class ValidityTester(object):
test_api = TEST_API
def __init__(self):
self._raw_proxies = None
self._usable_proxies = []
def set_raw_proxies(self, proxies):
self._raw_proxies = proxies
self._conn = RedisClient()
async def test_single_proxy(self, proxy):
"""
text one proxy, if valid, put them to usable_proxies.
"""
try:
async with aiohttp.ClientSession() as session:
try:
if isinstance(proxy, bytes):
proxy = proxy.decode('utf-8')
real_proxy = 'http://' + proxy
print('Testing', proxy)
async with session.get(self.test_api, proxy=real_proxy, timeout=get_proxy_timeout) as response:
if response.status == 200:
self._conn.put(proxy)
print('Valid proxy', proxy)
except (ProxyConnectionError, TimeoutError, ValueError):
print('Invalid proxy', proxy)
except (ServerDisconnectedError, ClientResponseError,ClientConnectorError) as s:
print(s)
pass
def test(self):
"""
aio test all proxies.
"""
print('ValidityTester is working')
try:
if self._raw_proxies:
loop = asyncio.get_event_loop()
tasks = [self.test_single_proxy(proxy) for proxy in self._raw_proxies]
loop.run_until_complete(asyncio.wait(tasks))
except ValueError:
print('Async Error')
class ValidityTester(object):
test_api = TEST_API
def __init__(self):
self._raw_proxies = None
self._usable_proxies = []
def set_raw_proxies(self, proxies):
self._raw_proxies = proxies
self._conn = RedisClient()
async def test_single_proxy(self, proxy):
"""
text one proxy, if valid, put them to usable_proxies.
"""
async with aiohttp.ClientSession() as session:
try:
if isinstance(proxy, bytes):
proxy = proxy.decode('utf-8')
real_proxy = 'http://' + proxy
print('Testing ' + real_proxy)
# 使用百度进行测试，如果能够访问，则说明代理ip可以用
async with session.get(self.test_api,
proxy=real_proxy,
timeout=15) as response:
if response.status == 200:
self._conn.put(proxy)
print('Vaild proxy', proxy)
except (ProxyConnectionError, TimeoutError, ValueError):
print('Invaild proxy', proxy)
def test(self):
"""
aio test all proxies.
"""
try:
loop = asyncio.get_event_loop()
tasks = [
self.test_single_proxy(proxy) for proxy in self._raw_proxies
]
loop.run_until_complete(asyncio.wait(tasks))
except ValueError:
print('asyncio error')
class ValidityTester(object):
test_api = TEST_API
def __init__(self):
self._raw_proxies = None
self._usable_proxies = []
def set_raw_proxies(self, proxies):
self._raw_proxies = proxies
self._conn = RedisClient()
async def test_single_proxy(self, proxy):
"""
text one proxy, if valid, put them to usable_proxies.
"""
try:
async with aiohttp.ClientSession() as session:
try:
if isinstance(proxy, bytes):
proxy = proxy.decode('utf-8')
real_proxy = 'http://' + proxy
print('Testing', proxy)
async with session.get(self.test_api, proxy=real_proxy, timeout=get_proxy_timeout) as response:
if response.status == 200:
self._conn.put(proxy)
print('Valid proxy', proxy)
except (ProxyConnectionError, TimeoutError, ValueError):
print('Invalid proxy', proxy)
except (ServerDisconnectedError, ClientResponseError,ClientConnectorError) as s:
print(s)
pass
def test(self):
"""
aio test all proxies.
"""
print('ValidityTester is working')
try:
loop = asyncio.get_event_loop()
tasks = [self.test_single_proxy(proxy) for proxy in self._raw_proxies]
loop.run_until_complete(asyncio.wait(tasks))
except ValueError:
print('Async Error')
class ValidityTester(object):
test_api = TEST_API
def __init__(self):
self._raw_proxies = None
self._usable_proxies = []
def set_raw_proxies(self, proxies):
self._raw_proxies = proxies
self._conn = RedisClient()
async def test_single_proxy(self, proxy):#定义协程函数
"""
text one proxy, if valid, put them to usable_proxies.
"""
try:
async with aiohttp.ClientSession() as session:#aiohttp异步请求库，相当于s = requests.Session() ，创建一个session对象，然后用session对象去打开网页
try:
if isinstance(proxy, bytes):#判断proxy是否是bytes类型
proxy = proxy.decode('utf-8')
real_proxy = 'http://' + proxy
print('Testing', proxy)
async with session.get(self.test_api, proxy=real_proxy, timeout=get_proxy_timeout) as response:#相当于response=urllib.request.urlopen(url)
if response.status == 200:
self._conn.put(proxy)
print('Valid proxy', proxy)
except (ProxyConnectionError, TimeoutError, ValueError):
print('Invalid proxy', proxy)
except (ServerDisconnectedError, ClientResponseError,ClientConnectorError) as s:
print(s)
pass
def test(self):
"""
aio test all proxies.
