|
1
|
|
|
#!/usr/bin/env python |
|
2
|
|
|
# -*- encoding: utf-8 -*- |
|
3
|
|
|
# vim: set et sw=4 ts=4 sts=4 ff=unix fenc=utf8: |
|
4
|
|
|
# Author: Binux<[email protected]> |
|
5
|
|
|
# http://binux.me |
|
6
|
|
|
# Created on 2012-12-17 11:07:19 |
|
7
|
|
|
|
|
8
|
|
|
from __future__ import unicode_literals |
|
9
|
|
|
|
|
10
|
|
|
import os |
|
11
|
|
|
import six |
|
12
|
|
|
import copy |
|
13
|
|
|
import time |
|
14
|
|
|
import json |
|
15
|
|
|
import logging |
|
16
|
|
|
import traceback |
|
17
|
|
|
import functools |
|
18
|
|
|
import threading |
|
19
|
|
|
import tornado.ioloop |
|
20
|
|
|
import tornado.httputil |
|
21
|
|
|
import tornado.httpclient |
|
22
|
|
|
import pyspider |
|
23
|
|
|
|
|
24
|
|
|
from six.moves import queue, http_cookies |
|
25
|
|
|
from six.moves.urllib.robotparser import RobotFileParser |
|
26
|
|
|
from requests import cookies |
|
27
|
|
|
from six.moves.urllib.parse import urljoin, urlsplit |
|
28
|
|
|
from tornado import gen |
|
29
|
|
|
from tornado.curl_httpclient import CurlAsyncHTTPClient |
|
30
|
|
|
from tornado.simple_httpclient import SimpleAsyncHTTPClient |
|
31
|
|
|
|
|
32
|
|
|
from pyspider.libs import utils, dataurl, counter |
|
33
|
|
|
from pyspider.libs.url import quote_chinese |
|
34
|
|
|
from .cookie_utils import extract_cookies_to_jar |
|
35
|
|
|
logger = logging.getLogger('fetcher') |
|
36
|
|
|
|
|
37
|
|
|
|
|
38
|
|
|
class MyCurlAsyncHTTPClient(CurlAsyncHTTPClient): |
|
39
|
|
|
|
|
40
|
|
|
def free_size(self): |
|
41
|
|
|
return len(self._free_list) |
|
42
|
|
|
|
|
43
|
|
|
def size(self): |
|
44
|
|
|
return len(self._curls) - self.free_size() |
|
45
|
|
|
|
|
46
|
|
|
|
|
47
|
|
|
class MySimpleAsyncHTTPClient(SimpleAsyncHTTPClient): |
|
48
|
|
|
|
|
49
|
|
|
def free_size(self): |
|
50
|
|
|
return self.max_clients - self.size() |
|
51
|
|
|
|
|
52
|
|
|
def size(self): |
|
53
|
|
|
return len(self.active) |
|
54
|
|
|
|
|
55
|
|
|
fetcher_output = { |
|
56
|
|
|
"status_code": int, |
|
57
|
|
|
"orig_url": str, |
|
58
|
|
|
"url": str, |
|
59
|
|
|
"headers": dict, |
|
60
|
|
|
"content": str, |
|
61
|
|
|
"cookies": dict, |
|
62
|
|
|
} |
|
63
|
|
|
|
|
64
|
|
|
|
|
65
|
|
|
class Fetcher(object): |
|
66
|
|
|
user_agent = "pyspider/%s (+http://pyspider.org/)" % pyspider.__version__ |
|
67
|
|
|
default_options = { |
|
68
|
|
|
'method': 'GET', |
|
69
|
|
|
'headers': { |
|
70
|
|
|
}, |
|
71
|
|
|
'use_gzip': True, |
|
72
|
|
|
'timeout': 120, |
|
73
|
|
|
'connect_timeout': 20, |
|
74
|
|
|
} |
|
75
|
|
|
phantomjs_proxy = None |
|
76
|
|
|
splash_endpoint = None |
|
77
|
|
|
splash_lua_source = open(os.path.join(os.path.dirname(__file__), "splash_fetcher.lua")).read() |
|
78
|
|
|
robot_txt_age = 60*60 # 1h |
|
79
|
|
|
|
|
80
|
|
|
def __init__(self, inqueue, outqueue, poolsize=100, proxy=None, async=True): |
|
81
|
|
|
self.inqueue = inqueue |
|
82
|
|
|
self.outqueue = outqueue |
|
83
|
|
|
|
|
84
|
|
|
self.poolsize = poolsize |
|
85
|
|
|
self._running = False |
|
86
|
|
|
self._quit = False |
|
87
|
|
|
self.proxy = proxy |
|
88
|
|
|
self.async = async |
|
89
|
|
|
self.ioloop = tornado.ioloop.IOLoop() |
|
90
|
|
|
|
|
91
|
|
|
self.robots_txt_cache = {} |
|
92
|
|
|
|
|
93
|
|
|
# binding io_loop to http_client here |
|
94
|
|
|
if self.async: |
|
95
|
|
|
self.http_client = MyCurlAsyncHTTPClient(max_clients=self.poolsize, |
|
96
|
|
|
io_loop=self.ioloop) |
|
97
|
|
|
else: |
|
98
|
|
|
self.http_client = tornado.httpclient.HTTPClient(MyCurlAsyncHTTPClient, max_clients=self.poolsize) |
|
99
|
|
|
|
|
100
|
|
|
self._cnt = { |
|
101
|
|
|
'5m': counter.CounterManager( |
|
102
|
|
|
lambda: counter.TimebaseAverageWindowCounter(30, 10)), |
|
103
|
|
|
'1h': counter.CounterManager( |
|
104
|
|
|
