Merge pull request #3214 from SamantazFox/fix-scheduled-streams

6 files changed, 167 insertions, 43 deletions

diff --git a/mocks b/mocks
-Subproject 020337194dd482c47ee2d53cd111d0ebf2831e5
+Subproject c401dd9203434b561022242c24b0c200d72284c
diff --git a/spec/invidious/videos/scheduled_live_extract_spec.cr b/spec/invidious/videos/scheduled_live_extract_spec.cr
new file mode 100644
index 00000000..6e531bbd
--- /dev/null
+++ b/spec/invidious/videos/scheduled_live_extract_spec.cr
@@ -0,0 +1,113 @@
+require "../../parsers_helper.cr"
+
+Spectator.describe Invidious::Hashtag do
+  it "parses scheduled livestreams data (test 1)" do
+    # Enable mock
+    _player = load_mock("video/scheduled_live_nintendo.player")
+    _next = load_mock("video/scheduled_live_nintendo.next")
+
+    raw_data = _player.merge!(_next)
+    info = parse_video_info("QMGibBzTu0g", raw_data)
+
+    # Some basic verifications
+    expect(typeof(info)).to eq(Hash(String, JSON::Any))
+
+    expect(info["shortDescription"].as_s).to eq(
+      "Tune in on 6/22 at 7 a.m. PT for a livestreamed Xenoblade Chronicles 3 Direct presentation featuring roughly 20 minutes of information about the upcoming RPG adventure for Nintendo Switch."
+    )
+    expect(info["descriptionHtml"].as_s).to eq(
+      "Tune in on 6/22 at 7 a.m. PT for a livestreamed Xenoblade Chronicles 3 Direct presentation featuring roughly 20 minutes of information about the upcoming RPG adventure for Nintendo Switch."
+    )
+
+    expect(info["likes"].as_i).to eq(2_283)
+
+    expect(info["genre"].as_s).to eq("Gaming")
+    expect(info["genreUrl"].raw).to be_nil
+    expect(info["genreUcid"].as_s).to be_empty
+    expect(info["license"].as_s).to be_empty
+
+    expect(info["authorThumbnail"].as_s).to eq(
+      "https://yt3.ggpht.com/ytc/AKedOLTt4vtjREUUNdHlyu9c4gtJjG90M9jQheRlLKy44A=s48-c-k-c0x00ffffff-no-rj"
+    )
+
+    expect(info["authorVerified"].as_bool).to be_true
+    expect(info["subCountText"].as_s).to eq("8.5M")
+
+    expect(info["relatedVideos"].as_a.size).to eq(20)
+
+    # related video #1
+    expect(info["relatedVideos"][3]["id"].as_s).to eq("a-SN3lLIUEo")
+    expect(info["relatedVideos"][3]["author"].as_s).to eq("Nintendo")
+    expect(info["relatedVideos"][3]["ucid"].as_s).to eq("UCGIY_O-8vW4rfX98KlMkvRg")
+    expect(info["relatedVideos"][3]["view_count"].as_s).to eq("147796")
+    expect(info["relatedVideos"][3]["short_view_count"].as_s).to eq("147K")
+    expect(info["relatedVideos"][3]["author_verified"].as_s).to eq("true")
+
+    # Related video #2
+    expect(info["relatedVideos"][16]["id"].as_s).to eq("l_uC1jFK0lo")
+    expect(info["relatedVideos"][16]["author"].as_s).to eq("Nintendo")
+    expect(info["relatedVideos"][16]["ucid"].as_s).to eq("UCGIY_O-8vW4rfX98KlMkvRg")
+    expect(info["relatedVideos"][16]["view_count"].as_s).to eq("53510")
+    expect(info["relatedVideos"][16]["short_view_count"].as_s).to eq("53K")
+    expect(info["relatedVideos"][16]["author_verified"].as_s).to eq("true")
+  end
+
+  it "parses scheduled livestreams data (test 2)" do
+    # Enable mock
+    _player = load_mock("video/scheduled_live_PBD-Podcast.player")
+    _next = load_mock("video/scheduled_live_PBD-Podcast.next")
+
+    raw_data = _player.merge!(_next)
+    info = parse_video_info("RG0cjYbXxME", raw_data)
+
+    # Some basic verifications
+    expect(typeof(info)).to eq(Hash(String, JSON::Any))
+
+    expect(info["shortDescription"].as_s).to start_with(
+      <<-TXT
+      PBD Podcast Episode 171. In this episode, Patrick Bet-David is joined by Dr. Patrick Moore and Adam Sosnick.
+
+      Join the channel to get exclusive access to perks: https://bit.ly/3Q9rSQL
+      TXT
+    )
+    expect(info["descriptionHtml"].as_s).to start_with(
+      <<-TXT
+      PBD Podcast Episode 171. In this episode, Patrick Bet-David is joined by Dr. Patrick Moore and Adam Sosnick.
+
+      Join the channel to get exclusive access to perks: <a href="https://bit.ly/3Q9rSQL">bit.ly/3Q9rSQL</a>
+      TXT
+    )
+
+    expect(info["likes"].as_i).to eq(22)
+
+    expect(info["genre"].as_s).to eq("Entertainment")
+    expect(info["genreUrl"].raw).to be_nil
+    expect(info["genreUcid"].as_s).to be_empty
+    expect(info["license"].as_s).to be_empty
+
+    expect(info["authorThumbnail"].as_s).to eq(
+      "https://yt3.ggpht.com/61ArDiQshJrvSXcGLhpFfIO3hlMabe2fksitcf6oGob0Mdr5gztdkXxRljICUodL4iuTSrtxW4A=s48-c-k-c0x00ffffff-no-rj"
+    )
+
+    expect(info["authorVerified"].as_bool).to be_false
+    expect(info["subCountText"].as_s).to eq("227K")
+
+    expect(info["relatedVideos"].as_a.size).to eq(20)
+
+    # related video #1
+    expect(info["relatedVideos"][2]["id"]).to eq("La9oLLoI5Rc")
+    expect(info["relatedVideos"][2]["author"]).to eq("Tom Bilyeu")
+    expect(info["relatedVideos"][2]["ucid"]).to eq("UCnYMOamNKLGVlJgRUbamveA")
+    expect(info["relatedVideos"][2]["view_count"]).to eq("13329149")
+    expect(info["relatedVideos"][2]["short_view_count"]).to eq("13M")
+    expect(info["relatedVideos"][2]["author_verified"]).to eq("true")
+
+    # Related video #2
+    expect(info["relatedVideos"][9]["id"]).to eq("IQ_4fvpzYuA")
+    expect(info["relatedVideos"][9]["author"]).to eq("Business Today")
+    expect(info["relatedVideos"][9]["ucid"]).to eq("UCaPHWiExfUWaKsUtENLCv5w")
+    expect(info["relatedVideos"][9]["view_count"]).to eq("26432")
+    expect(info["relatedVideos"][9]["short_view_count"]).to eq("26K")
+    expect(info["relatedVideos"][9]["author_verified"]).to eq("true")
+  end
+end
diff --git a/spec/parsers_helper.cr b/spec/parsers_helper.cr
index 6155fe33..e9154875 100644
--- a/spec/parsers_helper.cr
+++ b/spec/parsers_helper.cr
@@ -6,6 +6,7 @@ require "protodec/utils"
 
