoding=utf-8
import httplib
import urllib2
import time
from threading import Thread
import threading
from Queue import Queue
from time import sleep
import re
import copy
lang = "fr"
count = 0
class Location:
  # Location(False, "中国", "北京", "zh")
  # Location(True, "", "亚洲", "zh")
  def __init__(self, is_beyond_country, country_name, loc_name, lang):
    self.country_name = country_name
    self.loc_name = loc_name
    self.lang = lang
    self.is_beyond_country = is_beyond_country
prn_lock = threading.RLock()
def GetLocationURLs(url, recursive):
  global count
  if url.find("weather-forecast") != -1:
    count = count + 1
    if count % 500 == 0:
      prn_lock.acquire()
      print "count:%d" % (count)
      prn_lock.release()
    return [url]
  page = urllib2.urlopen(url).read()
  time.sleep(0.01)
  #"<h6><a href=\"http://www.accuweather.com/zh/browse-locations/afr\"><em>Africa</em></a></h6>"
  pattern = "<h6><a href=\"(.*)\"><em>(.*)</em></a></h6>"
  locs = re.findall(pattern, page)
  locs = [(url, name) for url, name in locs if url.find("browse-locations") != -1 or url.find("weather-forecast") != -1]
  if not recursive:
    urls = [url for url, name in locs]
    return urls
  urls = []
  for _url, _name in locs:
    lst = GetLocationURLs(_url, True)
    urls.extend(lst)
  return urls
#entry_url = "http://www.accuweather.com/zh/browse-locations"
entry_url = "http://www.accuweather.com/%s/browse-locations/eur/fr" % (lang)
#regions = ["afr", "ant", "arc", "asi", "cac", "eur", "mea", "nam", "ocn", "sam"]
#regions = ["eur"]
#region_urls = [ "%s/%s" % (entry_url, reg) for reg in regions]
#region_urls = ["http://www.accuweather.com/zh/browse-locations/eur/fr"]
sub_urls = GetLocationURLs(entry_url, False)
print len(sub_urls)
print sub_urls
q = Queue()
location_urls = []
ThreadNum = 5
lock = threading.RLock()
for url in sub_urls:
  q.put(url)
def working():
  while True:
    url = q.get()
    lst = GetLocationURLs(url, True)
    print "%s %d urls " % (url, len(lst))
    lock.acquire()
    location_urls.extend(lst)
    lock.release()
    q.task_done()
for i in range(ThreadNum):
  t = Thread(target=working)
  t.setDaemon(True)
  t.start()
q.join()  
fp = open('locations.txt', "w")
fp.write("\n".join(location_urls))
fp.close()
#for url in location_urls:
#  print url
#location_urls = GetLocationURLs(entry_url)
'''
def Fetch(url):
  try:
    print url
    web_path = url[0]
    local_name = url[1]   
    print "web_path:", web_path
    print "local_name:", local_name
    sContent = urllib2.urlopen(web_path).read()
    savePath = "D:\\Course\\NLP_Manning\\%s" % (local_name)
    print savePath
    file = open(savePath,'wb')
    file.write(sContent)
    file.close()
    print savePath + " saved";
  except:
    pass;
def working():
  while True:
    url = q.get()
    Fetch(url)
    sleep(10)
    q.task_done()
#root_url = "https://class.coursera.org/nlp/lecture/index"
root_url = "https://class.coursera.org/nlp/lecture/index"
page = urllib2.urlopen(root_url).read()
for i in range(NUM):
  t = Thread(target=working)
  t.setDaemon(True)
  t.start()
urls = copy.deepcopy(ppt_urls)
urls.extend(srt_urls)
urls.extend(video_urls)
print len(ppt_urls)
print len(srt_urls)
print len(video_urls)
print len(urls)
for url in urls:
  q.put(url)
q.join()
'''
'''
root_url = "http://www.accuweather.com/zh/cn/andingmen/57494/weather-forecast/57494"
page = urllib2.urlopen(root_url).read()
print page
'''

#encoding=utf-8
import sys
import httplib
import urllib2
import time
from threading import Thread
import threading
from Queue import Queue
from time import sleep
import re
import copy
from xml.dom import minidom
import HTMLParser
import datetime
q = Queue()
locks = [threading.RLock() for i in range(2)]
ThreadNumber = 20
locations = {}
conds = {}
def FindCountryBreadCrumbs(page):
  lines = page.splitlines()
  count = 0
  start = -1
  opened = False
  for line in lines:
    if line.find("<ul id=\"country-breadcrumbs\">") != -1:
      start = count
      opened = True
    if opened and line.find("</ul>") != -1:
      end = count
      opened = False
    count = count + 1
  return "\n".join(lines[start: (end + 1)])
def GetText(nodelist):
  rc = []
  for node in nodelist:
    if node.nodeType == node.TEXT_NODE:
      rc.append(HTMLParser.HTMLParser().unescape(node.data))
  return ''.join(rc)
def FindCondition(page):
  pat = "<span class=\"cond\">(.*"
  cds = re.findall(pat, page)
  cds = [HTMLParser.HTMLParser().unescape(cd).encode("utf-8") for cd in cds]
  return cds  
def ExtractInfo(url):
  try:
    page = urllib2.urlopen(url).read()
  except Exception, e:
    return []
  text = FindCountryBreadCrumbs(page)
  text = HTMLParser.HTMLParser().unescape(text)
  dom = minidom.parseString(text.encode("utf-8"))
  locs = []
  lis = dom.getElementsByTagName("li")
  for li in lis:
    adr_list = li.getElementsByTagName("a")
    if adr_list:
      locs.append(GetText(adr_list[0].childNodes).encode("utf-8"))
    strs = li.getElementsByTagName("strong")
    if strs:
      locs.append(GetText(strs[0].childNodes).encode("utf-8"))
  cds = FindCondition(page)
  return locs, cds
def AddMap(lst, m):
  for x in lst:
    if m.get(x) == None:
      m[x] = 1
def working():
  while True:
    urls = q.get()
    #print len(urls)
    m = {}
    m2 = {}
    count = 0
    for url in urls:
      count = count + 1
      #print "%d/%d" % (count, len(urls))
      locs, cds = ExtractInfo(url)
      AddMap(locs, m)
      AddMap(cds, m2)
    locks[1].acquire()
    AddMap(m.keys(), locations)
    AddMap(m2.keys(), conds)
    locks[1].release()
    q.task_done()
def main():
  if len(sys.argv) < 2:
    exit()
  loc_path = sys.argv[1]
  fp = open(loc_path, "r")
  urls = [line.strip() for line in fp]
  fp.close()
  #urls = urls[0:1000]
  blocks = len(urls) / ThreadNumber + 1
  for start in range(0, len(urls), blocks):
    end = start + blocks
    if end > len(urls):
      end = len(urls)
    q.put(urls[start:end])
  for i in range(ThreadNumber):
    t = Thread(target=working)
    t.setDaemon(True)
    t.start()
  q.join()
  fp = open("location_name.fr", "w")
  fp.write("\n".join(locations.keys()))
  fp.close()
  fp = open("conditions.fr", "w")
  fp.write("\n".join(conds.keys()))
  fp.close()
if __name__ == '__main__':
  main()