scrapy爬虫注意点（1）—— scrapy.FormRequest中formdata参数

1. 背景

• 在网页爬取的时候，有时候会使用scrapy.FormRequest向目标网站提交数据（表单提交）。参照scrapy官方文档的标准写法是：

# header信息unicornHeader = {    'Host': 'www.example.com',    'Referer': 'http://www.example.com/',}# 表单需要提交的数据myFormData = {'name': 'John Doe', 'age': '27'}# 自定义信息，向下层响应(response)传递下去customerData = {'key1': 'value1', 'key2': 'value2'}yield scrapy.FormRequest(url = "http://www.example.com/post/action",                         headers = unicornHeader,                         func = 'POST',             # GET or POST                         formdata = myFormData,       # 表单提交的数据                         meta = customerData,       # 自定义，向response传递数据                         callback = self.after_post,                         errback = self.error_handle,                         # 如果需要多次提交表单，且url一样，那么就必须加此参数dont_filter，防止被当成重复网页过滤掉了                         dont_filter = True                              )

• 但是，当表单提交数据myFormData 是形如字典内嵌字典的形式，又该如何写？

2. 案例 — 参数为字典

• 在做亚马逊网站爬取时，当进入商家店铺，爬取店铺内商品列表时，发现采取的方式是ajax请求，返回的是json数据。

• 请求信息如下：
  
  
  

• 响应信息如下：
  

• 如上图所示，From Data中的数据包含一个字典：

marketplaceID:ATVPDKIKX0DERseller:A2FE6D62A4WM6QproductSearchRequestData:{"marketplace":"ATVPDKIKX0DER","seller":"A2FE6D62A4WM6Q","url":"/sp/ajax/products","pageSize":12,"searchKeyword":"","extraRestrictions":{},"pageNumber":"1"}# formDate 必须构造如下：myFormData = {    'marketplaceID' : 'ATVPDKIKX0DER',    'seller' : 'A2FE6D62A4WM6Q',    # 注意下面这一行，内部字典是作为一个字符串的形式    'productSearchRequestData' :'{"marketplace":"ATVPDKIKX0DER","seller":"A2FE6D62A4WM6Q","url":"/sp/ajax/products","pageSize":12,"searchKeyword":"","extraRestrictions":{},"pageNumber":"1"}'}

• 在amazon中实际使用的构造方法如下：

def sendRequestForProducts(response):    ajaxParam = response.meta    for pageIdx in range(1, ajaxParam['totalPageNum']+1):        ajaxParam['isFirstAjax'] = False        ajaxParam['pageNumber'] = pageIdx        unicornHeader = {            'Host': 'www.amazon.com',            'Origin': 'https://www.amazon.com',            'Referer': ajaxParam['referUrl'],        }        '''        marketplaceID:ATVPDKIKX0DER        seller:AYZQAQRQKEXRP        productSearchRequestData:{"marketplace":"ATVPDKIKX0DER","seller":"AYZQAQRQKEXRP","url":"/sp/ajax/products","pageSize":12,"searchKeyword":"","extraRestrictions":{},"pageNumber":1}        '''        productSearchRequestData = '{"marketplace": "ATVPDKIKX0DER", "seller": "' + f'{ajaxParam["sellerID"]}' + '","url": "/sp/ajax/products", "pageSize": 12, "searchKeyword": "","extraRestrictions": {}, "pageNumber": "' + str(pageIdx) + '"}'        formdataProduct = {            'marketplaceID': ajaxParam['marketplaceID'],            'seller': ajaxParam['sellerID'],            'productSearchRequestData': productSearchRequestData        }        productAjaxMeta = ajaxParam        # 请求店铺商品列表        yield scrapy.FormRequest(            url = 'https://www.amazon.com/sp/ajax/products',            headers = unicornHeader,            formdata = formdataProduct,            func = 'POST',            meta = productAjaxMeta,            callback = self.solderProductAjax,            errback = self.error,  # 处理http error            dont_filter = True,    # 需要加此参数的        )

3. 原理分析

• 举例来说，目前有如下一笔数据：

formdata = {            'Field': {"pageIdx":99, "size":"10"},            'func': 'nextPage',            }