lambda: counter.TimebaseAverageWindowCounter(60, 60)), |
|
105
|
|
|
} |
|
106
|
|
|
|
|
107
|
|
|
def send_result(self, type, task, result): |
|
108
|
|
|
'''Send fetch result to processor''' |
|
109
|
|
|
if self.outqueue: |
|
110
|
|
|
try: |
|
111
|
|
|
self.outqueue.put((task, result)) |
|
112
|
|
|
except Exception as e: |
|
113
|
|
|
logger.exception(e) |
|
114
|
|
|
|
|
115
|
|
|
def fetch(self, task, callback=None): |
|
116
|
|
|
if self.async: |
|
117
|
|
|
return self.async_fetch(task, callback) |
|
118
|
|
|
else: |
|
119
|
|
|
return self.async_fetch(task, callback).result() |
|
120
|
|
|
|
|
121
|
|
|
@gen.coroutine |
|
122
|
|
|
def async_fetch(self, task, callback=None): |
|
123
|
|
|
'''Do one fetch''' |
|
124
|
|
|
url = task.get('url', 'data:,') |
|
125
|
|
|
if callback is None: |
|
126
|
|
|
callback = self.send_result |
|
127
|
|
|
|
|
128
|
|
|
type = 'None' |
|
129
|
|
|
start_time = time.time() |
|
130
|
|
|
try: |
|
131
|
|
|
if url.startswith('data:'): |
|
132
|
|
|
type = 'data' |
|
133
|
|
|
result = yield gen.maybe_future(self.data_fetch(url, task)) |
|
134
|
|
|
elif task.get('fetch', {}).get('fetch_type') in ('js', 'phantomjs'): |
|
135
|
|
|
type = 'phantomjs' |
|
136
|
|
|
result = yield self.phantomjs_fetch(url, task) |
|
137
|
|
|
elif task.get('fetch', {}).get('fetch_type') in ('splash', ): |
|
138
|
|
|
type = 'splash' |
|
139
|
|
|
result = yield self.splash_fetch(url, task) |
|
140
|
|
|
else: |
|
141
|
|
|
type = 'http' |
|
142
|
|
|
result = yield self.http_fetch(url, task) |
|
143
|
|
|
except Exception as e: |
|
144
|
|
|
logger.exception(e) |
|
145
|
|
|
result = self.handle_error(type, url, task, start_time, e) |
|
146
|
|
|
|
|
147
|
|
|
callback(type, task, result) |
|
148
|
|
|
self.on_result(type, task, result) |
|
149
|
|
|
raise gen.Return(result) |
|
150
|
|
|
|
|
151
|
|
|
def sync_fetch(self, task): |
|
152
|
|
|
'''Synchronization fetch, usually used in xmlrpc thread''' |
|
153
|
|
|
if not self._running: |
|
154
|
|
|
return self.ioloop.run_sync(functools.partial(self.async_fetch, task, lambda t, _, r: True)) |
|
155
|
|
|
|
|
156
|
|
|
wait_result = threading.Condition() |
|
157
|
|
|
_result = {} |
|
158
|
|
|
|
|
159
|
|
|
def callback(type, task, result): |
|
160
|
|
|
wait_result.acquire() |
|
161
|
|
|
_result['type'] = type |
|
162
|
|
|
_result['task'] = task |
|
163
|
|
|
_result['result'] = result |
|
164
|
|
|
wait_result.notify() |
|
165
|
|
|
wait_result.release() |
|
166
|
|
|
|
|
167
|
|
|
wait_result.acquire() |
|
168
|
|
|
self.ioloop.add_callback(self.fetch, task, callback) |
|
169
|
|
|
while 'result' not in _result: |
|
170
|
|
|
wait_result.wait() |
|
171
|
|
|
wait_result.release() |
|
172
|
|
|
return _result['result'] |
|
173
|
|
|
|
|
174
|
|
|
def data_fetch(self, url, task): |
|
175
|
|
|
'''A fake fetcher for dataurl''' |
|
176
|
|
|
self.on_fetch('data', task) |
|
177
|
|
|
result = {} |
|
178
|
|
|
result['orig_url'] = url |
|
179
|
|
|
result['content'] = dataurl.decode(url) |
|
180
|
|
|
result['headers'] = {} |
|
181
|
|
|
result['status_code'] = 200 |
|
182
|
|
|
result['url'] = url |
|
183
|
|
|
result['cookies'] = {} |
|
184
|
|
|
result['time'] = 0 |
|
185
|
|
|
result['save'] = task.get('fetch', {}).get('save') |
|
186
|
|
|
if len(result['content']) < 70: |
|
187
|
|
|
logger.info("[200] %s:%s %s 0s", task.get('project'), task.get('taskid'), url) |
|
188
|
|
|
else: |
|
189
|
|
|
logger.info( |
|
190
|
|
|
"[200] %s:%s data:,%s...[content:%d] 0s", |
|
191
|
|
|
task.get('project'), task.get('taskid'), |
|
192
|
|
|
result['content'][:70], |
|
193
|
|
|
len(result['content']) |
|
194
|
|
|
) |
|
195
|
|
|
|
|
196
|
|
|
return result |
|
197
|
|
|
|
|
198
|
|
|
def handle_error(self, type, url, task, start_time, error): |
|
199
|
