sarafg11

Untitled

Dec 30th, 2019
137
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 1.15 KB | None | 0 0
  1. {
  2. "sitemap_urls": [
  3. "https://www.sansperfumerias.com/1_es_0_sitemap.xml"
  4. ],
  5. "fields": {
  6. "name": "//h1[@itemprop='name']/text()",
  7. "price": "//div[@class='primary_block row']//p[@class='our_price_display']/span[@id='our_price_display']/text()",
  8. "price_old": "//div[contains(@class,'primary_block')]//p[@id='old_price']/span[@id='old_price_display']/text()",
  9. "reference": "",
  10. "image": "//img[@id='bigpic']/@src",
  11. "description": "//section[@id='idTab1']//text()|//div[@id='short_description_content']//text()",
  12. "brand": "",
  13. "brand_text": "//script[@type='text/javascript'][contains(.,'brand')]/text()",
  14. "categories": "(//div[contains(@class,'bread')]//a)[position()>1 and position()<=last()]/text()",
  15. "availability": "(//link[@itemprop='availability']/@href)[1]",
  16. "rating_average": "",
  17. "rating_count": "",
  18. "barcode": "",
  19. "stock": ""
  20. },
  21. "links": {
  22. "detail": "//h5/a[@class='product-name']",
  23. "next_page": "(//a[contains(@class,'next')])[last()]"
  24. },
  25. "regex": {
  26. "brand_text": ["brand\\\":\\\"\\\\\\\"(.*?)\\\\\\\"", 1]
  27. },
  28. "sitemap_rules":[
  29. [".html$", "parse"]
  30. ]
  31. }
Advertisement
Add Comment
Please, Sign In to add comment