"""
print('ValidityTester is working')
loop = asyncio.get_event_loop()#启动协程函数
tasks = [self.test_single_proxy(proxy) for proxy in self._raw_proxies]
loop.run_until_complete(asyncio.wait(tasks))#tasks是一个asyncio.ensure_future(协程函数（参数））的列表，相当于多任务，异步执行tasks里的所有任务
class ApiTestCase(unittest.TestCase):
def setUp(self):
self._app = app.test_client()
self._conn = RedisClient()
def tearDown(self):
self._conn.flush()
def test_get(self):
self._conn.put('aaa')
self._conn.put('bbb')
r = self._app.get('/get')
assert 'aaa' in str(r.data)
r = self._app.get('/get')
assert 'bbb' in str(r.data)
def test_count(self):
self._conn.put('aaa')
self._conn.put('bbb')
r = self._app.get('/count')
assert '2' in str(r.data)
self._conn.put('ccc')
self._conn.put('ddd')
r = self._app.get('/count')
assert '4' in str(r.data)
proxy = self._conn.pop()
r = self._app.get('/count')
assert '3' in str(r.data)
class RedisClientTestCase(unittest.TestCase):
def setUp(self):
self._conn = RedisClient()
def tearDown(self):
self._conn.flush()
def test_put_and_pop(self):
self._conn.put("label")
assert self._conn.pop() == "label"
def test_put_many(self):
self._conn.put_many(['a', 'b'])
assert self._conn.pop() == "a"
assert self._conn.pop() == "b"
def test_len(self):
self._conn.put_many(['a', 'b', 'c'])
assert self._conn.queue_len == 3
def test_get(self):
self._conn.put_many(['a', 'b', 'c', 'd'])
_ = self._conn.get(2)
assert self._conn.queue_len == 2
class ValidityTester(object):
test_api = TEST_API
def __init__(self):
self._raw_proxies = None # 原始代理
self._usable_proxies = [] # 可用的代理
def set_raw_proxies(self, proxies):
self._raw_proxies = proxies # 把接收到的proxies赋值给实例属性让内部使用
self._conn = RedisClient() # 创建redis连接
# 测试单个代理,async设置为异步函数(协程)
async def test_single_proxy(self, proxy):
"""
text one proxy, if valid, put them to usable_proxies.
"""
try:
# 创建一个会话对象session发请求
async with aiohttp.ClientSession() as session:
try:
# 设置代理为utf-8
if isinstance(proxy, bytes):
proxy = proxy.decode('utf-8')
# 拼接代理地址
real_proxy = 'http://' + proxy
print('Testing', proxy)
# 用session发送get请求,用百度测试,设置代理检测
async with session.get(
self.test_api,
proxy=real_proxy,
timeout=get_proxy_timeout) as response:
# 如果响应状态码是200,说明代理可用,加入到代理队列右边
if response.status == 200:
self._conn.put(proxy)
print('Valid proxy', proxy)
# 如果出现未连接成功,超时,值错误,那么就打印提示信息
except (ProxyConnectionError, TimeoutError, ValueError):
print('Invalid proxy', proxy)
# 捕获里面异常
except (ServerDisconnectedError, ClientResponseError,
ClientConnectorError) as s:
print(s)
pass
# 大量检测proxy
def test(self):
"""
aio test all proxies.
"""
print('ValidityTester is working')
try:
# 协程不能直接运行,需要创建一个事件循环
loop = asyncio.get_event_loop()
# 任务对象列表,把每个代理交给检测方法,生成一个任务对象,把这些任务对象放到一个任务对象列表中
tasks = [
self.test_single_proxy(proxy) for proxy in self._raw_proxies
]
# 调用run_until_complete方法,将协程注册到事件循环中,并启动事件循环
loop.run_until_complete(asyncio.wait(tasks))
except ValueError:
print('Async Error')
class ValidityTester(object):
test_api = TEST_API
def __init__(self):
self._raw_proxies = None
self._usable_proxies = []
def set_raw_proxies(self, proxies):
self._raw_proxies = proxies
self._conn = RedisClient()
# async def test_single_proxy(self, proxy):
# """
# text one proxy, if valid, put them to usable_proxies.
# """
# try:
# async with aiohttp.ClientSession() as session:
# try:
# if isinstance(proxy, bytes):
# proxy = proxy.decode('utf-8')
# print('正在测试:', proxy)
# async with session.get(self.test_api, proxy=proxy, timeout=get_proxy_timeout) as response:
# if response.status == 200:
# self._conn.put(proxy)
# print('可用代理:', proxy)
# except Exception as e:
# print(e)
# print('不可用代理:', proxy)
# except (ServerDisconnectedError, ClientResponseError,ClientConnectorError) as s:
# print(s)
# pass
def test_single_proxy(self, proxy):
"""
text one proxy, if valid, put them to usable_proxies.
"""
try:
try:
if isinstance(proxy, bytes):
proxy = proxy.decode('utf-8')
# real_proxy = 'http://' + proxy
print('正在测试:', proxy)
# async with session.get(self.test_api, proxy=real_proxy, timeout=get_proxy_timeout) as response:
pro = re.split(r'[\:\/\/]', proxy)
proxy_dict = {pro[0]: pro[3] + ':' + pro[4]}
response = requests.get(url=TEST_API,
proxies=proxy_dict,
timeout=5)
if response.status_code == 200:
self._conn.put(proxy)
print('可用代理:', proxy)
except Exception as e:
print(e)
print('不可用代理:', proxy)
except (ServerDisconnectedError, ClientResponseError,
ClientConnectorError) as s:
print(s)
pass
def test(self):
"""
aio test all proxies.