|
|
result = { |
|
200
|
|
|
'status_code': getattr(error, 'code', 599), |
|
201
|
|
|
'error': utils.text(error), |
|
202
|
|
|
'traceback': traceback.format_exc(), |
|
203
|
|
|
'content': "", |
|
204
|
|
|
'time': time.time() - start_time, |
|
205
|
|
|
'orig_url': url, |
|
206
|
|
|
'url': url, |
|
207
|
|
|
"save": task.get('fetch', {}).get('save') |
|
208
|
|
|
} |
|
209
|
|
|
logger.error("[%d] %s:%s %s, %r %.2fs", |
|
210
|
|
|
result['status_code'], task.get('project'), task.get('taskid'), |
|
211
|
|
|
url, error, result['time']) |
|
212
|
|
|
return result |
|
213
|
|
|
|
|
214
|
|
|
allowed_options = ['method', 'data', 'timeout', 'cookies', 'use_gzip', 'validate_cert'] |
|
215
|
|
|
|
|
216
|
|
|
def pack_tornado_request_parameters(self, url, task): |
|
217
|
|
|
fetch = copy.deepcopy(self.default_options) |
|
218
|
|
|
fetch['url'] = url |
|
219
|
|
|
fetch['headers'] = tornado.httputil.HTTPHeaders(fetch['headers']) |
|
220
|
|
|
fetch['headers']['User-Agent'] = self.user_agent |
|
221
|
|
|
task_fetch = task.get('fetch', {}) |
|
222
|
|
|
for each in self.allowed_options: |
|
223
|
|
|
if each in task_fetch: |
|
224
|
|
|
fetch[each] = task_fetch[each] |
|
225
|
|
|
fetch['headers'].update(task_fetch.get('headers', {})) |
|
226
|
|
|
|
|
227
|
|
|
if task.get('track'): |
|
228
|
|
|
track_headers = tornado.httputil.HTTPHeaders( |
|
229
|
|
|
task.get('track', {}).get('fetch', {}).get('headers') or {}) |
|
230
|
|
|
track_ok = task.get('track', {}).get('process', {}).get('ok', False) |
|
231
|
|
|
else: |
|
232
|
|
|
track_headers = {} |
|
233
|
|
|
track_ok = False |
|
234
|
|
|
# proxy |
|
235
|
|
|
proxy_string = None |
|
236
|
|
|
if isinstance(task_fetch.get('proxy'), six.string_types): |
|
237
|
|
|
proxy_string = task_fetch['proxy'] |
|
238
|
|
|
elif self.proxy and task_fetch.get('proxy', True): |
|
239
|
|
|
proxy_string = self.proxy |
|
240
|
|
|
if proxy_string: |
|
241
|
|
|
if '://' not in proxy_string: |
|
242
|
|
|
proxy_string = 'http://' + proxy_string |
|
243
|
|
|
proxy_splited = urlsplit(proxy_string) |
|
244
|
|
|
fetch['proxy_host'] = proxy_splited.hostname |
|
245
|
|
|
if proxy_splited.username: |
|
246
|
|
|
fetch['proxy_username'] = proxy_splited.username |
|
247
|
|
|
if proxy_splited.password: |
|
248
|
|
|
fetch['proxy_password'] = proxy_splited.password |
|
249
|
|
|
if six.PY2: |
|
250
|
|
|
for key in ('proxy_host', 'proxy_username', 'proxy_password'): |
|
251
|
|
|
if key in fetch: |
|
252
|
|
|
fetch[key] = fetch[key].encode('utf8') |
|
253
|
|
|
fetch['proxy_port'] = proxy_splited.port or 8080 |
|
254
|
|
|
|
|
255
|
|
|
# etag |
|
256
|
|
|
if task_fetch.get('etag', True): |
|
257
|
|
|
_t = None |
|
258
|
|
|
if isinstance(task_fetch.get('etag'), six.string_types): |
|
259
|
|
|
_t = task_fetch.get('etag') |
|
260
|
|
|
elif track_ok: |
|
261
|
|
|
_t = track_headers.get('etag') |
|
262
|
|
|
if _t and 'If-None-Match' not in fetch['headers']: |
|
263
|
|
|
fetch['headers']['If-None-Match'] = _t |
|
264
|
|
|
# last modifed |
|
265
|
|
|
if task_fetch.get('last_modified', task_fetch.get('last_modifed', True)): |
|
266
|
|
|
last_modified = task_fetch.get('last_modified', task_fetch.get('last_modifed', True)) |
|
267
|
|
|
_t = None |
|
268
|
|
|
if isinstance(last_modified, six.string_types): |
|
269
|
|
|
_t = last_modified |
|
270
|
|
|
elif track_ok: |
|
271
|
|
|
_t = track_headers.get('last-modified') |
|
272
|
|
|
if _t and 'If-Modified-Since' not in fetch['headers']: |
|
273
|
|
|
fetch['headers']['If-Modified-Since'] = _t |
|
274
|
|
|
# timeout |
|
275
|
|
|
if 'timeout' in fetch: |
|
276
|
|
|
fetch['request_timeout'] = fetch['timeout'] |
|
277
|
|
|
del fetch['timeout'] |
|
278
|
|
|
# data rename to body |
|
279
|
|
|
if 'data' in fetch: |
|
280
