def remove_ineffective_proxy(proxy):
url = random.choice(validate_urls)
try:
res = requests.get(url, headers=headers, proxies=proxy, timeout=3)
print(proxy, 'ok')
except:
host, port = re.findall(r'//(.+):(\d+)', proxy['http'])[0]
check_str = '{}|{}'.format(host, port)
RedisClient.remove_proxy(check_str)
print(proxy, 'fail')
def multiprocess_filter_proxy():
pool = multiprocessing.Pool(multiprocessing.cpu_count())
print(RedisClient.num_proxy_nofilter())
for _ in range(RedisClient.num_proxy_nofilter()):
item = RedisClient.pop_proxy_nofilter()
ip, port = item.split('|')
pool.apply_async(filter_proxy, args=(ip, port))
pool.close()
pool.join()
print('over')
def filter_proxy(ip, port):
proxy = {
'http': 'http://{}:{}'.format(ip, port),
'https': 'https://{}:{}'.format(ip, port),
}
validate = Validator()
result = validate.validate_proxy(proxy=proxy)
# print(result, ip, port)
if result == 4:
# 将结果存入redis
RedisClient.add_proxy(check_str='{}|{}'.format(ip, port), value=result)
def crawl_66ip():
"""
66ip 代理：http://www.66ip.cn
"""
url = 'http://www.66ip.cn/nmtq.php?getnum=100&isp=0&anonymoustype=4&start=&ports=&export=&ipaddress=&area=0&proxytype=2&api=66ip'
res = redefine_requests(url)
if not res:
return
items = re.findall(pattern=r'\d+\.\d+\.\d+\.\d+\:\d+', string=res.text)
for item in items:
ip, port = item.split(':')
RedisClient.add_proxy_nofilter(check_str='{}|{}'.format(ip, port))
def __init__(self):
self.db = RedisClient()
# self.db.recheck()
# self.db.clear_all()
self.crawlers = [xiciproxy.xiciproxy()
] #,xiciproxy.xiciproxy()_66proxy._66proxy(),
self.check_task = threading.Thread(target=self.check_useful_task)
self.check_task.daemon = True
self.check_task.start()
self.check_task = threading.Thread(target=self.interface)
self.check_task.daemon = True
self.check_task.start()
self.update()
def __init__(self, bind_ip, bind_port, process_id):
self.bind_ip = bind_ip
self.bind_port = bind_port
self.process_id = process_id
self.nid = get_rand_id()
# nodes 节点是一个双端队列
self.nodes = deque(maxlen=MAX_NODE_QSIZE)
# KRPC 协议是由 bencode 编码组成的一个简单的 RPC 结构，使用 UDP 报文发送。
self.udp = socket.socket(socket.AF_INET, socket.SOCK_DGRAM,
socket.IPPROTO_UDP)
# UDP 地址绑定
self.udp.bind((self.bind_ip, self.bind_port))
# redis 客户端
self.rc = RedisClient()
self.logger = get_logger("logger_{}".format(bind_port))
class AsyncAlarmThread(threading.Thread):
def __init__(self, **conf):
threading.Thread.__init__(self)
self._stop = threading.Event()
self.rs = RedisClient(db='alert', **conf)
def stop(self):
self._stop.set()
def run(self):
while not self._stop.isSet():
self.errors = self.rs.hgetall('total:alarm:errors')
message = dict()
for item, errors in self.errors.items():
item = item.split('|')[0]
error = json.loads(errors)
for e in error:
try:
status, info, product, idc, service, black = e.split('|')
if black == 'is_black': continue
title = item
content = '%s,%s,%s,%s,%s' % (status, info, product, idc, service)
content = content.replace(',,,', ',')
except Exception as e:
continue
if title not in message:message[title] = list()
message[title].append(content)
XMPPBOT.MESSAGE = message
