安裝Tornado
省事點可以直接用grequests庫，下面用的是tornado的異步client。 異步用到了tornado，根據官方文檔的例子修改得到一個簡單的異步爬蟲類。可以參考下最新的文檔學習下。
pip install tornado
異步爬蟲
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
|
#!/usr/bin/env python# -*- coding:utf-8 -*-import timefrom datetime import timedeltafrom tornado import httpclient, gen, ioloop, queuesimport tracebackclass AsySpider(object): """A simple class of asynchronous spider.""" def __init__(self, urls, concurrency=10, **kwargs): urls.reverse() self.urls = urls self.concurrency = concurrency self._q = queues.Queue() self._fetching = set() self._fetched = set() def fetch(self, url, **kwargs): fetch = getattr(httpclient.AsyncHTTPClient(), 'fetch') return fetch(url, **kwargs) def handle_html(self, url, html): """handle html page""" print(url) def handle_response(self, url, response): """inherit and rewrite this method""" if response.code == 200: self.handle_html(url, response.body) elif response.code == 599: # retry self._fetching.remove(url) self._q.put(url) @gen.coroutine def get_page(self, url): try: response = yield self.fetch(url) print('######fetched %s' % url) except Exception as e: print('Exception: %s %s' % (e, url)) raise gen.Return(e) raise gen.Return(response) @gen.coroutine def _run(self): @gen.coroutine def fetch_url(): current_url = yield self._q.get() try: if current_url in self._fetching: return print('fetching****** %s' % current_url) self._fetching.add(current_url) response = yield self.get_page(current_url) self.handle_response(current_url, response) # handle reponse self._fetched.add(current_url) for i in range(self.concurrency): if self.urls: yield self._q.put(self.urls.pop()) finally: self._q.task_done() @gen.coroutine def worker(): while True: yield fetch_url() self._q.put(self.urls.pop()) # add first url # Start workers, then wait for the work queue to be empty. for _ in range(self.concurrency): worker() yield self._q.join(timeout=timedelta(seconds=300000)) assert self._fetching == self._fetched def run(self): io_loop = ioloop.IOLoop.current() io_loop.run_sync(self._run)class MySpider(AsySpider): def fetch(self, url, **kwargs): """重寫父類fetch方法可以添加cookies，headers，timeout等信息""" cookies_str = "PHPSESSID=j1tt66a829idnms56ppb70jri4; pspt=%7B%22id%22%3A%2233153%22%2C%22pswd%22%3A%228835d2c1351d221b4ab016fbf9e8253f%22%2C%22_code%22%3A%22f779dcd011f4e2581c716d1e1b945861%22%7D; key=%E9%87%8D%E5%BA%86%E5%95%84%E6%9C%A8%E9%B8%9F%E7%BD%91%E7%BB%9C%E7%A7%91%E6%8A%80%E6%9C%89%E9%99%90%E5%85%AC%E5%8F%B8; think_language=zh-cn; SERVERID=a66d7d08fa1c8b2e37dbdc6ffff82d9e|1444973193|1444967835; CNZZDATA1254842228=1433864393-1442810831-%7C1444972138" # 從瀏覽器拷貝cookie字符串 headers = { 'User-Agent': 'mozilla/5.0 (compatible; baiduspider/2.0; +http://www.baidu.com/search/spider.html)', 'cookie': cookies_str } return super(MySpider, self).fetch( # 參數參考tornado文檔 url, headers=headers, request_timeout=1 ) def handle_html(self, url, html): print(url, html)def main(): urls = [] for page in range(1, 100): urls.append('http://www.baidu.com?page=%s' % page) s = MySpider(urls) s.run()if __name__ == '__main__': main() |
可以繼承這個類，塞一些url進去，然后重寫handle_page處理得到的頁面。
異步＋多進程爬蟲
還可以再變態點，加個進程池，使用了multiprocessing模塊。效率颼颼的,
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
|
