60 lines
1.9 KiB
YAML
60 lines
1.9 KiB
YAML
name: "HMV Mania"
|
|
sceneByURL:
|
|
- action: scrapeXPath
|
|
url:
|
|
- hmvmania.com
|
|
scraper: sceneScraper
|
|
|
|
sceneByName:
|
|
action: scrapeXPath
|
|
queryURL: https://hmvmania.com/?post_type=video&s={}
|
|
scraper: sceneSearch
|
|
sceneByQueryFragment:
|
|
action: scrapeXPath
|
|
queryURL: "{url}"
|
|
scraper: sceneScraper
|
|
|
|
xPathScrapers:
|
|
sceneSearch:
|
|
common:
|
|
$searchItem: //div[contains(@class,"main-container")]//div[contains(@class,"video-item-horizontal")]
|
|
scene:
|
|
Title: $searchItem//h4/text()
|
|
URL: $searchItem//a[h4]/@href
|
|
Image: $searchItem//img[not(@class)]/@data-src
|
|
Details: $searchItem//p/text()
|
|
sceneScraper:
|
|
common:
|
|
$infoNode: //div[@class="main-container"]//div[@class="row"]/div[1]
|
|
scene:
|
|
URL: //link[@rel="canonical"]/@href|//meta[@property="og:url"]/@content
|
|
Image: //meta[@property="og:image"]/@content
|
|
Title:
|
|
selector: $infoNode//h1[@class="video-entry-title"]/text()
|
|
postProcess:
|
|
- replace:
|
|
- regex: '^\[.+\]\s(.+)'
|
|
with: $1
|
|
Studio:
|
|
Name:
|
|
selector: $infoNode//h1[@class="video-entry-title"]/text()
|
|
postProcess:
|
|
- replace:
|
|
- regex: '^\[(.+)\]\s.+'
|
|
with: $1
|
|
URL: $infoNode//div[contains(@class,"content-creator")]/div/a/@href
|
|
Details:
|
|
selector: $infoNode//div[contains(@class,"entry-content")]/p/text()
|
|
concat: "\n"
|
|
postProcess:
|
|
- replace:
|
|
- regex: "^HMV Maker: .+"
|
|
with:
|
|
Tags:
|
|
Name: $infoNode//ul[contains(@class,"video-meta")]/li[i[contains(@class,"fa-tags")]]/a/text()
|
|
Date:
|
|
selector: $infoNode//ul[contains(@class,"video-meta")]/li[i[contains(@class,"fa-calendar-alt")]]/text()
|
|
postProcess:
|
|
- parseDate: "Jan 02 2006"
|
|
# Last Updated December 16, 2022
|