for i in xrange(30):
if not self._stop.isSet():
sleep(1)
def crawl_ip3366():
"""
云代理：http://www.ip3366.net
"""
for page in range(1, 6):
url = "http://www.ip3366.net/?stype=1&page={}".format(page)
res = redefine_requests(url)
if not res:
return
soup = BeautifulSoup(res.text, 'lxml')
items = soup.select('#list table tbody tr')
for item in items:
tds = item.select('td')
ip = tds[0].text.strip()
port = tds[1].text.strip()
RedisClient.add_proxy_nofilter(check_str='{}|{}'.format(ip, port))
def crawl_kuaidaili():
"""
快代理：https://www.kuaidaili.com
"""
for page in range(1, 6):
url = "https://www.kuaidaili.com/free/inha/{}/".format(page)
res = redefine_requests(url)
if not res:
continue
soup = BeautifulSoup(res.text, 'lxml')
items = soup.select('#list table tbody tr')
for item in items:
ip = item.find('td', {'data-title': 'IP'}).text.strip()
port = item.find('td', {'data-title': 'PORT'}).text.strip()
RedisClient.add_proxy_nofilter(check_str='{}|{}'.format(ip, port))
def crawl_swei360():
"""
360 代理：http://www.swei360.com
"""
for page in range(1, 2):
for style in [1, 3]:
print(page,style)
url = "http://www.swei360.com/free/?stype={}&page={}".format(style, page)
res = redefine_requests(url,timeout=20)
if not res:
continue
soup = BeautifulSoup(res.text, 'lxml')
items = soup.select('#list table tbody tr')
for item in items:
tds = item.select('td')
ip = tds[0].text.strip()
port = tds[1].text.strip()
RedisClient.add_proxy_nofilter(check_str='{}|{}'.format(ip, port))
def popular():
r = RedisClient.get_instance(dev=False)
pop = r.hgetall('popular')
sorted_searches = sorted(pop.items(), key=lambda x:int(x[1]), reverse=True)[0:10]
final_dict = {}
for sorted_search in sorted_searches:
final_dict[sorted_search[0].decode('utf-8')] = int(sorted_search[1].decode('utf-8'))
return jsonify(final_dict)
def crawl_data5u():
"""
无忧代理：http://www.data5u.com/
"""
urls = ["http://www.data5u.com/free/gwgn/index.shtml",
'http://www.data5u.com/free/gwpt/index.shtml',
'http://www.data5u.com/free/gwpt/index.shtml',
'http://www.data5u.com/free/gwpt/index.shtml',
'http://www.data5u.com/free/index.shtml'
]
for url in urls:
res = redefine_requests(url)
if not res:
return
soup = BeautifulSoup(res.text, 'lxml')
items = soup.find_all('ul', class_='l2')
for item in items:
spans = item.find_all('span')
ip = spans[0].text.strip()
port = spans[1].text.strip()
RedisClient.add_proxy_nofilter(check_str='{}|{}'.format(ip, port))
def crawl_xici():
"""
西刺代理：http://www.xicidaili.com
"""
urls = ['http://www.xicidaili.com/nn/',
'http://www.xicidaili.com/nt/',
'http://www.xicidaili.com/wn/',
'http://www.xicidaili.com/wt/'
]
for url in urls:
for page in range(1, 2):
res =redefine_requests('{}{}'.format(url, page))
if not res:
continue
soup = BeautifulSoup(res.text, 'lxml')
# print(soup)
# time.sleep(1000)
items = soup.select('#ip_list .odd')
for item in items:
tds = item.select('td')
ip = tds[1].text.strip()
port = tds[2].text.strip()
RedisClient.add_proxy_nofilter(check_str='{}|{}'.format(ip, port))
def run():
"""
定期验证代理是否还有效，删除失效的代理
:return:
"""
while True:
pool = multiprocessing.Pool(multiprocessing.cpu_count())