• 从网页上，可以看到请求数据如下：

Field=%7B%22pageIdx%22%3A99%2C%22size%22%3A%2210%22%7D&func=nextPage

• 第一种，按照如下方式发出请求，结果如下（正确）：

yield scrapy.FormRequest(            url = 'https://www.example.com/sp/ajax',            headers = unicornHeader,            formdata = {                        'Field': '{"pageIdx":99, "size":"10"}',                        'func': 'nextPage',                       },            func = 'POST',            callback = self.handleFunc,        )# 请求数据为：Field=%7B%22pageIdx%22%3A99%2C%22size%22%3A%2210%22%7D&func=nextPage

• 第二种，按照如下方式发出请求，结果如下（错误，无法获取到正确的数据）：

yield scrapy.FormRequest(            url = 'https://www.example.com/sp/ajax',            headers = unicornHeader,            formdata = {                        'Field': {"pageIdx":99, "size":"10"},                        'func': 'nextPage',                       },            func = 'POST',            callback = self.handleFunc,        )# 经过错误的编码之后，发送的请求为：Field=size&Field=pageIdx&func=nextPage

• 我们跟踪看一下scrapy中的源码：

# E:/Miniconda/Lib/site-packages/scrapy/http/request/form.py# FormRequestclass FormRequest(Request):    def __init__(self, *args, **kwargs):        formdata = kwargs.pop('formdata', None)        if formdata and kwargs.get('func') is None:            kwargs['func'] = 'POST'        super(FormRequest, self).__init__(*args, **kwargs)        if formdata:            items = formdata.items() if isinstance(formdata, dict) else formdata            querystr = _urlencode(items, self.encoding)            if self.func == '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)    # 关键函数 _urlencode    def _urlencode(seq, enc):        values = [(to_bytes(k, enc), to_bytes(v, enc))                  for k, vs in seq                  for v in (vs if is_listlike(vs) else [vs])]        return urlencode(values, doseq=1)

• 分析过程如下：

# 第一步：items = formdata.items() if isinstance(formdata, dict) else formdata# 第一步结果：经过items()方法执行后，原始的dict格式变成如下列表形式：     dict_items([('func', 'nextPage'), ('Field', {'size': '10', 'pageIdx': 99})])# 第二步：再经过后面的 _urlencode方法将items转换成如下：    [(b'func', b'nextPage'), (b'Field', b'size'), (b'Field', b'pageIdx')]# 可以看到就是在调用 _urlencode方法的时候出现了问题，上面的方法执行过后，会使字典形式的数据只保留了keys（value是字典的情况下，只保留了value字典中的key）.

• 解决方案： 就是将字典当成普通的字符串，然后编码（转换成bytes），进行传输，到达服务器端之后，服务器会反过来进行解码，得到这个字典字符串。然后服务器按照Dict进行解析。
• 拓展：对于其他特殊类型的数据，都按照这种方式打包成字符串进行传递。

4. 补充1 ——参数类型

• formdata的 参数值 必须是unicode , str 或者 bytes object，不能是整数。
  
  • 案例：

yield FormRequest(    url = 'https://www.amztracker.com/unicorn.php',    headers = unicornHeader,    # formdata 的参数必须是字符串    formdata={'rank': 10, 'category': productDetailInfo['topCategory']},    method = 'GET',    meta = {'productDetailInfo': productDetailInfo},    callback = self.amztrackerSale,    errback = self.error,  # 本项目中这里触发errback占绝大多数    dont_filter = True,  # 按理来说是不需要加此参数的)# 提示如下ERROR：Traceback (most recent call last):  File "E:\Miniconda\lib\site-packages\scrapy\utils\defer.py", line 102, in iter_errback    yield next(it)  File "E:\Miniconda\lib\site-packages\scrapy\spidermiddlewares\offsite.py", line 29, in process_spider_output    for x in result:  File "E:\Miniconda\lib\site-packages\scrapy\spidermiddlewares\referer.py", line 339, in <genexpr>    return (_set_referer(r) for r in result or ())  File "E:\Miniconda\lib\site-packages\scrapy\spidermiddlewares\urllength.py", line 37, in <genexpr>    return (r for r in result or () if _filter(r))  File "E:\Miniconda\lib\site-packages\scrapy\spidermiddlewares\depth.py", line 58, in <genexpr>    return (r for r in result or () if _filter(r))  File "E:\PyCharmCode\categorySelectorAmazon1\categorySelectorAmazon1\spiders\categorySelectorAmazon1Clawer.py", line 224, in parseProductDetail    dont_filter = True,  File "E:\Miniconda\lib\site-packages\scrapy\http\request\form.py", line 31, in __init__    querystr = _urlencode(items, self.encoding)  File "E:\Miniconda\lib\site-packages\scrapy\http\request\form.py", line 66, in _urlencode    for k, vs in seq  File "E:\Miniconda\lib\site-packages\scrapy\http\request\form.py", line 67, in <listcomp>    for v in (vs if is_listlike(vs) else [vs])]  File "E:\Miniconda\lib\site-packages\scrapy\utils\python.py", line 117, in to_bytes    'object, got %s' % type(text).__name__)TypeError: to_bytes must receive a unicode, str or bytes object, got int# 正确写法：formdata = {'rank': str(productDetailInfo['topRank']), 'category': productDetailInfo['topCategory']},

