def __init__(self, web_proxy_list=[], sustain=False, timeout=5, protocol=Protocol.HTTP):
self.userAgent = UserAgentManager()
self.logger = logging.getLogger()
self.logger.addHandler(handler)
self.logger.setLevel(0)
#####
# Each of the classes below implements a specific URL Parser
#####
parsers = list([])
parsers.append(FreeProxyParser('FreeProxy', 'http://free-proxy-list.net', timeout=timeout))
#parsers.append(ProxyForEuParser('ProxyForEU', 'http://proxyfor.eu/geo.php', 1.0, timeout=timeout)) <--doesn't work anymore
#parsers.append(RebroWeeblyParser('ReBro', 'http://rebro.weebly.com', timeout=timeout)) <--doesn't work anymore
parsers.append(PremProxyParser('PremProxy', 'https://premproxy.com', timeout=timeout))
self.logger.debug("=== Initialized Proxy Parsers ===")
for i in range(len(parsers)):
self.logger.debug("\t {0}".format(parsers[i].__str__()))
self.logger.debug("=================================")
self.sustain = sustain
self.parsers = parsers
self.proxy_list = web_proxy_list
for parser in parsers:
try:
size = len(self.proxy_list)
self.proxy_list += parser.parse_proxyList()
self.logger.debug('Added {} proxies from {}'.format(len(self.proxy_list)-size, parser.id))
except ReadTimeout:
self.logger.warning("Proxy Parser: '{}' TimedOut!".format(parser.url))
self.logger.debug('Total proxies = '+str(len(self.proxy_list)))
# filtering the list of available proxies according to user preferences
self.proxy_list = [p for p in self.proxy_list if protocol in p.protocols]
self.logger.debug('Filtered proxies = '+str(len(self.proxy_list)))
self.current_proxy = self.randomize_proxy()
def __init__(self, web_proxy_list=[], sustain=False, timeout=5):
self.userAgent = UserAgentManager()
self.logger = logging.getLogger()
self.logger.addHandler(handler)
self.logger.setLevel(0)
#####
# Each of the classes below implements a specific URL Parser
#####
parsers = list([])
parsers.append(FreeProxyParser('FreeProxy', 'http://free-proxy-list.net', timeout=timeout))
parsers.append(ProxyForEuParser('ProxyForEU', 'http://proxyfor.eu/geo.php', 1.0, timeout=timeout))
parsers.append(RebroWeeblyParser('ReBro', 'http://rebro.weebly.com', timeout=timeout))
parsers.append(SamairProxyParser('Samair', 'https://premproxy.com', timeout=timeout))
self.logger.debug("=== Initialized Proxy Parsers ===")
for i in range(len(parsers)):
self.logger.debug("\t {0}".format(parsers[i].__str__()))
self.logger.debug("=================================")
self.sustain = sustain
self.parsers = parsers
self.proxy_list = web_proxy_list
for i in range(len(parsers)):
try:
self.proxy_list += parsers[i].parse_proxyList()
except ReadTimeout:
self.logger.warning("Proxy Parser: '{}' TimedOut!".format(parsers[i].url))
self.current_proxy = self.randomize_proxy()
class TestBaseProxyParsers(unittest.TestCase):
def setUp(self):
agentsfile = os.path.join(
os.path.dirname(__file__),
'../http_request_randomizer/requests/data/user_agents.txt')
self.uafile = UserAgentManager(file=agentsfile)
self.uafake = UserAgentManager()
def test_agent_size(self):
self.assertTrue(self.uafile.get_len_user_agent() >= 899)
self.assertIsNone(self.uafake.get_len_user_agent())
def test_fist_user_agent(self):
expected = "Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.1; WOW64; Trident/6.0"
self.assertEqual(self.uafile.get_first_user_agent(), expected)
self.assertIsNone(self.uafake.get_first_user_agent())
def test_last_user_agent(self):
expected = "Opera/9.80 (Windows NT 5.1; U; ru) Presto/2.2.15 Version/10.0"