|
|
|
fetch['body'] = fetch['data'] |
|
281
|
|
|
del fetch['data'] |
|
282
|
|
|
|
|
283
|
|
|
return fetch |
|
284
|
|
|
|
|
285
|
|
|
@gen.coroutine |
|
286
|
|
|
def can_fetch(self, user_agent, url): |
|
287
|
|
|
parsed = urlsplit(url) |
|
288
|
|
|
domain = parsed.netloc |
|
289
|
|
|
if domain in self.robots_txt_cache: |
|
290
|
|
|
robot_txt = self.robots_txt_cache[domain] |
|
291
|
|
|
if time.time() - robot_txt.mtime() > self.robot_txt_age: |
|
292
|
|
|
robot_txt = None |
|
293
|
|
|
else: |
|
294
|
|
|
robot_txt = None |
|
295
|
|
|
|
|
296
|
|
|
if robot_txt is None: |
|
297
|
|
|
robot_txt = RobotFileParser() |
|
298
|
|
|
try: |
|
299
|
|
|
response = yield gen.maybe_future(self.http_client.fetch( |
|
300
|
|
|
urljoin(url, '/robots.txt'), connect_timeout=10, request_timeout=30)) |
|
301
|
|
|
content = response.body |
|
302
|
|
|
except tornado.httpclient.HTTPError as e: |
|
303
|
|
|
logger.error('load robots.txt from %s error: %r', domain, e) |
|
304
|
|
|
content = '' |
|
305
|
|
|
|
|
306
|
|
|
try: |
|
307
|
|
|
content = content.decode('utf8', 'ignore') |
|
308
|
|
|
except UnicodeDecodeError: |
|
309
|
|
|
content = '' |
|
310
|
|
|
|
|
311
|
|
|
robot_txt.parse(content.splitlines()) |
|
312
|
|
|
self.robots_txt_cache[domain] = robot_txt |
|
313
|
|
|
|
|
314
|
|
|
raise gen.Return(robot_txt.can_fetch(user_agent, url)) |
|
315
|
|
|
|
|
316
|
|
|
def clear_robot_txt_cache(self): |
|
317
|
|
|
now = time.time() |
|
318
|
|
|
for domain, robot_txt in self.robots_txt_cache.items(): |
|
319
|
|
|
if now - robot_txt.mtime() > self.robot_txt_age: |
|
320
|
|
|
del self.robots_txt_cache[domain] |
|
321
|
|
|
|
|
322
|
|
|
@gen.coroutine |
|
323
|
|
|
def http_fetch(self, url, task): |
|
324
|
|
|
'''HTTP fetcher''' |
|
325
|
|
|
start_time = time.time() |
|
326
|
|
|
self.on_fetch('http', task) |
|
327
|
|
|
handle_error = lambda x: self.handle_error('http', url, task, start_time, x) |
|
328
|
|
|
|
|
329
|
|
|
# setup request parameters |
|
330
|
|
|
fetch = self.pack_tornado_request_parameters(url, task) |
|
331
|
|
|
task_fetch = task.get('fetch', {}) |
|
332
|
|
|
|
|
333
|
|
|
session = cookies.RequestsCookieJar() |
|
334
|
|
|
# fix for tornado request obj |
|
335
|
|
|
if 'Cookie' in fetch['headers']: |
|
336
|
|
|
c = http_cookies.SimpleCookie() |
|
337
|
|
|
try: |
|
338
|
|
|
c.load(fetch['headers']['Cookie']) |
|
339
|
|
|
except AttributeError: |
|
340
|
|
|
c.load(utils.utf8(fetch['headers']['Cookie'])) |
|
341
|
|
|
for key in c: |
|
342
|
|
|
session.set(key, c[key]) |
|
343
|
|
|
del fetch['headers']['Cookie'] |
|
344
|
|
|
if 'cookies' in fetch: |
|
345
|
|
|
session.update(fetch['cookies']) |
|
346
|
|
|
del fetch['cookies'] |
|
347
|
|
|
|
|
348
|
|
|
max_redirects = task_fetch.get('max_redirects', 5) |
|
349
|
|
|
# we will handle redirects by hand to capture cookies |
|
350
|
|
|
fetch['follow_redirects'] = False |
|
351
|
|
|
|
|
352
|
|
|
# making requests |
|
353
|
|
|
while True: |
|
354
|
|
|
# robots.txt |
|
355
|
|
|
if task_fetch.get('robots_txt', False): |
|
356
|
|
|
can_fetch = yield self.can_fetch(fetch['headers']['User-Agent'], fetch['url']) |
|
357
|
|
|
if not can_fetch: |
|
358
|
|
|
error = tornado.httpclient.HTTPError(403, 'Disallowed by robots.txt') |
|
359
|
|
|
raise gen.Return(handle_error(error)) |
|
360
|
|
|
|
|
361
|
|
|
try: |
|
362
|
|
|
request = tornado.httpclient.HTTPRequest(**fetch) |
|
363
|
|
|
# if cookie already in header, get_cookie_header wouldn't work |
|
364
|
|
|
old_cookie_header = request.headers.get('Cookie') |
|
365
|
|
|
if old_cookie_header: |
|
366
|
|
|
del request.headers['Cookie'] |
|
367
|
|
|
cookie_header = cookies.get_cookie_header(session, request) |
|
368
|
|
|
if cookie_header: |
|
369
|
|
|