• 原理部分（源代码）：

# 第一阶段: 字典分解为items        if formdata:            items = formdata.items() if isinstance(formdata, dict) else formdata            querystr = _urlencode(items, self.encoding)# 第二阶段: 对value，调用 to_bytes 编码def _urlencode(seq, enc):    values = [(to_bytes(k, enc), to_bytes(v, enc))              for k, vs in seq              for v in (vs if is_listlike(vs) else [vs])]    return urlencode(values, doseq=1)# 第三阶段: 执行 to_bytes ，参数要求是bytes, strdef to_bytes(text, encoding=None, errors='strict'):    """Return the binary representation of `text`. If `text`    is already a bytes object, return it as-is."""    if isinstance(text, bytes):        return text    if not isinstance(text, six.string_types):        raise TypeError('to_bytes must receive a unicode, str or bytes '                        'object, got %s' % type(text).__name__)

5. 补充2 ——参数为中文

• formdata的 参数值 必须是unicode , str 或者 bytes object，不能是整数。

  • 以1688网站搜索产品为案例：

• 搜索信息如下（搜索关键词为：动漫周边）：
  
  
  

• 可以看到 动漫周边 == %B6%AF%C2%FE%D6%DC%B1%DF

# scrapy中这个请求的构造如下# python3 所有的字符串都是unicodeunicornHeaders = {    ':authority': 's.1688.com',    'Referer': 'https://www.1688.com/',}# python3 所有的字符串都是unicode# 动漫周边 tobyte为：%B6%AF%C2%FE%D6%DC%B1%DFformatStr = "动漫周边".encode('gbk')print(f"formatStr = {formatStr}")yield FormRequest(    url = 'https://s.1688.com/selloffer/offer_search.htm',    headers = unicornHeaders,    formdata = {'keywords': formatStr, 'n': 'y', 'spm': 'a260k.635.1998096057.d1'},    method = 'GET',    meta={},    callback = self.parseCategoryPage,    errback = self.error,  # 本项目中这里触发errback占绝大多数    dont_filter = True,  # 按理来说是不需要加此参数的)# 日志如下：formatStr = b'\xb6\xaf\xc2\xfe\xd6\xdc\xb1\xdf'2017-11-16 15:11:02 [scrapy.downloadermiddlewares.redirect] DEBUG: Redirecting (302) to <GET https://sec.1688.com/query.htm?smApp=searchweb2&smPolicy=searchweb2-selloffer-anti_Spider-seo-html-checklogin&smCharset=GBK&smTag=MTE1LjIxNi4xNjAuNDYsLDU5OWQ1NWIyZTk0NDQ1Y2E5ZDAzODRlOGM1MDI2OTZj&smReturn=https%3A%2F%2Fs.1688.com%2Fselloffer%2Foffer_search.htm%3Fkeywords%3D%25B6%25AF%25C2%25FE%25D6%25DC%25B1%25DF%26n%3Dy%26spm%3Da260k.635.1998096057.d1&smSign=05U0%2BJXfKLQmSbsnce55Yw%3D%3D> from <GET https://s.1688.com/selloffer/offer_search.htm?keywords=%B6%AF%C2%FE%D6%DC%B1%DF&n=y&spm=a260k.635.1998096057.d1># https://s.1688.com/selloffer/offer_search.htm?keywords=%B6%AF%C2%FE%D6%DC%B1%DF&n=y&spm=a260k.635.1998096057.d1