self.assertEqual(self.uafile.get_last_user_agent(), expected)
self.assertIsNone(self.uafake.get_last_user_agent())
def test_random_user_agent(self):
count = 0
for i in range(1, 101):
if self.uafile.get_random_user_agent(
) == self.uafile.get_random_user_agent():
count = count + 1
self.assertNotEqual(count, i)
class TestBaseProxyParsers(unittest.TestCase):
def setUp(self):
self.ua = UserAgentManager()
def test_agent_size(self):
self.assertTrue(self.ua.get_len_user_agent() >= 899)
def test_fist_user_agent(self):
expected = "Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.1; WOW64; Trident/6.0"
self.assertEquals(self.ua.get_first_user_agent(), expected)
def test_last_user_agent(self):
expected = "Opera/9.80 (Windows NT 5.1; U; ru) Presto/2.2.15 Version/10.0"
self.assertEquals(self.ua.get_last_user_agent(), expected)
def test_random_user_agent(self):
self.assertNotEqual(self.ua.get_random_user_agent(), self.ua.get_random_user_agent())
def __init__(self, web_proxy_list=[], sustain=False, timeout=5):
self.userAgent = UserAgentManager()
self.logger = logging.getLogger()
self.logger.addHandler(handler)
self.logger.setLevel(0)
#####
# Each of the classes below implements a specific URL Parser
#####
parsers = list([])
parsers.append(FreeProxyParser('FreeProxy', 'http://free-proxy-list.net', timeout=timeout))
parsers.append(ProxyForEuParser('ProxyForEU', 'http://proxyfor.eu/geo.php', 1.0, timeout=timeout))
parsers.append(RebroWeeblyParser('ReBro', 'http://rebro.weebly.com', timeout=timeout))
parsers.append(SamairProxyParser('Samair', 'https://premproxy.com', timeout=timeout))
self.logger.debug("=== Initialized Proxy Parsers ===")
class RequestProxy:
def __init__(self,
web_proxy_list=[],
sustain=False,
timeout=5,
protocol=Protocol.HTTP,
log_level=0):
self.logger = logging.getLogger()
self.logger.addHandler(handler)
self.logger.setLevel(log_level)
self.userAgent = UserAgentManager(file=os.path.join(
os.path.dirname(__file__), '../data/user_agents.txt'))
#####
# Each of the classes below implements a specific URL Parser
#####
parsers = list([])
parsers.append(
FreeProxyParser('FreeProxy',
'http://free-proxy-list.net',
timeout=timeout))
#parsers.append(ProxyForEuParser('ProxyForEU', 'http://proxyfor.eu/geo.php', 1.0, timeout=timeout)) <--doesn't work anymore
#parsers.append(RebroWeeblyParser('ReBro', 'http://rebro.weebly.com', timeout=timeout)) <--doesn't work anymore
parsers.append(
PremProxyParser('PremProxy',
'https://premproxy.com',
timeout=timeout))
parsers.append(
SslProxyParser('SslProxy',
'https://www.sslproxies.org',
timeout=timeout))
self.logger.debug("=== Initialized Proxy Parsers ===")
for i in range(len(parsers)):
self.logger.debug("\t {0}".format(parsers[i].__str__()))
self.logger.debug("=================================")
self.sustain = sustain
self.parsers = parsers
self.proxy_list = web_proxy_list
for parser in parsers:
try:
size = len(self.proxy_list)
self.proxy_list += parser.parse_proxyList()
self.logger.debug('Added {} proxies from {}'.format(
len(self.proxy_list) - size, parser.id))
except ReadTimeout:
self.logger.warning("Proxy Parser: '{}' TimedOut!".format(
parser.url))
self.logger.debug('Total proxies = ' + str(len(self.proxy_list)))
# filtering the list of available proxies according to user preferences
self.proxy_list = [
p for p in self.proxy_list if protocol in p.protocols
]
self.logger.debug('Filtered proxies = ' + str(len(self.proxy_list)))
self.current_proxy = self.randomize_proxy()
