Skip to content

Commit

Permalink
Update RandyBlue for new site layout
Browse files Browse the repository at this point in the history
  • Loading branch information
Maista6969 committed Jul 8, 2023
1 parent 7d99295 commit 01f5283
Showing 1 changed file with 16 additions and 13 deletions.
29 changes: 16 additions & 13 deletions scrapers/RandyBlue.yml
Original file line number Diff line number Diff line change
Expand Up @@ -2,29 +2,32 @@ name: "RandyBlue"
sceneByURL:
- action: scrapeXPath
url:
- randyblue.com/video/
- randyblue.com/scenes/
scraper: sceneScraper

xPathScrapers:
sceneScraper:
common:
$titleArea: //div[@class="title-zone"]
scene:
Title: //h2[@class="content-item-name"]
Title: $titleArea/h1
Date:
selector: //div[@id="info"]//li[@class="added"]/span
selector: $titleArea/div[@class="calendar"]
postProcess:
- parseDate: 01/02/2006
Details: //div[@class="short-description"]
Tags:
Name: //div[@class="tagcloud"]/a/text()
Performers:
Name: //a[@itemprop="actor"]
Image:
selector: "//script[contains(text(), 'image: ')]"
Details:
selector: //div[@id="collapseTwo"]
postProcess:
- replace:
- regex: '^.*image: "([^"]+)",.*$'
with: $1
- regex: \x{0020}|\x{00A0} # unicode SP, NBSP
with: " "
Tags:
Name: $titleArea/ul[@class="scene-tags"]/li/a
Performers:
Name: $titleArea/ul[@class="scene-models-list"]/li/a
Image: //meta[@itemprop="thumbnailUrl"]/@content
URL: //link[@rel="canonical"]/@href
Studio:
Name:
fixed: Randy Blue
# Last Updated December 17, 2021
# Last Updated July 8, 2023

0 comments on commit 01f5283

Please sign in to comment.