I would like to get the same result as this command line :
scrapy crawl linkedin_anonymous -a first=James -a last=Bond -o output.json

My script is as follows :

import scrapy
from linkedin_anonymous_spider import LinkedInAnonymousSpider
from scrapy.crawler import CrawlerProcess
from scrapy.utils.project import get_project_settings

spider = LinkedInAnonymousSpider(None, "James", "Bond")
process = CrawlerProcess(get_project_settings())
process.crawl(spider) ## <-------------- (1)
process.start()

I found out that process.crawl() in (1) is creating another LinkedInAnonymousSpider where first and last are None (printed in (2)), if so, then there is no point of creating the object spider and how is it possible to pass the arguments first and last to process.crawl()?

linkedin_anonymous :

from logging import INFO

import scrapy

class LinkedInAnonymousSpider(scrapy.Spider):
    name = "linkedin_anonymous"
    allowed_domains = ["linkedin.com"]
    start_urls = []

    base_url = "https://www.linkedin.com/pub/dir/?first=%s&last=%s&search=Search"

    def __init__(self, input = None, first= None, last=None):
        self.input = input  # source file name
        self.first = first
        self.last = last

    def start_requests(self):
        print self.first ## <------------- (2)
        if self.first and self.last: # taking input from command line parameters
                url = self.base_url % (self.first, self.last)
                yield self.make_requests_from_url(url)

    def parse(self, response): . . .

pass the spider arguments on the process.crawl method:

process.crawl(spider, input="inputargument", first="James", last="Bond")

You can do it the easy way:

from scrapy import cmdline

cmdline.execute("scrapy crawl linkedin_anonymous -a first=James -a last=Bond -o output.json".split())

if you have Scrapyd and you want to schedule the spider, do this

curl http://localhost:6800/schedule.json -d project=projectname -d spider=spidername -d first="James" -d last="Bond"