我有一系列定义网站结构的链接。从这些链接下载图像时,我想同时将下载的图像放在类似于网站结构的文件夹结构中,而不是只重命名(如Scrapy image download how to use custom filename中所述)
我的代码是这样的:
class MyImagesPipeline(ImagesPipeline):
"""Custom image pipeline to rename images as they are being downloaded"""
page_url=None
def image_key(self, url):
page_url=self.page_url
image_guid = url.split('/')[-1]
return '%s/%s/%s' % (page_url,image_guid.split('_')[0],image_guid)
def get_media_requests(self, item, info):
#http://store.abc.com/b/n/s/m
os.system('mkdir '+item['sku'][0].encode('ascii','ignore'))
self.page_url = urlparse(item['start_url']).path #I store the parent page's url in start_url Field
for image_url in item['image_urls']:
yield Request(image_url)
它会创建所需的文件夹结构,但是当我进入deapth中的文件夹时,我发现这些文件在文件夹中放错了位置。
我怀疑它正在发生,因为“get_media_requests”和“image_key”函数可能是异步执行的,因此“page_url”的值在“image_key”函数使用之前会发生变化。
答案 0 :(得分:1)
异步项处理阻止在管道中通过self
使用类变量,这是绝对正确的。您必须在每个请求中存储路径并覆盖更多方法(未经测试):
def image_key(self, url, page_url):
image_guid = url.split('/')[-1]
return '%s/%s/%s' % (page_url, image_guid.split('_')[0], image_guid)
def get_media_requests(self, item, info):
for image_url in item['image_urls']:
yield Request(image_url, meta=dict(page_url=urlparse(item['start_url']).path))
def get_images(self, response, request, info):
key = self.image_key(request.url, request.meta.get('page_url'))
...
def media_to_download(self, request, info):
...
key = self.image_key(request.url, request.meta.get('page_url'))
...
def media_downloaded(self, response, request, info):
...
try:
key = self.image_key(request.url, request.meta.get('page_url'))
...