"""
print('进行代理可用性测试:')
try:
# loop = asyncio.get_event_loop()
# tasks = [self.test_single_proxy(proxy) for proxy in self._raw_proxies]
# loop.run_until_complete(asyncio.wait(tasks))
for proxy in self._raw_proxies:
self.test_single_proxy(proxy)
except Exception as e:
print(e)
print('Async Error!')
class ValidityTester(object):
test_api = TEST_API
def __init__(self):
self._raw_proxies = None
self._usable_proxies = []
def set_raw_proxies(self, proxies):
self._raw_proxies = proxies
self._conn = RedisClient()
def set_timing_params(self):
self._conn = RedisClient()
self._all_ips_item = self._conn.getAll() #把现在所有的ip列表都拿出来做检查
self._post_url = ALIE_API
async def test_single_proxy(self, proxy):
"""
text one proxy, if valid, put them to usable_proxies.
"""
try:
async with aiohttp.ClientSession() as session:
try:
if isinstance(proxy, bytes):
proxy = proxy.decode('utf-8')
real_proxy = 'http://' + proxy
print('Testing', proxy)
async with session.get(
self.test_api,
proxy=real_proxy,
timeout=get_proxy_timeout) as response:
if response.status == 200:
self._conn.put(proxy)
print('Valid proxy', proxy)
except (ProxyConnectionError, TimeoutError, ValueError):
print('Invalid proxy', proxy)
except (ServerDisconnectedError, ClientResponseError,
ClientConnectorError) as s:
print(s)
pass
def test(self):
"""
aio test all proxies.
"""
print('ValidityTester is working')
try:
loop = asyncio.get_event_loop()
tasks = [
self.test_single_proxy(proxy) for proxy in self._raw_proxies
] #test_single_proxy 检验ip是否有效
loop.run_until_complete(asyncio.wait(tasks))
#loop.run_until_complete(asyncio.gather(self.test_single_proxy(proxy) for proxy in self._raw_proxies))
except ValueError:
print('Async Error')
async def TimingCheckFunction(self, proxy):
try:
async with aiohttp.ClientSession() as session:
try:
if isinstance(proxy, bytes): #bytes=str
proxy = proxy.decode('utf-8')
real_proxy = 'http://' + proxy
headers = {'User-Agent': choice(AGENTS)}
print('Timing Check Async Ip:' + str(proxy))
async with session.get(self._post_url,
proxy=real_proxy,
timeout=get_proxy_timeout,
headers=headers) as response:
if (response.status != 200):
self._conn.delete(proxy)
print('Delete Old Invalid Proxy', proxy)
else:
print('Keep Save IP', proxy)
except (ProxyConnectionError, TimeoutError, ValueError):
print('Foreach Delete Invalid Proxy Error', proxy)
self._conn.delete(proxy)
except (ServerDisconnectedError, ClientResponseError,
ClientConnectorError) as s:
print('-------')
print(s)
#self._conn.delete(proxy)
pass
def TimingCheck(self):
try:
loop = asyncio.get_event_loop()
tasks = [
self.TimingCheckFunction(proxy) for proxy in self._all_ips_item
] #test_single_proxy 检验ip是否有效
loop.run_until_complete(asyncio.wait(tasks))
except ValueError:
print('Timing Check Error')
class ValidityTester(object):
"""代理有效性测试器"""
# 测试用的url
test_api = TEST_API
def __init__(self):
# 为测试有效性的代理ip
self._raw_proxies = None
# 可用的代理ip
self._usable_proxies = []
def set_raw_proxies(self, proxies):
"""
设置为测试有效性的代理ip列表
:param proxies:代理ip列表
:return:
"""
self._raw_proxies = proxies
self._conn = RedisClient()
async def test_single_proxy(self, proxy):
"""
测试一个代理，如果有效就添加到_usable_proxies列表中
:param proxy: 代理
:return:
"""
try:
async with aiohttp.ClientSession() as session:
try:
if isinstance(proxy, bytes):
proxy = proxy.decode('utf-8')
real_proxy = 'http://' + proxy
print('Testing', proxy)
async with session.get(
self.test_api,
proxy=real_proxy,
timeout=get_proxy_timeout) as response:
if response.status == 200:
self._conn.put(proxy)
print('Valid proxy', proxy)
except (ProxyConnectionError, TimeoutError, ValueError,
ClientOSError):
print('Invalid proxy', proxy)
except (ServerDisconnectedError, ClientResponseError,
ClientConnectorError) as s:
print('Invalid proxy', proxy)
def test(self):
"""
异步测试所用代理
"""
print('ValidityTester is working')
try:
loop = asyncio.get_event_loop()
tasks = [
self.test_single_proxy(proxy) for proxy in self._raw_proxies
]
loop.run_until_complete(asyncio.wait(tasks))
except ValueError:
print('Async Error')
