python爬虫如何POST request payload形式的请求

1. 背景

• 最近在爬取某个站点时，发现在POST数据时，使用的数据格式是request payload，有别于之前常见的 POST数据格式（Form data）。而使用Form data数据的提交方式时，无法提交成功。
  
• 于是上网查了下二者的区别：http://xiaobaoqiu.github.io/blog/2014/09/04/form-data-vs-request-payload/，下面做了搬运工（侵权立删…）

1.1. Http请求中Form Data 和 Request Payload的区别

• AJAX Post请求中常用的两种传参数的形式：form data 和 request payload

1.1.1. Form data

• get请求的时候，我们的参数直接反映在url里面，形式为key1=value1&key2=value2形式，比如：

http://news.baidu.com/ns?word=NBA&tn=news&from=news&cl=2&rn=20&ct=1

• 而如果是post请求，那么表单参数是在请求体中，也是以key1=value1&key2=value2的形式在请求体中。通过chrome的开发者工具可以看到，如下：

RequestURL:http://127.0.0.1:8080/test/test.do Request Method:POST Status Code:200 OK

Request Headers Accept:text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8 Accept-Encoding:gzip,deflate,sdch Accept-Language:zh-CN,zh;q=0.8,en;q=0.6 AlexaToolbar-ALX_NS_PH:AlexaToolbar/alxg-3.2 Cache-Control:max-age=0 Connection:keep-alive Content-Length:25 Content-Type:application/x-www-form-urlencoded Cookie:JSESSIONID=74AC93F9F572980B6FC10474CD8EDD8D Host:127.0.0.1:8080 Origin:http://127.0.0.1:8080 Referer:http://127.0.0.1:8080/test/index.jsp User-Agent:Mozilla/5.0 (Windows NT 6.1)AppleWebKit/537.36 (KHTML, like Gecko) Chrome/33.0.1750.149 Safari/537.36

Form Data name:mikan address:street

Response Headers Content-Length:2 Date:Sun, 11 May 2014 11:05:33 GMT Server:Apache-Coyote/1.1

• 这里要注意post请求的Content-Type为application/x-www-form-urlencoded（默认的），参数是在请求体中，即上面请求中的Form Data。

• 前端代码：提交数据

xhr.setRequestHeader("Content-type","application/x-www-form-urlencoded"); xhr.send("name=foo&value=bar");

• 后端代码：接收提交的数据。在servlet中，可以通过request.getParameter(name)的形式来获取表单参数。

/** * 获取httpRequest的参数 * * @param request * @param name * @return */ protected String getParameterValue(HttpServletRequest request, String name) { return StringUtils.trimToEmpty(request.getParameter(name)); }

1.1.2. Request payload

• 如果使用原生AJAX POST请求的话，那么请求在chrome的开发者工具的表现如下，主要是参数在

Remote Address:192.168.234.240:80 Request URL:http://tuanbeta3.XXX.com/qimage/upload.htm Request Method:POST Status Code:200 OK

Request Headers Accept:application/json, text/javascript, */*; q=0.01 Accept-Encoding:gzip,deflate,sdch Accept-Language:zh-CN,zh;q=0.8,en;q=0.6 Connection:keep-alive Content-Length:151 Content-Type:application/json;charset=UTF-8 Cookie:JSESSIONID=E08388788943A651924CA0A10C7ACAD0 Host:tuanbeta3.XXX.com Origin:http://tuanbeta3.XXX.com Referer:http://tuanbeta3.XXX.com/qimage/customerlist.htm?menu=19 User-Agent:Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/35.0.1916.114 Safari/537.36 X–Requested–With:XMLHttpRequest

Request Payload [{widthEncode:NNNcaXN, heightEncode:NNNN5NN, displayUrl:201409/03/66I5P266rtT86oKq6,…}]

Response Headers Connection:keep–alive Content–Encoding:gzip Content–Type:application/json;charset=UTF-8 Date:Thu, 04 Sep 2014 06:49:44 GMT Server:nginx/1.4.7 Transfer–Encoding:chunked Vary:Accept–Encoding

• 注意请求的Content-Type是application/json;charset=UTF-8，而请求表单的参数在Request Payload中。
• 后端代码：获取数据（这里使用org.apache.commons.io.）：

/** * 从 request 获取 payload 数据 * * @param request * @return * @throws IOException */ private String getRequestPayload(HttpServletRequest request) throws IOException { return IOUtils.toString(request.getReader()); }

1.1.3. 二者区别

• 如果一个请求的Content-Type设置为application/x-www-form-urlencoded，那么这个Post请求会被认为是Http Post表单请求，那么请求主体将以一个标准的键值对和&的querystring形式出现。这种方式是HTML表单的默认设置，所以在过去这种方式更加常见。
• 其他形式的POST请求，是放到 Request payload 中（现在是为了方便阅读，使用了Json这样的数据格式），请求的Content-Type设置为application/json;charset=UTF-8或者不指定。