 require "spectator"
 
+require "../src/invidious/exceptions"
 require "../src/invidious/helpers/macros"
 require "../src/invidious/helpers/logger"
 require "../src/invidious/helpers/utils"
diff --git a/src/invidious/exceptions.cr b/src/invidious/exceptions.cr
index 471a199a..05be73a6 100644
--- a/src/invidious/exceptions.cr
+++ b/src/invidious/exceptions.cr
@@ -1,3 +1,11 @@
+# InfoExceptions are for displaying information to the user.
+#
+# An InfoException might or might not indicate that something went wrong.
+# Historically Invidious didn't differentiate between these two options, so to
+# maintain previous functionality InfoExceptions do not print backtraces.
+class InfoException < Exception
+end
+
 # Exception used to hold the bogus UCID during a channel search.
 class ChannelSearchException < InfoException
   getter channel : String
diff --git a/src/invidious/helpers/errors.cr b/src/invidious/helpers/errors.cr
index b80dcdaf..6e5a975d 100644
--- a/src/invidious/helpers/errors.cr
+++ b/src/invidious/helpers/errors.cr
@@ -1,11 +1,3 @@
-# InfoExceptions are for displaying information to the user.
-#
-# An InfoException might or might not indicate that something went wrong.
-# Historically Invidious didn't differentiate between these two options, so to
-# maintain previous functionality InfoExceptions do not print backtraces.
-class InfoException < Exception
-end
-
 # -------------------
 #  Issue template
 # -------------------
diff --git a/src/invidious/videos.cr b/src/invidious/videos.cr
index 19ee064c..f87c6b47 100644
--- a/src/invidious/videos.cr
+++ b/src/invidious/videos.cr
@@ -886,13 +886,13 @@ def parse_related_video(related : JSON::Any) : Hash(String, JSON::Any)?
 end
 
