Scrapy - Request Payload format and types

Question

This is the starting point of my scraping process.

https://www.storiaimoveis.com.br/alugar/brasil

This is the AJAX call which returns the data in a JSON format for every page.

https://www.storiaimoveis.com.br/api/search?fields=%24%24meta.geo.postalCodeAddress.city%2C%24%24meta.geo.postalCodeAddress.neighborhood%2C%24%24meta.geo.postalCodeAddress.street%2C%24%24meta.location%2C%24%24meta.created%2Caddress.number%2Caddress.postalCode%2Caddress.neighborhood%2Caddress.state%2Cmedia%2ClivingArea%2CtotalArea%2Ctypes%2Coperation%2CsalePrice%2CrentPrice%2CnewDevelopment%2CadministrationFee%2CyearlyTax%2Caccount.logoUrl%2Caccount.name%2Caccount.id%2Caccount.creci%2Cgarage%2Cbedrooms%2Csuites%2Cbathrooms%2Cref&optimizeMedia=true&size=20&from=0&sessionId=5ff29d7e-88d0-54d5-2641-e203cafd6f4e

My POST request fails with error 404. Those requests require payloads gave me trouble in the past. I always solved the problem somehow, but now I'm trying to understand what am I doing wrong with them.

My questions are;

• Does request payloads sent along with scrapy requests need a specific type or format?
• Do I need to call json.dumps(payload) before sending them, or send them as dictionaries?.
• Do I need to convert every key:value pair to strings before sending the payload?
• Could be any other reason why my request is failing?

This is my code's relevant parts.

class MySpider(CrawlSpider):

    name = 'myspider'

    start_urls = [
        'https://www.storiaimoveis.com.br/api/search?fields=%24%24meta.geo.postalCodeAddress.city%2C%24%24meta.geo.postalCodeAddress.neighborhood%2C%24%24meta.geo.postalCodeAddress.street%2C%24%24meta.location%2C%24%24meta.created%2Caddress.number%2Caddress.postalCode%2Caddress.neighborhood%2Caddress.state%2Cmedia%2ClivingArea%2CtotalArea%2Ctypes%2Coperation%2CsalePrice%2CrentPrice%2CnewDevelopment%2CadministrationFee%2CyearlyTax%2Caccount.logoUrl%2Caccount.name%2Caccount.id%2Caccount.creci%2Cgarage%2Cbedrooms%2Csuites%2Cbathrooms%2Cref&optimizeMedia=true&size=20&from=0&sessionId=5ff29d7e-88d0-54d5-2641-e203cafd6f4e'
    ]

    page = 1
    payload = {"locations":[{"geo":{"top_left":{"lat":5.2717863,
                                                "lon":-73.982817},
                                    "bottom_right":{"lat":-34.0891,
                                                    "lon":-28.650543}},
                             "placeId":"ChIJzyjM68dZnAARYz4p8gYVWik",
                             "keywords":"Brasil",
                             "address":{"label":"Brasil","country":"BR"}}],
               "operation":["RENT"],
               "bathrooms":[],
               "bedrooms":[],
               "garage":[],
               "features":[]}
    headers = {
        'Accept': 'application/json',
        'Content-Type': 'application/json',
        'Referer': 'https://www.storiaimoveis.com.br/alugar/brasil',
        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/75.0.3770.100 Safari/537.36'
    }


    def parse(self, response):
        for url in self.start_urls:
            yield scrapy.Request(url=url,
                                 method='POST',
                                 headers=self.headers,
                                 body=json.dumps(self.payload),
                                 callback=self.parse_items)

    def parse_items(self, response):
        from scrapy.shell import inspect_response
        inspect_response(response, self)
        print response.text

Answer 1

Yes, you need to call json.dumps(payload) because the request body needs to be str or unicode as stated in the documentation: https://docs.scrapy.org/en/latest/topics/request-response.html#request-objects

But, in your case, your request fails because of these 2 missing headers: Content-Type and Referer .

What I usually do in order to get the right request headers is this:

1. Inspect the headers in Chrome dev tools:

2. Use curl or Postman to make the request until I get the right headers. In this case, Content-Type and Referer seem to be enough for an HTTP 200 response status: