Search code examples
encodingproxypython-3.5brokeraiohttp

Using proxybroker in python3.5 throws an encoding error


I am trying to use proxybroker to generate a file with active proxies for certain countries. I always get the same error trying to fetch the proxies. The error seems to be an encoding/decoding error in a packe used by proxbroker. But I suspect there might be better ways to use proxybroker.

This is the code that causes problems:

def gather_proxies(countries):
"""
This method uses the proxybroker package to asynchronously get two new proxies per specified country
and returns the proxies as a list of country and proxy.

:param countries: The ISO style country codes to fetch proxies for. Countries is a list of two letter strings.
:return: A list of proxies that are themself a list with  two paramters[Location, proxy address].
"""
proxy_list = []
types = ['HTTP']
for country in countries:
    loop = asyncio.get_event_loop()

    proxies = asyncio.Queue(loop=loop)
    broker = Broker(proxies, loop=loop,)

    loop.run_until_complete(broker.find(limit=2, countries=country, types=types))

    while True:
        proxy = proxies.get_nowait()
        if proxy is None:
            break
        print(str(proxy))
        proxy_list.append([country, proxy.host + ":" + str(proxy.port)])
return proxy_list

and the error message:

../app/main/download_thread.py:344: in update_proxies
proxy_list = gather_proxies(country_list)
../app/main/download_thread.py:368: in gather_proxies
    loop.run_until_complete(broker.find(limit=2, countries=country, types=types))
/usr/lib/python3.5/asyncio/base_events.py:387: in run_until_complete
    return future.result()
/usr/lib/python3.5/asyncio/futures.py:274: in result
    raise self._exception
/usr/lib/python3.5/asyncio/tasks.py:241: in _step
    result = coro.throw(exc)
../venv/lib/python3.5/site-packages/proxybroker/api.py:108: in find
    await self._run(self._checker.check_judges(), action)
../venv/lib/python3.5/site-packages/proxybroker/api.py:114: in _run
    await tasks
/usr/lib/python3.5/asyncio/futures.py:361: in __iter__
    yield self  # This tells Task to wait for completion.
/usr/lib/python3.5/asyncio/tasks.py:296: in _wakeup
    future.result()
/usr/lib/python3.5/asyncio/futures.py:274: in result
    raise self._exception
/usr/lib/python3.5/asyncio/tasks.py:241: in _step
    result = coro.throw(exc)
../venv/lib/python3.5/site-packages/proxybroker/checker.py:26: in check_judges
    await asyncio.gather(*[j.check() for j in self._judges])
/usr/lib/python3.5/asyncio/futures.py:361: in __iter__
    yield self  # This tells Task to wait for completion.
/usr/lib/python3.5/asyncio/tasks.py:296: in _wakeup
    future.result()
/usr/lib/python3.5/asyncio/futures.py:274: in result
    raise self._exception
/usr/lib/python3.5/asyncio/tasks.py:239: in _step
    result = coro.send(None)
../venv/lib/python3.5/site-packages/proxybroker/judge.py:62: in check
    page = await resp.text()
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <ClientResponse(http://ip.spys.ru/) [200 OK]>
<CIMultiDictProxy('Date': 'Thu, 18 Aug 2016 11:02:53 GMT', 'Server': 'Ap...': 'no-cache', 'Vary': 'Accept-Encoding', 'Transfer-Encoding': 'chunked', 'Content-Type': 'text/html; charset=UTF-8')>

encoding = 'utf-8'

    @asyncio.coroutine
    def text(self, encoding=None):
        """Read response payload and decode."""
        if self._content is None:
            yield from self.read()
    
        if encoding is None:
            encoding = self._get_encoding()
    
>       return self._content.decode(encoding)
E       UnicodeDecodeError: 'utf-8' codec can't decode byte 0xd6 in position 5568: invalid continuation byte

../venv/lib/python3.5/site-packages/aiohttp/client_reqrep.py:758: UnicodeDecodeError

The problem seems to be within the proxybroker or rather the aiohttp package. But since it is supposedly a tested package the problem is probably my code.

Can anyone see what I did wrong or does anyone have a suggestion regarding the useage of proxybroker?


Solution

  • The problem is in resp.text() call. It retrieves html page as text. aiohttp tries to determine proper encoding using chardet library but for malformed pages it's not possible.

    I think resp.text() should be replaced with resp.read() for fetching page as bytes without decoding to str.