 def extract_video_info(video_id : String, proxy_region : String? = nil, context_screen : String? = nil)
-  params = {} of String => JSON::Any
-
+  # Init client config for the API
   client_config = YoutubeAPI::ClientConfig.new(proxy_region: proxy_region)
   if context_screen == "embed"
     client_config.client_type = YoutubeAPI::ClientType::TvHtml5ScreenEmbed
   end
 
+  # Fetch data from the player endpoint
   player_response = YoutubeAPI.player(video_id: video_id, params: "", client_config: client_config)
 
   playability_status = player_response.dig?("playabilityStatus", "status").try &.as_s
@@ -903,26 +903,29 @@ def extract_video_info(video_id : String, proxy_region : String? = nil, context_
     reason ||= subreason.try &.[]("runs").as_a.map(&.[]("text")).join("")
     reason ||= player_response.dig("playabilityStatus", "reason").as_s
 
-    params["reason"] = JSON::Any.new(reason)
-
     # Stop here if video is not a scheduled livestream
     if playability_status != "LIVE_STREAM_OFFLINE"
-      return params
+      return {
+        "reason" => JSON::Any.new(reason),
+      }
     end
+  else
+    reason = nil
   end
 
-  params["shortDescription"] = player_response.dig?("videoDetails", "shortDescription") || JSON::Any.new(nil)
-
   # Don't fetch the next endpoint if the video is unavailable.
-  if !params["reason"]?
+  if {"OK", "LIVE_STREAM_OFFLINE"}.any?(playability_status)
     next_response = YoutubeAPI.next({"videoId": video_id, "params": ""})
     player_response = player_response.merge(next_response)
   end
 
+  params = parse_video_info(video_id, player_response)
+  params["reason"] = JSON::Any.new(reason) if reason
+
   # Fetch the video streams using an Android client in order to get the decrypted URLs and
   # maybe fix throttling issues (#2194).See for the explanation about the decrypted URLs:
   # https://github.com/TeamNewPipe/NewPipeExtractor/issues/562
-  if !params["reason"]?
+  if reason.nil?
     if context_screen == "embed"
       client_config.client_type = YoutubeAPI::ClientType::AndroidScreenEmbed
     else
@@ -940,10 +943,15 @@ def extract_video_info(video_id : String, proxy_region : String? = nil, context_
     end
   end
 
+  # TODO: clean that up
   {"captions", "microformat", "playabilityStatus", "storyboards", "videoDetails"}.each do |f|
     params[f] = player_response[f] if player_response[f]?
   end
 
+  return params
+end
+
+def parse_video_info(video_id : String, player_response : Hash(String, JSON::Any)) : Hash(String, JSON::Any)
   # Top level elements
 
   main_results = player_response.dig?("contents", "twoColumnWatchNextResults")
@@ -997,8 +1005,6 @@ def extract_video_info(video_id : String, proxy_region : String? = nil, context_
     end
   end
 
-  params["relatedVideos"] = JSON::Any.new(related)
-
   # Likes
 
   toplevel_buttons = video_primary_renderer
@@ -1019,42 +1025,36 @@ def extract_video_info(video_id : String, proxy_region : String? = nil, context_
     end
   end
 
-  params["likes"] = JSON::Any.new(likes || 0_i64)
-  params["dislikes"] = JSON::Any.new(0_i64)
-
   # Description
 