def set_logger_level(self, level):
self.logger.setLevel(level)
def get_proxy_list(self):
return self.proxy_list
def generate_random_request_headers(self):
headers = {
"Connection": "close", # another way to cover tracks
"User-Agent": self.userAgent.get_random_user_agent()
} # select a random user agent
return headers
def randomize_proxy(self):
if len(self.proxy_list) == 0:
raise ProxyListException("list is empty")
rand_proxy = random.choice(self.proxy_list)
while not rand_proxy:
rand_proxy = random.choice(self.proxy_list)
self.current_proxy = rand_proxy
return rand_proxy
#####
# Proxy format:
# http://<USERNAME>:<PASSWORD>@<IP-ADDR>:<PORT>
#####
def generate_proxied_request(self,
url,
method="GET",
params={},
data={},
headers={},
req_timeout=30):
try:
random.shuffle(self.proxy_list)
# req_headers = dict(params.items() + self.generate_random_request_headers().items())
req_headers = dict(params.items())
req_headers_random = dict(
self.generate_random_request_headers().items())
req_headers.update(req_headers_random)
if not self.sustain:
self.randomize_proxy()
headers.update(req_headers)
self.logger.debug("Using headers: {0}".format(str(headers)))
self.logger.debug("Using proxy: {0}".format(str(
self.current_proxy)))
request = requests.request(
method,
url,
headers=headers,
data=data,
params=params,
timeout=req_timeout,
proxies={
"http":
"http://{0}".format(self.current_proxy.get_address()),
"https":
"https://{0}".format(self.current_proxy.get_address())
})
# Avoid HTTP request errors
if request.status_code == 409:
raise ConnectionError(
"HTTP Response [409] - Possible Cloudflare DNS resolution error"
)
elif request.status_code == 403:
raise ConnectionError(
"HTTP Response [403] - Permission denied error")
elif request.status_code == 503:
raise ConnectionError(
"HTTP Response [503] - Service unavailable error")
self.logger.info('RR Status {}'.format(request.status_code))
return request
except ConnectionError:
try:
self.proxy_list.remove(self.current_proxy)
except ValueError:
pass
self.logger.debug(
"Proxy unreachable - Removed Straggling proxy: {0} PL Size = {1}"
.format(self.current_proxy, len(self.proxy_list)))
self.randomize_proxy()
except ReadTimeout:
try:
self.proxy_list.remove(self.current_proxy)
except ValueError:
pass
self.logger.debug(
"Read timed out - Removed Straggling proxy: {0} PL Size = {1}".
format(self.current_proxy, len(self.proxy_list)))
self.randomize_proxy()
except ChunkedEncodingError:
try:
self.proxy_list.remove(self.current_proxy)
except ValueError:
pass
self.logger.debug(
"Wrong server chunked encoding - Removed Straggling proxy: {0} PL Size = {1}"
.format(self.current_proxy, len(self.proxy_list)))
self.randomize_proxy()
except TooManyRedirects:
try:
self.proxy_list.remove(self.current_proxy)
except ValueError:
pass
self.logger.debug(
"Too many redirects - Removed Straggling proxy: {0} PL Size = {1}"
.format(self.current_proxy, len(self.proxy_list)))
self.randomize_proxy()
class RequestProxy:
def __init__(self, web_proxy_list=[], sustain=False, timeout=5):
self.userAgent = UserAgentManager()
self.logger = logging.getLogger()
self.logger.addHandler(handler)
self.logger.setLevel(0)
#####
# Each of the classes below implements a specific URL Parser
#####
parsers = list([])
parsers.append(
FreeProxyParser('http://free-proxy-list.net', timeout=timeout))
parsers.append(
ProxyForEuParser('http://proxyfor.eu/geo.php',
1.0,
timeout=timeout))
parsers.append(