2. 环境

• python 3.6.1
• 系统：win7
• IDE：pycharm
• requests 2.14.2
• scrapy 1.4.0

3. 使用requests模块post payload请求

import json import requests import datetime

postUrl = 'https://sellercentral.amazon.com/fba/profitabilitycalculator/getafnfee?profitcalcToken=en2kXFaY81m513NydhTZ9sdb6hoj3D' # payloadData数据 payloadData = { 'afnPriceStr': 10, 'currency':'USD', 'productInfoMapping': { 'asin': 'B072JW3Z6L', 'dimensionUnit': 'inches', } } # 请求头设置 payloadHeader = { 'Host': 'sellercentral.amazon.com', 'Content-Type': 'application/json', } # 下载超时 timeOut = 25 # 代理 proxy = "183.12.50.118:8080" proxies = { "http": proxy, "https": proxy, } r = requests.post(postUrl, data=json.dumps(payloadData), headers=payloadHeader) dumpJsonData = json.dumps(payloadData) print(f"dumpJsonData = {dumpJsonData}") res = requests.post(postUrl, data=dumpJsonData, headers=payloadHeader, timeout=timeOut, proxies=proxies, allow_redirects=True) # 下面这种直接填充json参数的方式也OK # res = requests.post(postUrl, json=payloadData, headers=header) print(f"responseTime = {datetime.datetime.now()}, statusCode = {res.status_code}, res text = {res.text}")

4. 在scrapy中post payload请求

• 这儿有个坏消息，那就是scrapy目前还不支持payload这种request请求。而且scrapy对formdata的请求也有很严格的要求，具体可以参考这篇文章：https://blog.csdn.net/zwq912318834/article/details/78292536

4.1. 分析scrapy源码

• 参考注解

# 文件：E:\Miniconda\Lib\site-packages\scrapy\http\request\form.py class FormRequest(Request):

def __init__(self, *args, **kwargs): formdata = kwargs.pop('formdata', None) if formdata and kwargs.get('method') is None: kwargs['method'] = 'POST'

super(FormRequest, self).__init__(*args, **kwargs)

if formdata: items = formdata.items() if isinstance(formdata, dict) else formdata querystr = _urlencode(items, self.encoding) # 这儿写死了，当提交数据时，设置好Content-Type，也就是form data类型 # 就算改写这儿，后面也没有对 json数据解析的处理 if self.method == 'POST': self.headers.setdefault(b'Content-Type', b'application/x-www-form-urlencoded') self._set_body(querystr) else: self._set_url(self.url + ('&' if '?' in self.url else '?') + querystr)

4.2. 思路：在scrapy中嵌入requests模块

• 分析请求
  
  
  