proxies = RedisClient.get_all_proxies()
print(len(proxies))
for proxy in proxies:
pool.apply_async(remove_ineffective_proxy, args=(proxy, ))
pool.close()
pool.join()
time.sleep(10)
def search():
data = json.loads(request.data.decode('utf-8'))
query = data['q']
prefs = data['prefs']
es = ElasticStorage.get_instance(dev=False)
r = RedisClient.get_instance(dev=False)
if r.hexists('popular', query.lower()):
r.hincrby('popular', query.lower())
else:
r.hset('popular', query.lower(), 1)
articles = es.query_articles(query, prefs)
articles = list(articles)
articles = list({article['title']:article for article in articles}.values())
for article in articles:
for key, value in source_map.items():
if key in article['url']:
article['source'] = value
return jsonify(
articles=articles
)
class ProxyPool(object):
MAX_ERROR_NUM = 2
def __init__(self):
self.db = RedisClient()
# self.db.recheck()
# self.db.clear_all()
self.crawlers = [xiciproxy.xiciproxy()
] #,xiciproxy.xiciproxy()_66proxy._66proxy(),
self.check_task = threading.Thread(target=self.check_useful_task)
self.check_task.daemon = True
self.check_task.start()
self.check_task = threading.Thread(target=self.interface)
self.check_task.daemon = True
self.check_task.start()
self.update()
def update(self):
threading.Thread(target=self.get_data).start()
def get_data(self):
for crawler in self.crawlers:
self.db.adds_temp_buffer(crawler.start())
def start_one_check(self, check_buffer, eval_data, type):
assert type == 'http' or type == 'https', 'type 为 http或者https'
if type == 'http':
msgs = [
(requests.get, queue.Queue(), 'http://www.qq.com', {
'timeout': 5,
'proxies': {
'http': eval_data[0]
}
}),
]
check_buffer.append((
eval_data,
msgs[0][1],
'http',
))
else:
msgs = [
(requests.get, queue.Queue(), 'https://www.baidu.com', {
'timeout': 5,
'proxies': {
'https': eval_data[0]
}
}),
]
check_buffer.append((eval_data, msgs[0][1], 'https'))
return msgs
def check_useful_task(self):
check_buffer = []
count = 0
while True:
count += 1
data = self.db.pop_temp_buffer()
if data:
eval_data = eval(data)
if eval_data[4] < self.MAX_ERROR_NUM and eval_data[
5] < self.MAX_ERROR_NUM:
if count // 2:
msgs = self.start_one_check(check_buffer, eval_data,
'http')
else:
msgs = self.start_one_check(check_buffer, eval_data,
'https')
else:
if eval_data[4] < self.MAX_ERROR_NUM:
msgs = self.start_one_check(check_buffer, eval_data,
'http')
elif eval_data[5] < self.MAX_ERROR_NUM:
msgs = self.start_one_check(check_buffer, eval_data,
'https')
else:
continue
downloader(msgs)
else:
for item in check_buffer:
if not item[1].empty():
res = item[1].get()
if res:
res.encoding = res.apparent_encoding
if item[2] == 'http':
if re.findall("<title>腾讯首页</title>", res.text):
item[0][4] = self.MAX_ERROR_NUM
self.db.adds_http_pool(
((*item[0][:4], 0, item[0][6]), ))
else:
item[0][4] += 1
elif item[2] == 'https':
if re.findall("<title>百度一下，你就知道</title>",
res.text):
item[0][5] = self.MAX_ERROR_NUM
self.db.adds_https_pool(
((*item[0][:4], 0, item[0][6]), ))
else:
item[0][5] += 1
else:
if item[2] == 'http':
item[0][4] += 1
elif item[2] == 'https':
item[0][5] += 1
