如何修复CSV / JSON的Scrapy词典输出格式

Question

My code is below. I'm looking to extract results to CSV. However, the scrapy results in a dictionary with 2 keys, and all the values lumped together in each key. The output does not look good. How do I fix this. Can this be done through pipelines/itemloaders etc...

Thanks very much.

import scrapy
from scrapy.linkextractors import LinkExtractor
from scrapy.spiders import CrawlSpider, Rule
from scrapy.loader import ItemLoader
from scrapy.loader.processors import TakeFirst, MapCompose, Join
from gumtree1.items import GumtreeItems

class AdItemLoader(ItemLoader):
    jobs_in = MapCompose(unicode.strip)

class GumtreeEasySpider(CrawlSpider):
    name = 'gumtree_easy'
    allowed_domains = ['gumtree.com.au']
    start_urls = ['http://www.gumtree.com.au/s-jobs/page-2/c9302?ad=offering']

    rules = (
        Rule(LinkExtractor(restrict_xpaths='//a[@class="rs-paginator-btn next"]'), callback='parse_item', follow=True),
    )

    def parse_item(self, response):
        loader = AdItemLoader(item=GumtreeItems(), response=response)
        loader.add_xpath('jobs','//div[@id="recent-sr-title"]/following-sibling::*//*[@itemprop="name"]/text()')
        loader.add_xpath('location', '//div[@id="recent-sr-title"]/following-sibling::*//*[@class="rs-ad-location-area"]/text()')
        yield loader.load_item() 

The result:

2016-03-16 01:51:32 [scrapy] DEBUG: Scraped from <200 http://www.gumtree.com.au/s-jobs/page-5/c9302?ad=offering>
{'jobs': [u'Technical Account Manager',
          u'Service & Maintenance Advisor',
          u'we are hiring motorbike driver delivery leaflet.Strat NOW(BE...',
          u'Casual Gardner/landscape maintenance labourer',
          u'Seeking for Experienced Builders Cleaners with white card',
          u'Babysitter / home help for approx 2 weeks',
          u'Toothing brickwork | Dapto',
          u'EXPERIENCED CHEF',
          u'ChildCare Trainee Wanted',
          u'Skilled Pipelayers & Drainer- Sydney Region',
          u'Casual staff required for Royal Easter Show',
          u'Fencing contractor',
          u'Excavator & Loader Operator',
          u'***EXPERIENCED STRAWBERRY AND RASPBERRY PICKERS WANTED***',
          u'Kitchenhand required for Indian restaurant',
          u'Taxi Driver Wanted',
          u'Full time nanny/sitter',
          u'Kitchen hand and meal packing',
          u'Depot Assistant Required',
          u'hairdresser Junior apprentice required for salon in Randwick',
          u'Insulation Installers Required',
          u'The Knox is seeking a new apprentice',
          u'Medical Receptionist Needed in Bankstown Area - Night Shifts',
          u'On Call Easy Work, Do you live in Berala, Lidcombe or Auburn...',
          u'Looking for farm jon'],
 'location': [u'Melbourne City',
              u'Eastern Suburbs',
              u'Rockdale Area',
              u'Logan Area',
              u'Greater Dandenong',
              u'Brisbane North East',
              u'Kiama Area',
              u'Byron Area',
              u'Dardanup Area',
              u'Blacktown Area',
              u'Auburn Area',
              u'Kingston Area',
              u'Inner Sydney',
              u'Northern Midlands',
              u'Inner Sydney',
              u'Hume Area',
              u'Maribyrnong Area',
              u'Perth City',
              u'Brisbane South East',
              u'Eastern Suburbs',
              u'Gold Coast South',
              u'North Canberra',
              u'Bankstown Area',
              u'Auburn Area',
              u'Gingin Area']}

Should it be like this instead. Jobs and Location as individual Dicts? This writes correctly to CSV with Jobs and Location in separate cells but I find that using for loops and zip shoult not the best way.

import scrapy
from gumtree1.items import GumtreeItems

class AussieGum1Spider(scrapy.Spider):
    name = "aussie_gum1"
    allowed_domains = ["gumtree.com.au"]
    start_urls = (
        'http://www.gumtree.com.au/s-jobs/page-2/c9302?ad=offering',
    )

    def parse(self, response):
        item = GumtreeItems()
        jobs = response.xpath('//div[@id="recent-sr-title"]/following-sibling::*//*[@itemprop="name"]/text()').extract()
        location = response.xpath('//div[@id="recent-sr-title"]/following-sibling::*//*[@class="rs-ad-location-area"]/text()').extract()
        for j, l in zip(jobs, location):
            item['jobs'] = j.strip()
            item['location'] = l
            yield item

partial results below.

2016-03-16 02:20:46 [scrapy] DEBUG: Crawled (200) <GET http://www.gumtree.com.au/s-jobs/page-3/c9302?ad=offering> (referer: http://www.gumtree.com.au/s-jobs/page-2/c9302?ad=offering)
2016-03-16 02:20:46 [scrapy] DEBUG: Scraped from <200 http://www.gumtree.com.au/s-jobs/page-3/c9302?ad=offering>
{'jobs': u'Live In Au pair-Urgent', 'location': u'Wanneroo Area'}
2016-03-16 02:20:46 [scrapy] DEBUG: Scraped from <200 http://www.gumtree.com.au/s-jobs/page-3/c9302?ad=offering>
{'jobs': u'live in carer', 'location': u'Fraser Coast'}
2016-03-16 02:20:46 [scrapy] DEBUG: Scraped from <200 http://www.gumtree.com.au/s-jobs/page-3/c9302?ad=offering>
{'jobs': u'Mental Health Nurse', 'location': u'Perth Region'}
2016-03-16 02:20:46 [scrapy] DEBUG: Scraped from <200 http://www.gumtree.com.au/s-jobs/page-3/c9302?ad=offering>
{'jobs': u'Experienced NBN pit and pipe installers/node and cabinet wor...',
 'location': u'Marrickville Area'}
2016-03-16 02:20:46 [scrapy] DEBUG: Scraped from <200 http://www.gumtree.com.au/s-jobs/page-3/c9302?ad=offering>
{'jobs': u'Delivery Driver / Pizza Maker Job - Dominos Pizza',
 'location': u'Hurstville Area'}

Thanks very much.

Answer 1

To be honest, using a for loop is the right way, but you can workaround it on a pipeline:

from scrapy.http import Response
from gumtree1.items import GumtreeItems, CustomItem
from scrapy.exceptions import DropItem


class CustomPipeline(object):

    def __init__(self, crawler):
        self.crawler = crawler

    @classmethod
    def from_crawler(cls, crawler):
        return cls(crawler)

    def process_item(self, item, spider):
        if isinstance(item, GumtreeItems):
            for i, jobs in enumerate(item['jobs']):
                self.crawler.engine.scraper._process_spidermw_output(
                    CustomItem(jobs=jobs, location=item['location'][i]), None, Response(''), spider)
            raise DropItem("main item dropped")
        return item

also add the custom item:

class CustomItem(scrapy.Item):
    jobs = scrapy.Field()
    location = scrapy.Field()

Hope this helped, again I think you should use the loop.

Answer 2

Have a parent selector for every item and extract job and location relative to it:

rows = response.xpath('//div[@id="recent-sr-title"]/following-sibling::*')
for row in rows:
    item = GumtreeItems()
    item['jobs'] = row.xpath('.//*[@itemprop="name"]/text()').extract_first().strip()
    item['location'] = row.xpath('.//*[@class="rs-ad-location-area"]/text()').extract_first().strip()
    yield item