[shopify] skip consecutive duplicate products
Not filtering duplicate URLs anymore caused the archive ID uniqueness test to fail.
This commit is contained in:
@@ -89,10 +89,14 @@ class ShopifyCollectionExtractor(ShopifyExtractor):
|
|||||||
while True:
|
while True:
|
||||||
page = self.request(self.item_url, params=params).text
|
page = self.request(self.item_url, params=params).text
|
||||||
urls = search_re.findall(page)
|
urls = search_re.findall(page)
|
||||||
|
last = None
|
||||||
|
|
||||||
if not urls:
|
if not urls:
|
||||||
return
|
return
|
||||||
for path in urls:
|
for path in urls:
|
||||||
|
if last == path:
|
||||||
|
continue
|
||||||
|
last = path
|
||||||
yield self.root + path
|
yield self.root + path
|
||||||
params["page"] += 1
|
params["page"] += 1
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user