if item[0][4] < self.MAX_ERROR_NUM or item[0][
5] < self.MAX_ERROR_NUM:
self.db.adds_temp_buffer((item[0], ))
check_buffer.remove(item)
time.sleep(0.01)
def interface(self):
while True:
instr = input('>>>')
if instr == 'exit':
sys.exit()
elif instr == 'get_http_one':
print(self.db.get_http_one())
elif instr == 'get_https_one':
print(self.db.get_https_one())
else:
print('请重新输入')
#!/usr/bin/env python
# coding=utf-8
from sanic import Sanic
from sanic.response import json, html
import os
import sys
base_dir = os.path.join(os.path.dirname(os.path.abspath(__file__)))
sys.path.insert(0, base_dir)
from database import RedisClient
app = Sanic()
redis_conn = RedisClient()
@app.route("/")
async def index(request):
return html('<h2>Welcome to Proxy Pool System</h2>')
@app.route("/pop")
async def pop_proxy(request):
proxy = await redis_conn.pop_proxy()
if proxy:
proxy = proxy.decode('utf-8')
if proxy[:5] == "https":
return json({"https": proxy})
else:
return json({"http": proxy})
def __init__(self):
self.redis = RedisClient()
def __init__(self, **conf):
threading.Thread.__init__(self)
self._stop = threading.Event()
self.rs = RedisClient(db='alert', **conf)
class DHTServer:
def __init__(self, bind_ip, bind_port, process_id):
self.bind_ip = bind_ip
self.bind_port = bind_port
self.process_id = process_id
self.nid = get_rand_id()
# nodes 节点是一个双端队列
self.nodes = deque(maxlen=MAX_NODE_QSIZE)
# KRPC 协议是由 bencode 编码组成的一个简单的 RPC 结构，使用 UDP 报文发送。
self.udp = socket.socket(socket.AF_INET, socket.SOCK_DGRAM,
socket.IPPROTO_UDP)
# UDP 地址绑定
self.udp.bind((self.bind_ip, self.bind_port))
# redis 客户端
self.rc = RedisClient()
self.logger = get_logger("logger_{}".format(bind_port))
def bootstrap(self):
"""
利用 tracker 服务器，伪装成 DHT 节点，加入 DHT 网络
"""
for address in BOOTSTRAP_NODES:
self.send_find_node(address)
def bs_timer(self):
"""
定时执行 bootstrap()
"""
t = 1
while True:
if t % PER_SEC_BS_TIMER == 0:
t = 1
self.bootstrap()
t += 1
time.sleep(1)
def send_krpc(self, msg, address):
"""
发送 krpc 协议
:param msg: 发送 UDP 报文信息
:param address: 发送地址，(ip, port) 元组
"""
try:
# msg 要经过 bencode 编码
self.udp.sendto(bencoder.bencode(msg), address)
except:
pass
def send_error(self, tid, address):
"""
发送错误回复
"""
msg = dict(t=tid, y="e", e=[202, "Server Error"])
self.send_krpc(msg, address)
def send_find_node(self, address, nid=None):
"""
发送 find_node 请求。
`find_node 请求`
find_node 被用来查找给定 ID 的节点的联系信息。这时 KPRC 协议中的
"q" == "find_node"。find_node 请求包含 2 个参数，第一个参数是 id，
包含了请求节点的 ID。第二个参数是 target，包含了请求者正在查找的
节点的 ID。当一个节点接收到了 find_node 的请求，他应该给出对应的
回复，回复中包含 2 个关键字 id 和 nodes，nodes 是字符串类型，
包含了被请求节点的路由表中最接近目标节点的 K(8) 个最接近的节点的联系信息。
`示例`
参数: {"id" : "<querying nodes id>", "target" : "<id of target node>"}
回复: {"id" : "<queried nodes id>", "nodes" : "<compact node info>"}
:param address: 地址元组（ip, port)
:param nid: 节点 id
"""
nid = get_neighbor(nid) if nid else self.nid
tid = get_rand_id()
msg = dict(
t=tid,
y="q",
q="find_node", # 指定请求为 find_node
a=dict(id=nid, target=get_rand_id()),
)
self.send_krpc(msg, address)
def send_find_node_forever(self):
"""
循环发送 find_node 请求
"""
self.logger.info("send find node forever...")
