forki 0.2.15 → 0.2.17

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 252b9aab5a0269780c129e75249f339231febbdf72f27b93b18d637370a2d395
-  data.tar.gz: 1594e70cf96cf0f9da6724b5b64eb39a19f8c1d9cbde5d8621166240e3a3212a
+  metadata.gz: 1d9500f6b5531d22bd8f1a227c9f7c241bacd594221bf4ce9d717d98090411b1
+  data.tar.gz: 1bd7ac04cce0047d98151671cdf7812e99d50701ed3ba7877c569256f7de40f9
 SHA512:
-  metadata.gz: b07a698ca5052ac2391f6e1814cba9610bec5743dd53b4f69fd0ef2592856590a309fd9f10947b512e8f1c20f9b32880d9b908c51005bdfffc56f509735cda30
-  data.tar.gz: db40445e250f8ecb84578653b2355a63266a5f138d92e1b0a1bc4961643ceab5ee925ea7c074e96ae449268d98ee87684a5cc9532a7e51f1076484013421c7fd
+  metadata.gz: 46c9fa11210476a43629ff6870207227b657912c3cacd9ca4b84d63d1391219ac8964d9ca9581e3c24e5ceec0afe816c252605bb07175a1d9f81f00507b01c7d
+  data.tar.gz: 53adaa95825c4f27987d9fe2ee5dd873eb4da28c2415961a61a092240c0691482ba79fecbc1681f4aa8df22c5fff826ad1c1f94c25c5a191bbc9f8aa1b884fcc

data/lib/forki/scrapers/post_scraper.rb

@@ -33,12 +33,12 @@ module Forki
 
       if post_is_text_only
         extract_text_post_data(graphql_objects)
+      elsif post_has_image && !post_has_video && !post_has_video_in_comment_stream
+        extract_image_post_data(graphql_objects)
       elsif post_has_video
         extract_video_post_data(graphql_strings)
       elsif post_has_video_in_comment_stream
         extract_video_comment_post_data(graphql_objects)
-      elsif post_has_image
-        extract_image_post_data(graphql_objects)
       else
         extract_image_post_data(graphql_objects)
       end
@@ -84,6 +84,7 @@ module Forki
         return true unless graphql_object.fetch("image", nil).nil? # so long as the associated values are not nil
         return true unless graphql_object.fetch("currMedia", nil).nil?
         return true unless graphql_object.fetch("photo_image", nil).nil?
+        return true unless graphql_object.fetch("large_share_image", nil).nil?
 
         # This is a complicated form for `web.facebook.com` posts
         if !graphql_object.dig("node", "comet_sections", "content", "story", "attachments").nil?
@@ -329,11 +330,10 @@ module Forki
         reshare_warning = feedback_object["comet_ufi_summary_and_actions_renderer"]["feedback"]["should_show_reshare_warning"]
       end
 
-      if !video_object.key?("browser_native_hd_url") && !video_object.key?("browser_native_sd_url") && video_object.key?("videoDeliveryLegacyFields")
-        video_url = video_object["videoDeliveryLegacyFields"]["browser_native_hd_url"] || video_object["videoDeliveryLegacyFields"]["browser_native_sd_url"]
-      else
-        video_url = video_object["browser_native_hd_url"] || video_object["browser_native_sd_url"]
-      end
+      video_object_url_subsearch = video_object
+      video_object_url_subsearch = video_object_url_subsearch["videoDeliveryLegacyFields"] if video_object_url_subsearch.has_key?("videoDeliveryLegacyFields")
+      # if video_object.key?("browser_native_hd_url") || video_object.key?("browser_native_sd_url")
+      video_url = video_object_url_subsearch["browser_native_hd_url"] || video_object_url_subsearch["browser_native_sd_url"]
 
       post_details = {
         id: video_object["id"],
@@ -442,6 +442,10 @@ module Forki
           if image_url.nil?
             image_url = attachments.first&.dig("styles", "attachment", "media", "large_share_image", "uri")
           end
+
+          if image_url.nil?
+            image_url = attachments.first&.dig("styles", "attachment", "media", "image", "uri")
+          end
         end
 
         text = graphql_object["node"]["comet_sections"]["content"]["story"].dig("message", "text")

data/lib/forki/scrapers/sieves/video_sieves/video_sieve_watch_tab.rb

@@ -46,9 +46,16 @@ class VideoSieveWatchTab < VideoSieve
     video_url = video_object.dig("short_form_video_context", "playback_video", "browser_native_hd_url") if video_url.nil?
     video_url = video_object.dig("short_form_video_context", "playback_video", "browser_native_sd_url") if video_url.nil?
 
+    video_url = video_object["attachments"]&.first.dig("media", "videoDeliveryLegacyFields", "browser_native_hd_url") if video_url.nil?
+    video_url = video_object["attachments"]&.first.dig("media", "videoDeliveryLegacyFields", "browser_native_sd_url") if video_url.nil?
+
+    raise VideoSieveFailedError.new(sieve_class: "VideoSieveWatchTab") if video_url.nil?
+
     video_preview_image_url = video_object["attachments"]&.first.dig("media", "preferred_thumbnail", "image", "uri")
     video_preview_image_url = video_object["short_form_video_context"]["video"]["first_frame_thumbnail"] if video_preview_image_url.nil?
 
+    raise VideoSieveFailedError.new(sieve_class: "VideoSieveWatchTab") if video_preview_image_url.nil?
+
     if !video_object["feedback_context"].nil?
       feedback_object = video_object["feedback_context"]["feedback_target_with_context"]
     else

data/lib/forki/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Forki
-  VERSION = "0.2.15"
+  VERSION = "0.2.17"
 end

data/lib/forki.rb

@@ -48,6 +48,13 @@ module Forki
     end
   end
 
+  class VideoSieveFailedError < StandardError
+    def initialize(msg = "Video sieve failed to find a video", sieve_class: VideoSieve)
+      self.msg = "#{sieve_class} failed to find a video" if msg.nil?
+      super
+    end
+  end
+
   define_setting :temp_storage_location, "tmp/forki"
 
   # Extract the file extension from a media URL

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: forki
 version: !ruby/object:Gem::Version
-  version: 0.2.15
+  version: 0.2.17
 platform: ruby
 authors:
 - ''
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2024-10-25 00:00:00.000000000 Z
+date: 2024-11-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: capybara