request.headers['Cookie'] = cookie_header |
|
370
|
|
|
elif old_cookie_header: |
|
371
|
|
|
request.headers['Cookie'] = old_cookie_header |
|
372
|
|
|
except Exception as e: |
|
373
|
|
|
logger.exception(fetch) |
|
374
|
|
|
raise gen.Return(handle_error(e)) |
|
375
|
|
|
|
|
376
|
|
|
try: |
|
377
|
|
|
response = yield gen.maybe_future(self.http_client.fetch(request)) |
|
378
|
|
|
except tornado.httpclient.HTTPError as e: |
|
379
|
|
|
if e.response: |
|
380
|
|
|
response = e.response |
|
381
|
|
|
else: |
|
382
|
|
|
raise gen.Return(handle_error(e)) |
|
383
|
|
|
|
|
384
|
|
|
extract_cookies_to_jar(session, response.request, response.headers) |
|
385
|
|
|
if (response.code in (301, 302, 303, 307) |
|
386
|
|
|
and response.headers.get('Location') |
|
387
|
|
|
and task_fetch.get('allow_redirects', True)): |
|
388
|
|
|
if max_redirects <= 0: |
|
389
|
|
|
error = tornado.httpclient.HTTPError( |
|
390
|
|
|
599, 'Maximum (%d) redirects followed' % task_fetch.get('max_redirects', 5), |
|
391
|
|
|
response) |
|
392
|
|
|
raise gen.Return(handle_error(error)) |
|
393
|
|
|
if response.code in (302, 303): |
|
394
|
|
|
fetch['method'] = 'GET' |
|
395
|
|
|
if 'body' in fetch: |
|
396
|
|
|
del fetch['body'] |
|
397
|
|
|
fetch['url'] = quote_chinese(urljoin(fetch['url'], response.headers['Location'])) |
|
398
|
|
|
fetch['request_timeout'] -= time.time() - start_time |
|
399
|
|
|
if fetch['request_timeout'] < 0: |
|
400
|
|
|
fetch['request_timeout'] = 0.1 |
|
401
|
|
|
max_redirects -= 1 |
|
402
|
|
|
continue |
|
403
|
|
|
|
|
404
|
|
|
result = {} |
|
405
|
|
|
result['orig_url'] = url |
|
406
|
|
|
result['content'] = response.body or '' |
|
407
|
|
|
result['headers'] = dict(response.headers) |
|
408
|
|
|
result['status_code'] = response.code |
|
409
|
|
|
result['url'] = response.effective_url or url |
|
410
|
|
|
result['time'] = time.time() - start_time |
|
411
|
|
|
result['cookies'] = session.get_dict() |
|
412
|
|
|
result['save'] = task_fetch.get('save') |
|
413
|
|
|
if response.error: |
|
414
|
|
|
result['error'] = utils.text(response.error) |
|
415
|
|
|
if 200 <= response.code < 300: |
|
416
|
|
|
logger.info("[%d] %s:%s %s %.2fs", response.code, |
|
417
|
|
|
task.get('project'), task.get('taskid'), |
|
418
|
|
|
url, result['time']) |
|
419
|
|
|
else: |
|
420
|
|
|
logger.warning("[%d] %s:%s %s %.2fs", response.code, |
|
421
|
|
|
task.get('project'), task.get('taskid'), |
|
422
|
|
|
url, result['time']) |
|
423
|
|
|
|
|
424
|
|
|
raise gen.Return(result) |
|
425
|
|
|
|
|
426
|
|
|
@gen.coroutine |
|
427
|
|
|
def phantomjs_fetch(self, url, task): |
|
428
|
|
|
'''Fetch with phantomjs proxy''' |
|
429
|
|
|
start_time = time.time() |
|
430
|
|
|
self.on_fetch('phantomjs', task) |
|
431
|
|
|
handle_error = lambda x: self.handle_error('phantomjs', url, task, start_time, x) |
|
432
|
|
|
|
|
433
|
|
|
# check phantomjs proxy is enabled |
|
434
|
|
|
if not self.phantomjs_proxy: |
|
435
|
|
|
result = { |
|
436
|
|
|
"orig_url": url, |
|
437
|
|
|
"content": "phantomjs is not enabled.", |
|
438
|
|
|
"headers": {}, |
|
439
|
|
|
"status_code": 501, |
|
440
|
|
|
"url": url, |
|
441
|
|
|
"time": time.time() - start_time, |
|
442
|
|
|
"cookies": {}, |
|
443
|
|
|
"save": task.get('fetch', {}).get('save') |
|
444
|
|
|
} |
|
445
|
|
|
logger.warning("[501] %s:%s %s 0s", task.get('project'), task.get('taskid'), url) |
|
446
|
|
|
raise gen.Return(result) |
|
447
|
|
|
|
|
448
|
|
|
# setup request parameters |
|
449
|
|
|
fetch = self.pack_tornado_request_parameters(url, task) |
|
450
|
|
|
task_fetch = task.get('fetch', {}) |
|
451
|
|
|
for each in task_fetch: |
|
452
|
|
|
if each not in fetch: |
|
453
|
|
|
fetch[each] = task_fetch[each] |
|
454
|
|
|
|
|
455
|
|
|
# robots.txt |
|
456
|
|
|
