Merge pull request #1533 from smcallah/patch-1

Updated AVJiali.yml and Tenshigao.yml with sceneByName scrapers and cleaned up the code
stashapp · Nov 7, 2023 · 50a4f64 · 50a4f64
2 parents beeaadd + 6b73583
commit 50a4f64
Show file tree

Hide file tree

Showing 2 changed files with 60 additions and 17 deletions.
diff --git a/scrapers/AVJiali.yml b/scrapers/AVJiali.yml
@@ -9,6 +9,14 @@ performerByURL:
     url:
       - avjiali.com
     scraper: performerScraper
+sceneByName:
+  action: scrapeXPath
+  queryURL: https://avjiali.com/?s={}
+  scraper: sceneSearch
+sceneByQueryFragment:
+  action: scrapeXPath
+  queryURL: "{url}"
+  scraper: sceneScraper
 xPathScrapers:
   sceneScraper:
     scene:
@@ -38,6 +46,7 @@ xPathScrapers:
           - replace:
               - regex: ^//
                 with: "https://"
+      URL: //link[@rel="canonical"]/@href
       Studio:
         Name:
           fixed: AV Jiali
@@ -101,43 +110,59 @@ xPathScrapers:
                 with: $1
               - regex: None
                 with:
-      Details: 
-        selector: //*[@id="main"]/div[1]/div[2]/p/text()
+      Details:
+        selector: //p[span[@class="readmore"]]
+        postProcess:
+          - replace:
+              - regex: ... Read More
+                with:
       Piercings:
         selector: $profile[contains(strong, "Piercings:")]//text()
         postProcess:
           - replace:
               - regex: .*Piercings:\s*(.*)\s*$
                 with: $1
-              - regex: None
-                with:
-              - regex: No Piercings
+              - regex: None|^No$|No Piercing.*
                 with:
       Tattoos:
         selector: $profile[contains(strong, "Tattoo:")]//text()
         postProcess:
           - replace:
               - regex: .*Tattoo:\s*(.*)\s*$
                 with: $1
-              - regex: None
+              - regex: None|^No$|No Tattoo.*
                 with:
       HairColor:
         selector: $profile[contains(strong, "Hair color:")]//text()
         postProcess:
           - replace:
-              - regex: .*Hair Color:?\s*(.*)\s*$
-                with: $1
-              - regex: .*Hair color:?\s*(.*)\s*$
+              - regex: (?i).*hair color:?\s*(.*)\s*$
                 with: $1
               - regex: None
                 with:
       URL: //link[@rel="canonical"][1]/@href
       Image:
-        selector: //img[@class="rounded"]/@src
+        selector: //div[@class="model-photo"]/img[@class="rounded"]/@src
         postProcess:
           - replace:
               - regex: ^//
                 with: https://
-              - regex: 160x160
-                with: 500x500
-# Last Updated October 03, 2023
+              - regex: 460x640
+                with: 690x960
+      Ethnicity:
+        fixed: Asian
+      Gender:
+        fixed: Female
+      Country:
+        fixed: Taiwan
+  sceneSearch:
+    common:
+      $videos: //div[@class="thumb"]/a
+    scene:
+      Title: $videos/@title 
+      URL: $videos/@href
+      Image: $videos/img/@src
+      Studio:
+        Name:
+          fixed: AVJiali
+# Last Updated November 07, 2023
diff --git a/scrapers/Tenshigao.yml b/scrapers/Tenshigao.yml
@@ -9,6 +9,14 @@ performerByURL:
     url:
       - tenshigao.com
     scraper: performerScraper
+sceneByName:
+  action: scrapeXPath
+  queryURL: https://tenshigao.com/?s={}
+  scraper: sceneSearch
+sceneByQueryFragment:
+  action: scrapeXPath
+  queryURL: "{url}"
+  scraper: sceneScraper
 xPathScrapers:
   sceneScraper:
     common: 
@@ -105,7 +113,7 @@ xPathScrapers:
               - regex: None
                 with:
       Details: 
-        selector: //p[span[@class="readmore"]]
+        selector: //p[span[@class="readmore"]] | //div[@class="intro"]/p
         postProcess:
           - replace:
               - regex: ... Read More
@@ -116,15 +124,15 @@ xPathScrapers:
           - replace:
               - regex: .*Piercings:\s*(.*)\s*$
                 with: $1
-              - regex: None|^No$|No Piercings
+              - regex: None|^No$|No Piercing.*
                 with:
       Tattoos:
         selector: $profile[contains(strong, "Tattoo:")]//text()
         postProcess:
           - replace:
               - regex: .*Tattoo:\s*(.*)\s*$
                 with: $1
-              - regex: None|^No$|No Tattoos
+              - regex: None|^No$|No Tattoo.*
                 with:
       HairColor:
         selector: $profile[contains(strong, "Hair color:")]//text()
@@ -149,4 +157,14 @@ xPathScrapers:
                 with: https://
               - regex: 160x160
                 with: 500x500
-# Last Updated October 24, 2023
+  sceneSearch:
+    common:
+      $videos: //div[@class="thumb"]/a
+    scene:
+      Title: $videos/@title 
+      URL: $videos/@href
+      Image: $videos/img/@src
+      Studio:
+        Name:
+          fixed: Tenshigao
+# Last Updated November 07, 2023