Understandin how rename images scrapy works

2019-02-21 05:50发布

I see all questions here, but i dont understand yet.

Actualy with de code bellow i do what i need, except rename de image, so i try change name in the items.py file, pls check comments inside.

settings.py

SPIDER_MODULES = ['xxx.spiders']
NEWSPIDER_MODULE = 'xxx.spiders'
ITEM_PIPELINES = {'scrapy.pipelines.images.ImagesPipeline': 1}
IMAGES_STORE = '/home/magicnt/xxx/images'

items.py

class XxxItem(scrapy.Item):
    # define the fields for your item here like:
    # name = scrapy.Field()
    title = scrapy.Field()
    image_urls = scrapy.Field()
    #images = scrapy.Field()<---with that code work with default name images
    images = title<--- I try rename here, but not work

spider.py

from xxx.items import XxxItem
import scrapy
from scrapy.pipelines.images import ImagesPipeline
from scrapy.exceptions import DropItem

class CoverSpider(scrapy.Spider):
    name = "pyimagesearch-cover-spider"
    start_urls = ['https://xxx.com.br/product']
    def parse(self, response):
        for bimb in response.css('#mod_imoveis_result'):
            imageURL = bimb.xpath('./div[@id="g-img-imo"]/div[@class="img_p_results"]/img/@src').extract_first()
            title = bimb.css('#titulo_imovel::text').extract_first()
            yield {
                'image_urls' : [response.urljoin(imageURL)],
                'title' : title
            }
        next_page = response.xpath('//a[contains(@class, "num_pages") and contains(@class, "pg_number_next")]/@href').extract_first()
        yield response.follow(next_page, self.parse)

My goal is rename downloaded images with the title from item. Any tip for this goal are welcome.

I'm totally new to python and oo, I usually scrape with structural php but realize what a good scrapy it can be, ask for a little patience and help.

标签: python scrapy
1条回答
Explosion°爆炸
2楼-- · 2019-02-21 06:26

My code is based on Scrapy Image Pipeline: How to rename images? I tested it a week ago and it works on my own spiders.

# This pipeline is designed for an item with multiple images
class ImagesWithNamesPipeline(ImagesPipeline):
    def get_media_requests(self, item, info):
        # values in field "image_name" must have suffix ".jpg"
        # you can only change "image_name" to your own image name filed "images"
        # however it should be a list
        for (image_url, image_name) in zip(item[self.IMAGES_URLS_FIELD], item["image_names"]):
            yield scrapy.Request(url=image_url, meta={"image_name": image_name})

    def file_path(self, request, response=None, info=None):
        image_name = request.meta["image_name"]
        return image_name

Here is how the ImagePipeline works:

The pipeline will execute image_downloaded -> get_images -> file_path in order. ("->" means invokes)

  • image_downloaded: save images that get_images return by invoking persist_file
  • get_images: convert images to JPEG
  • file_path: return the relative path of image

I scaned through the source code of ImagePipeline and found no special field for rename an image. Scrapy will rename it in this way:

def file_path(self, request, response=None, info=None):
    image_guid = hashlib.sha1(to_bytes(url)).hexdigest()  # change to request.url after deprecation
    return 'full/%s.jpg' % (image_guid)

Therefore we should override method file_path. According to the source code of FilePipeline which ImagePipeline inherits, we only need to return relative paths and persist_file will get things done.

查看更多
登录 后发表回答