if task_fetch.get('robots_txt', False): |
|
457
|
|
|
user_agent = fetch['headers']['User-Agent'] |
|
458
|
|
|
can_fetch = yield self.can_fetch(user_agent, url) |
|
459
|
|
|
if not can_fetch: |
|
460
|
|
|
error = tornado.httpclient.HTTPError(403, 'Disallowed by robots.txt') |
|
461
|
|
|
raise gen.Return(handle_error(error)) |
|
462
|
|
|
|
|
463
|
|
|
request_conf = { |
|
464
|
|
|
'follow_redirects': False |
|
465
|
|
|
} |
|
466
|
|
|
request_conf['connect_timeout'] = fetch.get('connect_timeout', 20) |
|
467
|
|
|
request_conf['request_timeout'] = fetch.get('request_timeout', 120) + 1 |
|
468
|
|
|
|
|
469
|
|
|
session = cookies.RequestsCookieJar() |
|
470
|
|
|
request = tornado.httpclient.HTTPRequest(url=fetch['url']) |
|
471
|
|
|
if fetch.get('cookies'): |
|
472
|
|
|
session.update(fetch['cookies']) |
|
473
|
|
|
if 'Cookie' in request.headers: |
|
474
|
|
|
del request.headers['Cookie'] |
|
475
|
|
|
fetch['headers']['Cookie'] = cookies.get_cookie_header(session, request) |
|
476
|
|
|
|
|
477
|
|
|
# making requests |
|
478
|
|
|
fetch['headers'] = dict(fetch['headers']) |
|
479
|
|
|
try: |
|
480
|
|
|
request = tornado.httpclient.HTTPRequest( |
|
481
|
|
|
url=self.phantomjs_proxy, method="POST", |
|
482
|
|
|
body=json.dumps(fetch), **request_conf) |
|
483
|
|
|
except Exception as e: |
|
484
|
|
|
raise gen.Return(handle_error(e)) |
|
485
|
|
|
|
|
486
|
|
|
try: |
|
487
|
|
|
response = yield gen.maybe_future(self.http_client.fetch(request)) |
|
488
|
|
|
except tornado.httpclient.HTTPError as e: |
|
489
|
|
|
if e.response: |
|
490
|
|
|
response = e.response |
|
491
|
|
|
else: |
|
492
|
|
|
raise gen.Return(handle_error(e)) |
|
493
|
|
|
|
|
494
|
|
|
if not response.body: |
|
495
|
|
|
raise gen.Return(handle_error(Exception('no response from phantomjs'))) |
|
496
|
|
|
|
|
497
|
|
|
result = {} |
|
498
|
|
|
try: |
|
499
|
|
|
result = json.loads(utils.text(response.body)) |
|
500
|
|
|
assert 'status_code' in result, result |
|
501
|
|
|
except Exception as e: |
|
502
|
|
|
if response.error: |
|
503
|
|
|
result['error'] = utils.text(response.error) |
|
504
|
|
|
raise gen.Return(handle_error(e)) |
|
505
|
|
|
|
|
506
|
|
|
if result.get('status_code', 200): |
|
507
|
|
|
logger.info("[%d] %s:%s %s %.2fs", result['status_code'], |
|
508
|
|
|
task.get('project'), task.get('taskid'), url, result['time']) |
|
509
|
|
|
else: |
|
510
|
|
|
logger.error("[%d] %s:%s %s, %r %.2fs", result['status_code'], |
|
511
|
|
|
task.get('project'), task.get('taskid'), |
|
512
|
|
|
url, result['content'], result['time']) |
|
513
|
|
|
|
|
514
|
|
|
raise gen.Return(result) |
|
515
|
|
|
|
|
516
|
|
|
@gen.coroutine |
|
517
|
|
|
def splash_fetch(self, url, task): |
|
518
|
|
|
'''Fetch with splash''' |
|
519
|
|
|
start_time = time.time() |
|
520
|
|
|
self.on_fetch('splash', task) |
|
521
|
|
|
handle_error = lambda x: self.handle_error('splash', url, task, start_time, x) |
|
522
|
|
|
|
|
523
|
|
|
# check phantomjs proxy is enabled |
|
524
|
|
|
if not self.splash_endpoint: |
|
525
|
|
|
result = { |
|
526
|
|
|
"orig_url": url, |
|
527
|
|
|
"content": "splash is not enabled.", |
|
528
|
|
|
"headers": {}, |
|
529
|
|
|
"status_code": 501, |
|
530
|
|
|
"url": url, |
|
531
|
|
|
"time": time.time() - start_time, |
|
532
|
|
|
"cookies": {}, |
|
533
|
|
|
"save": task.get('fetch', {}).get('save') |
|
534
|
|
|
} |
|
535
|
|
|
logger.warning("[501] %s:%s %s 0s", task.get('project'), task.get('taskid'), url) |
|
536
|
|
|
raise gen.Return(result) |
|
537
|
|
|
|
|
538
|
|
|
# setup request parameters |
|
539
|
|
|
fetch = self.pack_tornado_request_parameters(url, task) |
|
540
|
|
|
task_fetch = task.get('fetch', {}) |
|
541
|
|
|
for each in task_fetch: |
|
542
|
|
|
if each not in fetch: |
|
543
|