RebroWeeblyParser('http://rebro.weebly.com', timeout=timeout))
parsers.append(
SamairProxyParser('http://samair.ru/proxy/time-01.htm',
timeout=timeout))
self.logger.debug("=== Initialized Proxy Parsers ===")
for i in range(len(parsers)):
self.logger.debug("\t {0}".format(parsers[i].__str__()))
self.logger.debug("=================================")
self.sustain = sustain
self.parsers = parsers
self.proxy_list = web_proxy_list
for i in range(len(parsers)):
try:
self.proxy_list += parsers[i].parse_proxyList()
except ReadTimeout:
self.logger.warn("Proxy Parser: '{}' TimedOut!".format(
parsers[i].url))
self.current_proxy = self.randomize_proxy()
def set_logger_level(self, level):
self.logger.setLevel(level)
def get_proxy_list(self):
return self.proxy_list
def current_proxy_ip(self):
return str(self.current_proxy)
def generate_random_request_headers(self):
headers = {
"Connection": "close", # another way to cover tracks
"User-Agent": self.userAgent.get_random_user_agent()
} # select a random user agent
return headers
def randomize_proxy(self):
if len(self.proxy_list) == 0:
raise ProxyListException("list is empty")
rand_proxy = random.choice(self.proxy_list)
while not rand_proxy:
rand_proxy = random.choice(self.proxy_list)
self.current_proxy = rand_proxy
return rand_proxy
#####
# Proxy format:
# http://<USERNAME>:<PASSWORD>@<IP-ADDR>:<PORT>
#####
def generate_proxied_request(self,
url,
method="GET",
params={},
data={},
headers={},
req_timeout=30):
try:
random.shuffle(self.proxy_list)
# req_headers = dict(params.items() + self.generate_random_request_headers().items())
req_headers = dict(params.items())
req_headers_random = dict(
self.generate_random_request_headers().items())
req_headers.update(req_headers_random)
if not self.sustain:
self.randomize_proxy()
headers.update(req_headers)
self.logger.debug("Using proxy: {0}".format(str(
self.current_proxy)))
request = requests.request(method,
url,
proxies={"http": self.current_proxy},
headers=headers,
data=data,
params=params,
timeout=req_timeout)
# Avoid HTTP request errors
if request.status_code == 409:
raise ConnectionError(
"HTTP Response [409] - Possible Cloudflare DNS resolution error"
)
elif request.status_code == 403:
raise ConnectionError(
"HTTP Response [403] - Permission denied error")
elif request.status_code == 503:
raise ConnectionError(
"HTTP Response [503] - Service unavailable error")
print('RR Status {}'.format(request.status_code))
return request
except ConnectionError:
try:
self.proxy_list.remove(self.current_proxy)
except ValueError:
pass
self.logger.debug(
"Proxy unreachable - Removed Straggling proxy: {0} PL Size = {1}"
.format(self.current_proxy, len(self.proxy_list)))
self.randomize_proxy()
except ReadTimeout:
try:
self.proxy_list.remove(self.current_proxy)
except ValueError:
pass
self.logger.debug(
"Read timed out - Removed Straggling proxy: {0} PL Size = {1}".
format(self.current_proxy, len(self.proxy_list)))
self.randomize_proxy()
except ChunkedEncodingError:
try:
self.proxy_list.remove(self.current_proxy)
except ValueError:
pass
self.logger.debug(
"Wrong server chunked encoding - Removed Straggling proxy: {0} PL Size = {1}"
.format(self.current_proxy, len(self.proxy_list)))
self.randomize_proxy()
def setUp(self):
agentsfile = os.path.join(
os.path.dirname(__file__),
'../http_request_randomizer/requests/data/user_agents.txt')
self.uafile = UserAgentManager(file=agentsfile)
self.uafake = UserAgentManager()
def setUp(self):
self.ua = UserAgentManager()
