97 lines
2.7 KiB
YAML
97 lines
2.7 KiB
YAML
name: GirlsOutWest
|
|
sceneByURL:
|
|
- action: scrapeXPath
|
|
url:
|
|
- tour.girlsoutwest.com/
|
|
scraper: sceneScraper
|
|
performerByURL:
|
|
- action: scrapeXPath
|
|
url:
|
|
- tour.girlsoutwest.com/models/
|
|
scraper: performerScraper
|
|
galleryByURL:
|
|
- action: scrapeXPath
|
|
url:
|
|
- tour.girlsoutwest.com/
|
|
scraper: galleryScraper
|
|
xPathScrapers:
|
|
sceneScraper:
|
|
scene:
|
|
Title:
|
|
selector: //div[@class='trailerTitle']/div/h3/text()
|
|
# uncomment the following to strip performer names from titles
|
|
#postProcess:
|
|
# - replace:
|
|
# - regex: .+-
|
|
# with:
|
|
Performers:
|
|
Name:
|
|
selector: //div[@class='trailer topSpace']/div//p/a/text()
|
|
URL:
|
|
selector: //div[@class='trailer topSpace']/div//p/a/@href
|
|
Image:
|
|
# use 1x or 2x for low/medium resolution images
|
|
selector: //div[@class='videoplayer']/img/@src0_3x
|
|
postProcess:
|
|
- replace:
|
|
- regex: ^
|
|
with: https://tour.girlsoutwest.com
|
|
Date:
|
|
selector: //div[@class='trailer topSpace']/div[@class='centerwrap clear']/p/text()
|
|
concat: " "
|
|
postProcess:
|
|
- replace:
|
|
- regex: (.+)(\d{2}\/\d{2}\/\d{4})(.+)
|
|
with: $2
|
|
- parseDate: 01/02/2006
|
|
Studio:
|
|
Name:
|
|
fixed: Girls Out West
|
|
URL:
|
|
fixed: https://girlsoutwest.com
|
|
performerScraper:
|
|
performer:
|
|
Name:
|
|
selector: //div[@class='profileDetails clear']/h2/text()
|
|
Gender:
|
|
fixed: Female
|
|
Image:
|
|
# use 1x or 2x for low/medium resolution images
|
|
selector: //div[@class='profilePic']/img/@src0_3x
|
|
postProcess:
|
|
- replace:
|
|
- regex: ^
|
|
with: https://tour.girlsoutwest.com
|
|
galleryScraper:
|
|
gallery:
|
|
Title:
|
|
selector: //div[@class='trailerTitle']/div/h3/text()
|
|
Performers:
|
|
Name:
|
|
selector: //div[@class='trailer topSpace']/div//p/a/text()
|
|
URL:
|
|
selector: //div[@class='trailer topSpace']/div//p/a/@href
|
|
Date:
|
|
selector: //div[@class='trailer topSpace']/div[@class='centerwrap clear']/p/text()
|
|
concat: " "
|
|
postProcess:
|
|
- replace:
|
|
- regex: (.+)(\d{2}\/\d{2}\/\d{4})(.+)
|
|
with: $2
|
|
- parseDate: 01/02/2006
|
|
Studio:
|
|
Name:
|
|
fixed: Girls Out West
|
|
URL:
|
|
fixed: https://girlsoutwest.com
|
|
|
|
driver:
|
|
cookies:
|
|
- CookieURL: "https://tour.girlsoutwest.com"
|
|
Cookies:
|
|
- Name: "numheader"
|
|
Domain: "tour.girlsoutwest.com"
|
|
Value: "5"
|
|
Path: "/"
|
|
# Last Updated December 31, 2020
|