#!/usr/bin/env python# -*- coding:utf-8 -*-import timefrom multiprocessing import Poolfrom datetime import timedeltafrom tornado import httpclient, gen, ioloop, queuesclass AsySpider(object): """A simple class of asynchronous spider.""" def __init__(self, urls, concurrency): urls.reverse() self.urls = urls self.concurrency = concurrency self._q = queues.Queue() self._fetching = set() self._fetched = set() def handle_page(self, url, html): filename = url.rsplit('/', 1)[1] with open(filename, 'w+') as f: f.write(html) @gen.coroutine def get_page(self, url): try: response = yield httpclient.AsyncHTTPClient().fetch(url) print('######fetched %s' % url) except Exception as e: print('Exception: %s %s' % (e, url)) raise gen.Return('') raise gen.Return(response.body) @gen.coroutine def _run(self): @gen.coroutine def fetch_url(): current_url = yield self._q.get() try: if current_url in self._fetching: return print('fetching****** %s' % current_url) self._fetching.add(current_url) html = yield self.get_page(current_url) self._fetched.add(current_url) self.handle_page(current_url, html) for i in range(self.concurrency): if self.urls: yield self._q.put(self.urls.pop()) finally: self._q.task_done() @gen.coroutine def worker(): while True: yield fetch_url() self._q.put(self.urls.pop()) # Start workers, then wait for the work queue to be empty. for _ in range(self.concurrency): worker() yield self._q.join(timeout=timedelta(seconds=300000)) assert self._fetching == self._fetched def run(self): io_loop = ioloop.IOLoop.current() io_loop.run_sync(self._run)def run_spider(beg, end): urls = [] for page in range(beg, end): urls.append('http://127.0.0.1/%s.htm' % page) s = AsySpider(urls, 10) s.run()def main(): _st = time.time() p = Pool() all_num = 73000 num = 4 # number of cpu cores per_num, left = divmod(all_num, num) s = range(0, all_num, per_num) res = [] for i in range(len(s)-1): res.append((s[i], s[i+1])) res.append((s[len(s)-1], all_num)) print res for i in res: p.apply_async(run_spider, args=(i[0], i[1],)) p.close() p.join() print time.time()-_stif __name__ == '__main__': main() |
多線程爬蟲
線程池實現.
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
|
#!/usr/bin/env python# -*- coding:utf-8 -*-import Queueimport sysimport requestsimport osimport threadingimport timeclass Worker(threading.Thread): # 處理工作請求 def __init__(self, workQueue, resultQueue, **kwds): threading.Thread.__init__(self, **kwds) self.setDaemon(True) self.workQueue = workQueue self.resultQueue = resultQueue def run(self): while 1: try: callable, args, kwds = self.workQueue.get(False) # get task res = callable(*args, **kwds) self.resultQueue.put(res) # put result except Queue.Empty: breakclass WorkManager: # 線程池管理,創建 def __init__(self, num_of_workers=10): self.workQueue = Queue.Queue() # 請求隊列 self.resultQueue = Queue.Queue() # 輸出結果的隊列 self.workers = [] self._recruitThreads(num_of_workers) def _recruitThreads(self, num_of_workers): for i in range(num_of_workers): worker = Worker(self.workQueue, self.resultQueue) # 創建工作線程 self.workers.append(worker) # 加入到線程隊列 def start(self): for w in self.workers: w.start() def wait_for_complete(self): while len(self.workers): worker = self.workers.pop() # 從池中取出一個線程處理請求 worker.join() if worker.isAlive() and not self.workQueue.empty(): self.workers.append(worker) # 重新加入線程池中 print 'All jobs were complete.' def add_job(self, callable, *args, **kwds): self.workQueue.put((callable, args, kwds)) # 向工作隊列中加入請求 def get_result(self, *args, **kwds): return self.resultQueue.get(*args, **kwds)def download_file(url): #print 'beg download', url requests.get(url).textdef main(): try: num_of_threads = int(sys.argv[1]) except: num_of_threads = 10 _st = time.time() wm = WorkManager(num_of_threads) print num_of_threads urls = ['http://www.baidu.com'] * 1000 for i in urls: wm.add_job(download_file, i) wm.start() wm.wait_for_complete() print time.time() - _stif __name__ == '__main__': main() |
這三種隨便一種都有很高的效率，但是這么跑會給網站服務器不小的壓力，尤其是小站點，還是有點節操為好。