+  short_description = player_response.dig?("videoDetails", "shortDescription")
+
   description_html = video_secondary_renderer.try &.dig?("description", "runs")
     .try &.as_a.try { |t| content_to_comment_html(t, video_id) }
 
-  params["descriptionHtml"] = JSON::Any.new(description_html || "<p></p>")
-
   # Video metadata
 
   metadata = video_secondary_renderer
     .try &.dig?("metadataRowContainer", "metadataRowContainerRenderer", "rows")
       .try &.as_a
 
-  params["genre"] = params["microformat"]?.try &.["playerMicroformatRenderer"]?.try &.["category"]? || JSON::Any.new("")
-  params["genreUrl"] = JSON::Any.new(nil)
+  genre = player_response.dig?("microformat", "playerMicroformatRenderer", "category")
+  genre_ucid = nil
+  license = nil
 
   metadata.try &.each do |row|
-    title = row["metadataRowRenderer"]?.try &.["title"]?.try &.["simpleText"]?.try &.as_s
+    metadata_title = row.dig?("metadataRowRenderer", "title", "simpleText").try &.as_s
     contents = row.dig?("metadataRowRenderer", "contents", 0)
 
-    if title.try &.== "Category"
+    if metadata_title == "Category"
       contents = contents.try &.dig?("runs", 0)
 
-      params["genre"] = JSON::Any.new(contents.try &.["text"]?.try &.as_s || "")
-      params["genreUcid"] = JSON::Any.new(contents.try &.["navigationEndpoint"]?.try &.["browseEndpoint"]?
-        .try &.["browseId"]?.try &.as_s || "")
-    elsif title.try &.== "License"
-      contents = contents.try &.["runs"]?
-        .try &.as_a[0]?
-
-      params["license"] = JSON::Any.new(contents.try &.["text"]?.try &.as_s || "")
-    elsif title.try &.== "Licensed to YouTube by"
-      params["license"] = JSON::Any.new(contents.try &.["simpleText"]?.try &.as_s || "")
+      genre = contents.try &.["text"]?
+      genre_ucid = contents.try &.dig?("navigationEndpoint", "browseEndpoint", "browseId")
+    elsif metadata_title == "License"
+      license = contents.try &.dig?("runs", 0, "text")
+    elsif metadata_title == "Licensed to YouTube by"
+      license = contents.try &.["simpleText"]?
     end
   end
 
@@ -1062,20 +1062,30 @@ def extract_video_info(video_id : String, proxy_region : String? = nil, context_
 
   if author_info = video_secondary_renderer.try &.dig?("owner", "videoOwnerRenderer")
     author_thumbnail = author_info.dig?("thumbnail", "thumbnails", 0, "url")
-    params["authorThumbnail"] = JSON::Any.new(author_thumbnail.try &.as_s || "")
-
     author_verified = has_verified_badge?(author_info["badges"]?)
-    params["authorVerified"] = JSON::Any.new(author_verified)
 
     subs_text = author_info["subscriberCountText"]?
       .try { |t| t["simpleText"]? || t.dig?("runs", 0, "text") }
       .try &.as_s.split(" ", 2)[0]
-
-    params["subCountText"] = JSON::Any.new(subs_text || "-")
   end
 
   # Return data
 
+  params = {
+    "shortDescription" => JSON::Any.new(short_description.try &.as_s || nil),
+    "relatedVideos"    => JSON::Any.new(related),
+    "likes"            => JSON::Any.new(likes || 0_i64),
+    "dislikes"         => JSON::Any.new(0_i64),
+    "descriptionHtml"  => JSON::Any.new(description_html || "<p></p>"),
+    "genre"            => JSON::Any.new(genre.try &.as_s || ""),
+    "genreUrl"         => JSON::Any.new(nil),
+    "genreUcid"        => JSON::Any.new(genre_ucid.try &.as_s || ""),
+    "license"          => JSON::Any.new(license.try &.as_s || ""),
+    "authorThumbnail"  => JSON::Any.new(author_thumbnail.try &.as_s || ""),
+    "authorVerified"   => JSON::Any.new(author_verified),
+    "subCountText"     => JSON::Any.new(subs_text || "-"),
+  }
+
   return params
 end