while True:
try:
# 弹出一个节点
node = self.nodes.popleft()
self.send_find_node((node.ip, node.port), node.nid)
time.sleep(SLEEP_TIME)
except IndexError:
# 一旦节点队列为空，则重新加入 DHT 网络
self.bootstrap()
def save_magnet(self, info_hash):
"""
将磁力链接保存到数据库
:param info_hash: 磁力链接的 info_hash
"""
# 使用 codecs 解码 info_hash
hex_info_hash = codecs.getencoder("hex")(info_hash)[0].decode()
magnet = MAGNET_PER.format(hex_info_hash)
self.rc.add_magnet(magnet)
# self.logger.info("pid " + str(self.process_id) + " - " + magnet)
self.logger.info("pid_{0} - {1}".format(self.process_id, magnet))
def on_message(self, msg, address):
"""
负责返回信息的处理
:param msg: 报文信息
:param address: 报文地址
"""
try:
# `回复`
# 对应于 KPRC 消息字典中的 y 关键字的值是 r，包含了一个附加的关键字 r。
# 关键字 r 是字典类型，包含了返回的值。发送回复消息是在正确解析了请求消息的
# 基础上完成的。
if msg[b"y"] == b"r":
# nodes 是字符串类型，包含了被请求节点的路由表中最接近目标节点
# 的 K个最接近的节点的联系信息。
if msg[b"r"].get(b"nodes", None):
self.on_find_node_response(msg)
# `请求`
# 对应于 KPRC 消息字典中的 y 关键字的值是 q，它包含 2 个附加的关键字
# q 和 a。关键字 q 是字符串类型，包含了请求的方法名字。关键字 a 一个字典
# 类型包含了请求所附加的参数。
# 而实际上我们只需要获取这两者中的 info hash，用于构造磁力链接进而获取种子。
elif msg[b"y"] == b"q":
# get_peers 与 torrent 文件的 info_hash 有关。这时 KPRC 协议中的
# "q" = "get_peers"。get_peers 请求包含 2 个参数。第一个参数是 id，
# 包含了请求节点的 ID。第二个参数是 info_hash，它代表 torrent 文件的 info_hash
if msg[b"q"] == b"get_peers":
self.on_get_peers_request(msg, address)
# announce_peer 表明请求的节点正在某个端口下载 torrent
# 文件。announce_peer 包含 4 个参数。第一个参数是 id，包含了请求节点的 ID；
# 第二个参数是 info_hash，包含了 torrent 文件的 info_hash；第三个参数是 port
# 包含了整型的端口号，表明 peer 在哪个端口下载；第四个参数数是 token，
# 这是在之前的 get_peers 请求中收到的回复中包含的。
elif msg[b"q"] == b"announce_peer":
self.on_announce_peer_request(msg, address)
except KeyError:
pass
def on_find_node_response(self, msg):
"""
解码 nodes 节点信息，并存储在双端队列
:param msg: 节点报文信息
"""
nodes = get_nodes_info(msg[b"r"][b"nodes"])
for node in nodes:
nid, ip, port = node
# 进行节点有效性判断
if len(nid) != PER_NID_LEN or ip == self.bind_ip:
continue
# 将节点加入双端队列
self.nodes.append(HNode(nid, ip, port))
def on_get_peers_request(self, msg, address):
"""
处理 get_peers 请求，获取 info hash
:param msg: 节点报文信息
:param address: 节点地址
"""
tid = msg[b"t"]
try:
info_hash = msg[b"a"][b"info_hash"]
self.save_magnet(info_hash)
except KeyError:
# 没有对应的 info hash，发送错误回复
self.send_error(tid, address)
def on_announce_peer_request(self, msg, address):
"""
处理 get_announce 请求，获取 info hash，address, port
本爬虫目的暂时只是爬取磁链，所以忽略 address, port 有需要的
开发者可自行完善这部分内容
:param msg: 节点报文信息
:param address: 节点地址
"""
tid = msg[b"t"]
try:
info_hash = msg[b"a"][b"info_hash"]
self.save_magnet(info_hash)
except KeyError:
# 没有对应的 info hash，发送错误回复
self.send_error(tid, address)
def receive_response_forever(self):
"""
循环接受 udp 数据
"""
self.logger.info("receive response forever {}:{}".format(
self.bind_ip, self.bind_port))
# 首先加入到 DHT 网络
self.bootstrap()
while True:
try:
# 接受返回报文
data, address = self.udp.recvfrom(UDP_RECV_BUFFSIZE)
# 使用 bdecode 解码返回数据
msg = bencoder.bdecode(data)
# 处理返回信息
self.on_message(msg, address)
time.sleep(SLEEP_TIME)
except Exception as e:
self.logger.warning(e)
