Scrapy 1.8.0 returns 错误 500，但 Python 代码 returns 成功 200

Question

地址是：'https://planningapi.agileapplications.co.uk/api/application/search?reference=GDO+19%2F12'

我可以通过 Python 请求库轻松下载此页面：

headers = {
            'x-client': 'EXMOOR',
            'x-product': 'CITIZENPORTAL',
            'x-service': 'PA',
            }
url='https://planningapi.agileapplications.co.uk/api/application/search?reference=GDO+19%2F12'
resp = requests.get(url, headers=headers)

或者我可以通过 CURL 轻松下载页面：

curl 'https://planningapi.agileapplications.co.uk/api/application/search?reference=GDO+19%2F12' -H 'x-product: CITIZENPORTAL' -H 'x-service: PA' -H 'x-client: EXMOOR'

他们都 return 状态 200 结果：

{"total":1,"results":[{"id":18468,"reference":"GDO 19/12","proposal":"Prior notification for excavations to bury tanks and trenches to lay water pipes","location":"Land North West of North and South Ley, Exford, Minehead, Somerset.","username":"","applicantSurname":"Mr & Mrs M Burnett","agentName":"JCH Planning Limited","decisionText":null,"registrationDate":"2019-10-04","decisionDate":"2019-10-30","finalGrantDate":null,"appealLodgedDate":null,"appealDecisionDate":null,"areaId":[],"wardId":[],"parishId":[3],"responded":null,"lastLetterDate":null,"targetResponseDate":null}]}

但是Scrapy return的状态500错误：

formdata = {'reference': 'GDO 19/12', }

headers = {
    'x-client': 'EXMOOR',
    'x-product': 'CITIZENPORTAL',
    'x-service': 'PA',
    }

fr = scrapy.FormRequest(
    url='https://planningapi.agileapplications.co.uk/api/application/search',
    method='GET',
    meta=response.meta,
    headers=headers,
    formdata=formdata,
    dont_filter=True,
    callback=self.ref_result_2,
    )

yield fr

可能是因为 Scrapy 将 headers 键大写（我试过 un-capitalizing 它们，但后来 Twisted 做了同样的事情 - 它再次将它们大写），也许是出于其他原因。

如何调整我的 Scrapy 1.8.0 代码以成功获得与 Python 请求相同的结果？

Answer 1

确实是Scrapy将header字段大写造成的。如果您尝试在 cURL 命令中将 then 大写，您将得到与使用 Scrapy 时相同的错误（您可以在蜘蛛 class 中的 Scrapy 设置 handle_httpstatus_list 中测试它并打印 response.text 在解析方法中）。正如您也已经说过的那样，Twisted 也是如此，因此覆盖 scrapy.http.Headers 不是解决方案。

但是，根据 this issue comment:

，您可以采取一些技巧使 Twisted 不将特定的 header 大写

# -*- coding: utf-8 -*-
from pprint import pprint
import scrapy
from twisted.web.http_headers import Headers as TwistedHeaders

TwistedHeaders._caseMappings.update({
    b'x-client': b'x-client',
    b'x-product': b'x-product',
    b'x-service': b'x-service',
})

class Foo(scrapy.Spider):
    name = 'foo'
    handle_httpstatus_list = [500]

    def start_requests(self):
        formdata = {'reference': 'GDO 19/12'}
        headers = {
            'x-client': 'EXMOOR',
            'x-product': 'CITIZENPORTAL',
            'x-service': 'PA'
        }
        yield scrapy.FormRequest(
            'https://planningapi.agileapplications.co.uk/api/application/search',
            method='GET', headers=headers, formdata=formdata, callback=self.parse)

    def parse(self, response):
        pprint(response.text)

现在你会得到结果。另一方面，根据 RFC 7230，第 3.2 节，header 字段应为 case-insensitive.

Scrapy 1.8.0 returns 错误 500，但 Python 代码 returns 成功 200

Scrapy 1.8.0 returns error 500, but Python code returns success 200

python

python-3.x

scrapy

header

twisted