|
|
fetch[each] = task_fetch[each] |
|
544
|
|
|
|
|
545
|
|
|
# robots.txt |
|
546
|
|
|
if task_fetch.get('robots_txt', False): |
|
547
|
|
|
user_agent = fetch['headers']['User-Agent'] |
|
548
|
|
|
can_fetch = yield self.can_fetch(user_agent, url) |
|
549
|
|
|
if not can_fetch: |
|
550
|
|
|
error = tornado.httpclient.HTTPError(403, 'Disallowed by robots.txt') |
|
551
|
|
|
raise gen.Return(handle_error(error)) |
|
552
|
|
|
|
|
553
|
|
|
request_conf = { |
|
554
|
|
|
'follow_redirects': False, |
|
555
|
|
|
'headers': { |
|
556
|
|
|
'Content-Type': 'application/json', |
|
557
|
|
|
} |
|
558
|
|
|
} |
|
559
|
|
|
request_conf['connect_timeout'] = fetch.get('connect_timeout', 20) |
|
560
|
|
|
request_conf['request_timeout'] = fetch.get('request_timeout', 120) + 1 |
|
561
|
|
|
|
|
562
|
|
|
session = cookies.RequestsCookieJar() |
|
563
|
|
|
request = tornado.httpclient.HTTPRequest(url=fetch['url']) |
|
564
|
|
|
if fetch.get('cookies'): |
|
565
|
|
|
session.update(fetch['cookies']) |
|
566
|
|
|
if 'Cookie' in request.headers: |
|
567
|
|
|
del request.headers['Cookie'] |
|
568
|
|
|
fetch['headers']['Cookie'] = cookies.get_cookie_header(session, request) |
|
569
|
|
|
|
|
570
|
|
|
# making requests |
|
571
|
|
|
fetch['lua_source'] = self.splash_lua_source |
|
572
|
|
|
fetch['headers'] = dict(fetch['headers']) |
|
573
|
|
|
try: |
|
574
|
|
|
request = tornado.httpclient.HTTPRequest( |
|
575
|
|
|
url=self.splash_endpoint, method="POST", |
|
576
|
|
|
body=json.dumps(fetch), **request_conf) |
|
577
|
|
|
except Exception as e: |
|
578
|
|
|
raise gen.Return(handle_error(e)) |
|
579
|
|
|
|
|
580
|
|
|
try: |
|
581
|
|
|
response = yield gen.maybe_future(self.http_client.fetch(request)) |
|
582
|
|
|
except tornado.httpclient.HTTPError as e: |
|
583
|
|
|
if e.response: |
|
584
|
|
|
response = e.response |
|
585
|
|
|
else: |
|
586
|
|
|
raise gen.Return(handle_error(e)) |
|
587
|
|
|
|
|
588
|
|
|
if not response.body: |
|
589
|
|
|
raise gen.Return(handle_error(Exception('no response from phantomjs'))) |
|
590
|
|
|
|
|
591
|
|
|
result = {} |
|
592
|
|
|
try: |
|
593
|
|
|
result = json.loads(utils.text(response.body)) |
|
594
|
|
|
assert 'status_code' in result, result |
|
595
|
|
|
except ValueError as e: |
|
596
|
|
|
logger.error("result is not json: %r", response.body[:500]) |
|
597
|
|
|
raise gen.Return(handle_error(e)) |
|
598
|
|
|
except Exception as e: |
|
599
|
|
|
if response.error: |
|
600
|
|
|
result['error'] = utils.text(response.error) |
|
601
|
|
|
raise gen.Return(handle_error(e)) |
|
602
|
|
|
|
|
603
|
|
|
if result.get('status_code', 200): |
|
604
|
|
|
logger.info("[%d] %s:%s %s %.2fs", result['status_code'], |
|
605
|
|
|
task.get('project'), task.get('taskid'), url, result['time']) |
|
606
|
|
|
else: |
|
607
|
|
|
logger.error("[%d] %s:%s %s, %r %.2fs", result['status_code'], |
|
608
|
|
|
task.get('project'), task.get('taskid'), |
|
609
|
|
|
url, result['content'], result['time']) |
|
610
|
|
|
|
|
611
|
|
|
raise gen.Return(result) |
|
612
|
|
|
|
|
613
|
|
|
def run(self): |
|
614
|
|
|
'''Run loop''' |
|
615
|
|
|
logger.info("fetcher starting...") |
|
616
|
|
|
|
|
617
|
|
|
def queue_loop(): |
|
618
|
|
|
if not self.outqueue or not self.inqueue: |
|
619
|
|
|
return |
|
620
|
|
|
while not self._quit: |
|
621
|
|
|
try: |
|
622
|
|
|
if self.outqueue.full(): |
|
623
|
|
|
break |
|
624
|
|
|
if self.http_client.free_size() <= 0: |
|
625
|
|
|
break |
|
626
|
|
|
task = self.inqueue.get_nowait() |
|
627
|
|
|
# FIXME: decode unicode_obj should used after data selete from |
|
628
|
|
|
# database, it's used here for performance |
|
629
|
|
|
task = utils.decode_unicode_obj(task) |
|
630
|
|
|
self.fetch(task) |
|
631
|
|
|
except queue.Empty: |
|
632
|
|
|
