[Python網路程式設計]gevent httpclient以及網頁編碼
阿新 • • 發佈：2019-01-28
之前看到geventhttpclient這個專案，https://github.com/gwik/geventhttpclient，官方文件說非常快，由於響應使用了C的解析，所以我一直想把這玩意用到專案中，
這兩天一直在糾結這玩意，說實在一句話，比較難用，封裝的不給力，最大缺陷如下：
1.不支援重定向，重定向需要自己來寫，很費事
2.新建的httpclient物件只能傳送同域名的請求
這相當的蛋疼，我花了一點時間封裝了一下，解決了上面的兩個問題，還增加了自動編解碼問題，程式碼如下：
#!/usr/bin/env python #-*-encoding:UTF-8-*- import re from geventhttpclient.url import URL from geventhttpclient.client import HTTPClient,HTTPClientPool from urlparse import urljoin #from core.common import urljoin HEADERS = { 'Accept' : 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8', 'User-Agent': 'Mozilla/5.0 (Windows NT 6.1; WOW64; rv:29.0) Gecko/20100101 Firefox/29.0' } DEFAULF_METHOD = "GET" MAX_REDIRECT_TIME = 10 DEFAULT_PAGE_ENCODING = "utf8" class DifferDomainException(Exception): """ if request different domain url,geventhttpclient will throw it, see gevent.client 'raise ValueError("Invalid host in URL")' """ def __init__(self,uri): self.args = uri self.uri = uri class MaxRedirectException(Exception): def __init__(self,response): self.args = response self.response = response class HTTP(HTTPClient): def request(self,request_uri, method=DEFAULF_METHOD,body=b"", headers={},follow_redirect=True,redirects=MAX_REDIRECT_TIME): if body and method == DEFAULF_METHOD: method = "POST" h = [ k.title() for k in headers.iterkeys() ] headers.update(dict( [ (k,v) for k,v in HEADERS.iteritems() if k not in h ] )) response = super(HTTP,self).request(method, request_uri, body, headers) if follow_redirect and response.status_code in (301,302,303,307) and response.method in ("GET","POST"): if redirects: location = response.get('location') or response.get('content-location') or response.get('uri') if location: location = urljoin(request_uri,location) if not location.startswith(self._base_url_string): raise DifferDomainException(location) return self.request(location, method, body, headers, follow_redirect,redirects-1) else: raise MaxRedirectException(response) return response class HTTPPool(HTTPClientPool): def get_client(self, url): if not isinstance(url, URL): url = URL(url) client_key = url.host, url.port try: return self.clients[client_key] except KeyError: client = HTTP.from_url(url, **self.client_args) self.clients[client_key] = client return client _POLL = HTTPPool(network_timeout=100,connection_timeout=100) META_CHARSET_REGEX = re.compile(r'(?si)<head>.*<meta http-equiv="?content-type"?[^>]+charset=(?P<result>[^">]+).*</head>') def decodePage(content,content_type): httpCharset, metaCharset = None, None if content_type and content_type.find("charset=") != -1: httpCharset = content_type.split("charset=")[-1] match = META_CHARSET_REGEX.search(content) if match: metaCharset = match.group('result') print httpCharset,metaCharset charset = httpCharset or metaCharset or DEFAULT_PAGE_ENCODING return content.decode(charset).encode(DEFAULT_PAGE_ENCODING) def request(request_uri, method=DEFAULF_METHOD,body=b"", headers={},follow_redirect=True,auto_read=True): client = _POLL.get_client(request_uri) response = None try: response = client.request(request_uri,method,body,headers,follow_redirect) except DifferDomainException,e: print "DifferDomainException:"+e.uri response = request(e.uri,method,body,headers,follow_redirect) except MaxRedirectException,e: print "max redirect" response = e.response # will return previous response,of course redirect response except Exception,e: print str(e) if auto_read and response: with response: response.content = decodePage(response.read(),response.get('content-type')) return response def test(): # print request("http://127.0.0.1/re.php",follow_redirect=False) # print request("http://127.0.0.1/re.php",follow_redirect=True).content r=request("http://www.baidu.com/",follow_redirect=False) #baidu utf8 utf8 print r.content[:10] r=request("http://www.163.com/",follow_redirect=False) #163 gbk gb2312 print r.content[:10] test()
在測試網頁編碼問題遇到了一些問題，看下面：
由於頭部的請求先到，所以我們一般認為返回的內容編碼是先根據頭部的，如果沒有再看頁面編碼。
我們看網易的編碼，頭部為gbk，網頁為gb2312，但用gb2312解碼竟然有問題，？？？ 我很不解，各位大大們為啥呢？
但用頭部gbk解碼是正常的，這也證明了頭部編碼優先。按理說網頁編碼是告訴瀏覽器以gb2312顯示，但明顯有問題，瀏覽器怎麼做到的？
我們再看新浪的，這更讓我鬱悶了，誰來拯救我啊？