• 返回的查询结果
  
• 第一步：在爬虫中构造请求，把所有的参数以及必要信息带进去。

# 文件 mySpider.py中

payloadData = {} payloadData['afnPriceStr'] = 0 payloadData['currency'] = asinInfo['currencyCodeHidden'] payloadData['futureFeeDate'] = asinInfo['futureFeeDateHidden'] payloadData['hasFutureFee'] = False payloadData['hasTaxPage'] = True payloadData['marketPlaceId'] = asinInfo['marketplaceIdHidden'] payloadData['mfnPriceStr'] = 0 payloadData['mfnShippingPriceStr'] = 0 payloadData['productInfoMapping'] = {} payloadData['productInfoMapping']['asin'] = dataFieldJson['asin'] payloadData['productInfoMapping']['binding'] = dataFieldJson['binding'] payloadData['productInfoMapping']['dimensionUnit'] = dataFieldJson['dimensionUnit'] payloadData['productInfoMapping']['dimensionUnitString'] = dataFieldJson['dimensionUnitString'] payloadData['productInfoMapping']['encryptedMarketplaceId'] = dataFieldJson['encryptedMarketplaceId'] payloadData['productInfoMapping']['gl'] = dataFieldJson['gl'] payloadData['productInfoMapping']['height'] = dataFieldJson['height'] payloadData['productInfoMapping']['imageUrl'] = dataFieldJson['imageUrl'] payloadData['productInfoMapping']['isAsinLimits'] = dataFieldJson['isAsinLimits'] payloadData['productInfoMapping']['isWhiteGloveRequired'] = dataFieldJson['isWhiteGloveRequired'] payloadData['productInfoMapping']['length'] = dataFieldJson['length'] payloadData['productInfoMapping']['link'] = dataFieldJson['link'] payloadData['productInfoMapping']['originalUrl'] = dataFieldJson['originalUrl'] payloadData['productInfoMapping']['productGroup'] = dataFieldJson['productGroup'] payloadData['productInfoMapping']['subCategory'] = dataFieldJson['subCategory'] payloadData['productInfoMapping']['thumbStringUrl'] = dataFieldJson['thumbStringUrl'] payloadData['productInfoMapping']['title'] = dataFieldJson['title'] payloadData['productInfoMapping']['weight'] = dataFieldJson['weight'] payloadData['productInfoMapping']['weightUnit'] = dataFieldJson['weightUnit'] payloadData['productInfoMapping']['weightUnitString'] = dataFieldJson['weightUnitString'] payloadData['productInfoMapping']['width'] = dataFieldJson['width'] # https://sellercentral.amazon.com/fba/profitabilitycalculator/getafnfee?profitcalcToken=en2kXFaY81m513NydhTZ9sdb6hoj3D postUrl = f"https://sellercentral.amazon.com/fba/profitabilitycalculator/getafnfee?profitcalcToken={asinInfo['tokenValue']}" payloadHeader = { 'Host': 'sellercentral.amazon.com', 'Content-Type': 'application/json', } # scrapy源码：self.headers.setdefault(b'Content-Type', b'application/x-www-form-urlencoded') print(f"payloadData = {payloadData}") # 这个request并不真正用来调度，去发出请求，因为这种方式构造方式，是无法提交成功的，会返回404错误 # 这样构造主要是把查询参数提交出去，在下载中间件部分用request模块下载，用 “payloadFlag” 标记这种request yield Request(url = postUrl, headers = payloadHeader, meta = {'payloadFlag': True, 'payloadData': payloadData, 'headers': payloadHeader, 'asinInfo': asinInfo}, callback = self.parseAsinSearchFinallyRes, errback = self.error, dont_filter = True )

• 第二步：在中间件中，用requests模块处理这个请求

# 文件：middlewares.py

class PayLoadRequestMiddleware: def process_request(self, request, spider): # 如果有的请求是带有payload请求的，在这个里面处理掉 if request.meta.get('payloadFlag', False): print(f"PayLoadRequestMiddleware enter") postUrl = request.url headers = request.meta.get('headers', {}) payloadData = request.meta.get('payloadData', {}) proxy = request.meta['proxy'] proxies = { "http": proxy, "https": proxy, } timeOut = request.meta.get('download_timeout', 25) allow_redirects = request.meta.get('dont_redirect', False) dumpJsonData = json.dumps(payloadData) print(f"dumpJsonData = {dumpJsonData}") # 发现这个居然是个同步 阻塞的过程，太过影响速度了 res = requests.post(postUrl, data=dumpJsonData, headers=headers, timeout=timeOut, proxies=proxies, allow_redirects=allow_redirects) # res = requests.post(postUrl, json=payloadData, headers=header) print(f"responseTime = {datetime.datetime.now()}, res text = {res.text}, statusCode = {res.status_code}") if res.status_code > 199 and res.status_code < 300: # 返回Response，就进入callback函数处理，不会再去下载这个请求 return HtmlResponse(url=request.url, body=res.content, request=request, # 最好根据网页的具体编码而定 encoding='utf-8', status=200) else: print(f"request mode getting page error, Exception = {e}") return HtmlResponse(url=request.url, status=500, request=request)

4.3. 遗留下的问题

• scrapy之所以强大，就是因为并发度高。大家都知道，由于Python GIL的原因，导致python无法通过多线程来提高性能。但是至少可以做到下载与解析同步的过程，在下载空档的时候，进行数据的解析，调度等等，这都归功于scrapy采用的异步结构。
• 但是，我们在中间件中使用requests模块进行网页下载，因为这是个同步过程，所以会阻塞在这个地方，拉低了整个爬虫的效率。
• 所以，需要根据项目具体的情况，来决定合适的方案。当然这里又涉及到一个新的话题，就是scrapy提供的两种爬取模式：深度优先模式和广度优先模式。如何尽可能最大限度的利用scrapy的并发？在环境不稳定的情形下如何保证尽可能稳定的拿到数据？
• 深度优先模式和广度优先模式是在settings中设置的。

# 文件： settings.py

# DEPTH_PRIORITY（默认值为0）设置为一个正值后，Scrapy的调度器的队列就会从LIFO变成FIFO，因此抓取规则就由DFO（深度优先）变成了BFO（广度优先） DEPTH_PRIORITY = 1, # 广度优先（肯呢个会累积大量的request，累计占有大量的内存，最终数据也在最后一批爬取）

• 深度优先：DEPTH_PRIORITY = 0
  
• 广度优先：DEPTH_PRIORITY = 1
  
• 想将这个过程做成异步的，一直没有思路，欢迎大神提出好的想法