break |
|
633
|
|
|
except KeyboardInterrupt: |
|
634
|
|
|
break |
|
635
|
|
|
except Exception as e: |
|
636
|
|
|
logger.exception(e) |
|
637
|
|
|
break |
|
638
|
|
|
|
|
639
|
|
|
tornado.ioloop.PeriodicCallback(queue_loop, 100, io_loop=self.ioloop).start() |
|
640
|
|
|
tornado.ioloop.PeriodicCallback(self.clear_robot_txt_cache, 10000, io_loop=self.ioloop).start() |
|
641
|
|
|
self._running = True |
|
642
|
|
|
|
|
643
|
|
|
try: |
|
644
|
|
|
self.ioloop.start() |
|
645
|
|
|
except KeyboardInterrupt: |
|
646
|
|
|
pass |
|
647
|
|
|
|
|
648
|
|
|
logger.info("fetcher exiting...") |
|
649
|
|
|
|
|
650
|
|
|
def quit(self): |
|
651
|
|
|
'''Quit fetcher''' |
|
652
|
|
|
self._running = False |
|
653
|
|
|
self._quit = True |
|
654
|
|
|
self.ioloop.add_callback(self.ioloop.stop) |
|
655
|
|
|
if hasattr(self, 'xmlrpc_server'): |
|
656
|
|
|
self.xmlrpc_ioloop.add_callback(self.xmlrpc_server.stop) |
|
657
|
|
|
self.xmlrpc_ioloop.add_callback(self.xmlrpc_ioloop.stop) |
|
658
|
|
|
|
|
659
|
|
|
def size(self): |
|
660
|
|
|
return self.http_client.size() |
|
661
|
|
|
|
|
662
|
|
|
def xmlrpc_run(self, port=24444, bind='127.0.0.1', logRequests=False): |
|
663
|
|
|
'''Run xmlrpc server''' |
|
664
|
|
|
import umsgpack |
|
665
|
|
|
from pyspider.libs.wsgi_xmlrpc import WSGIXMLRPCApplication |
|
666
|
|
|
try: |
|
667
|
|
|
from xmlrpc.client import Binary |
|
668
|
|
|
except ImportError: |
|
669
|
|
|
from xmlrpclib import Binary |
|
670
|
|
|
|
|
671
|
|
|
application = WSGIXMLRPCApplication() |
|
672
|
|
|
|
|
673
|
|
|
application.register_function(self.quit, '_quit') |
|
674
|
|
|
application.register_function(self.size) |
|
675
|
|
|
|
|
676
|
|
|
def sync_fetch(task): |
|
677
|
|
|
result = self.sync_fetch(task) |
|
678
|
|
|
result = Binary(umsgpack.packb(result)) |
|
679
|
|
|
return result |
|
680
|
|
|
application.register_function(sync_fetch, 'fetch') |
|
681
|
|
|
|
|
682
|
|
|
def dump_counter(_time, _type): |
|
683
|
|
|
return self._cnt[_time].to_dict(_type) |
|
684
|
|
|
application.register_function(dump_counter, 'counter') |
|
685
|
|
|
|
|
686
|
|
|
import tornado.wsgi |
|
687
|
|
|
import tornado.ioloop |
|
688
|
|
|
import tornado.httpserver |
|
689
|
|
|
|
|
690
|
|
|
container = tornado.wsgi.WSGIContainer(application) |
|
691
|
|
|
self.xmlrpc_ioloop = tornado.ioloop.IOLoop() |
|
692
|
|
|
self.xmlrpc_server = tornado.httpserver.HTTPServer(container, io_loop=self.xmlrpc_ioloop) |
|
693
|
|
|
self.xmlrpc_server.listen(port=port, address=bind) |
|
694
|
|
|
logger.info('fetcher.xmlrpc listening on %s:%s', bind, port) |
|
695
|
|
|
self.xmlrpc_ioloop.start() |
|
696
|
|
|
|
|
697
|
|
|
def on_fetch(self, type, task): |
|
698
|
|
|
'''Called before task fetch''' |
|
699
|
|
|
pass |
|
700
|
|
|
|
|
701
|
|
|
def on_result(self, type, task, result): |
|
702
|
|
|
'''Called after task fetched''' |
|
703
|
|
|
status_code = result.get('status_code', 599) |
|
704
|
|
|
if status_code != 599: |
|
705
|
|
|
status_code = (int(status_code) / 100 * 100) |
|
706
|
|
|
self._cnt['5m'].event((task.get('project'), status_code), +1) |
|
707
|
|
|
self._cnt['1h'].event((task.get('project'), status_code), +1) |
|
708
|
|
|
|
|
709
|
|
|
if type in ('http', 'phantomjs') and result.get('time'): |
|
710
|
|
|
content_len = len(result.get('content', '')) |
|
711
|
|
|
self._cnt['5m'].event((task.get('project'), 'speed'), |
|
712
|
|
|
float(content_len) / result.get('time')) |
|
713
|
|
|
self._cnt['1h'].event((task.get('project'), 'speed'), |
|
714
|
|
|
float(content_len) / result.get('time')) |
|
715
|
|
|
self._cnt['5m'].event((task.get('project'), 'time'), result.get('time')) |
|
716
|
|
|
self._cnt['1h'].event((task.get('project'), 'time'), result.get('time')) |
|
717
|
|
|
|
binux /
pyspider
