Exemplo n.º 1
0
def download(url):
    try:
        r = requests.get(url=url,
                         headers=config.get_header(),
                         timeout=config.TIMEOUT)
        r.encoding = chardet.detect(r.content)['encoding']
        if (not r.ok) or len(r.content) < 500:
            raise ConnectionError
        else:
            return r.text

    except Exception:
        count = 0  # 重试次数

        start = random.choice([num for num in range(1, 100)])
        allIPS = session.query(IP).count()
        start_index = start * allIPS / 100
        end_index = start_index + 30
        if end_index > allIPS:
            end_index = allIPS

        proxylist = session.query(IP).all()[start_index:end_index]

        if not proxylist:
            return None

        while count < config.RETRY_TIME:
            try:
                proxy = random.choice(proxylist)
                ip = proxy.ip
                port = proxy.port
                proxies = {
                    "http": "http://%s:%s" % (ip, port),
                    "https": "http://%s:%s" % (ip, port)
                }

                r = requests.get(url=url,
                                 headers=config.get_header(),
                                 timeout=config.TIMEOUT,
                                 proxies=proxies)
                r.encoding = chardet.detect(r.content)['encoding']
                if (not r.ok) or len(r.content) < 500:
                    raise ConnectionError
                else:
                    return r.text
            except Exception:
                g_count += 1
Exemplo n.º 2
0
def check_and_store_ValidIP(ip, port, type, protocol):
    if check_ip(ip):
        print("ip: %s port: %s type: %d protocol: %d" %
              (ip, port, type, protocol))
        proxy = IP(id=session.query(IP).count() + 1,
                   ip=ip,
                   port=port,
                   types=int(type),
                   protocol=int(protocol))
        session.add(proxy)
        try:
            session.commit()
        except Exception:
            session.rollback()
Exemplo n.º 3
0
def download(url):
    try:
        r = requests.get(url=url,
                         headers=config.get_header(),
                         timeout=config.TIMEOUT)
        r.encoding = chardet.detect(r.content)['encoding']
        if (not r.ok) or len(r.content) < 500:
            raise ConnectionError
        else:
            return r.text

    except Exception:
        count = 0  # 重试次数
        # proxylist = sqlhelper.select(10)
        proxylist = session.query(IP).all()[1:20]

        if not proxylist:
            return None

        while count < config.RETRY_TIME:
            try:
                proxy = random.choice(proxylist)
                ip = proxy.ip
                port = proxy.port
                proxies = {
                    "http": "http://%s:%s" % (ip, port),
                    "https": "http://%s:%s" % (ip, port)
                }

                r = requests.get(url=url,
                                 headers=config.get_header(),
                                 timeout=config.TIMEOUT,
                                 proxies=proxies)
                r.encoding = chardet.detect(r.content)['encoding']
                if (not r.ok) or len(r.content) < 500:
                    raise ConnectionError
                else:
                    return r.text
            except Exception:
                count += 1
Exemplo n.º 4
0
def getfreeip(num):
    return session.query(IP).all()[1:num]