From 43bd331e48ad1a19cd3c7a6d5beb72e3127c5edc Mon Sep 17 00:00:00 2001
From: Samantaz Fox <coding@samantaz.fr>
Date: Thu, 1 Apr 2021 02:36:43 +0000
Subject: [PATCH] Multiple youtube_api.cr helper fixes

Add documentation
Bump web client version string
Add charset=UTF-8 to the 'content-type' header
Parse JSON and return it as a Hash
Handle API error messages
---
 src/invidious/channels.cr            | 40 +++++++---------------------
 src/invidious/helpers/youtube_api.cr | 29 +++++++++++++++++---
 src/invidious/playlists.cr           |  2 +-
 src/invidious/search.cr              |  3 +--
 4 files changed, 36 insertions(+), 38 deletions(-)

diff --git a/src/invidious/channels.cr b/src/invidious/channels.cr
index 3109b508..bbef3d4f 100644
--- a/src/invidious/channels.cr
+++ b/src/invidious/channels.cr
@@ -229,22 +229,8 @@ def fetch_channel(ucid, db, pull_all_videos = true, locale = nil)
   page = 1
 
   LOGGER.trace("fetch_channel: #{ucid} : Downloading channel videos page")
-  response_body = get_channel_videos_response(ucid, page, auto_generated: auto_generated)
-
-  videos = [] of SearchVideo
-  begin
-    initial_data = JSON.parse(response_body)
-    raise InfoException.new("Could not extract channel JSON") if !initial_data
-
-    LOGGER.trace("fetch_channel: #{ucid} : Extracting videos from channel videos page initial_data")
-    videos = extract_videos(initial_data.as_h, author, ucid)
-  rescue ex
-    if response_body.includes?("To continue with your YouTube experience, please fill out the form below.") ||
-       response_body.includes?("https://www.google.com/sorry/index")
-      raise InfoException.new("Could not extract channel info. Instance is likely blocked.")
-    end
-    raise ex
-  end
+  initial_data = get_channel_videos_response(ucid, page, auto_generated: auto_generated)
+  videos = extract_videos(initial_data, author, ucid)
 
   LOGGER.trace("fetch_channel: #{ucid} : Extracting videos from channel RSS feed")
   rss.xpath_nodes("//feed/entry").each do |entry|
@@ -304,10 +290,8 @@ def fetch_channel(ucid, db, pull_all_videos = true, locale = nil)
     ids = [] of String
 
     loop do
-      response_body = get_channel_videos_response(ucid, page, auto_generated: auto_generated)
-      initial_data = JSON.parse(response_body)
-      raise InfoException.new("Could not extract channel JSON") if !initial_data
-      videos = extract_videos(initial_data.as_h, author, ucid)
+      initial_data = get_channel_videos_response(ucid, page, auto_generated: auto_generated)
+      videos = extract_videos(initial_data, author, ucid)
 
       count = videos.size
       videos = videos.map { |video| ChannelVideo.new({
@@ -358,8 +342,7 @@ end
 def fetch_channel_playlists(ucid, author, continuation, sort_by)
   if continuation
     response_json = request_youtube_api_browse(continuation)
-    result = JSON.parse(response_json)
-    continuationItems = result["onResponseReceivedActions"]?
+    continuationItems = response_json["onResponseReceivedActions"]?
       .try &.[0]["appendContinuationItemsAction"]["continuationItems"]
 
     return [] of SearchItem, nil if !continuationItems
@@ -964,21 +947,16 @@ def get_60_videos(ucid, author, page, auto_generated, sort_by = "newest")
   videos = [] of SearchVideo
 
   2.times do |i|
-    response_json = get_channel_videos_response(ucid, page * 2 + (i - 1), auto_generated: auto_generated, sort_by: sort_by)
-    initial_data = JSON.parse(response_json)
-    break if !initial_data
-    videos.concat extract_videos(initial_data.as_h, author, ucid)
+    initial_data = get_channel_videos_response(ucid, page * 2 + (i - 1), auto_generated: auto_generated, sort_by: sort_by)
+    videos.concat extract_videos(initial_data, author, ucid)
   end
 
   return videos.size, videos
 end
 
 def get_latest_videos(ucid)
-  response_json = get_channel_videos_response(ucid)
-  initial_data = JSON.parse(response_json)
-  return [] of SearchVideo if !initial_data
+  initial_data = get_channel_videos_response(ucid)
   author = initial_data["metadata"]?.try &.["channelMetadataRenderer"]?.try &.["title"]?.try &.as_s
-  items = extract_videos(initial_data.as_h, author, ucid)
 
-  return items
+  return extract_videos(initial_data, author, ucid)
 end
diff --git a/src/invidious/helpers/youtube_api.cr b/src/invidious/helpers/youtube_api.cr
index 30413532..84e0c38f 100644
--- a/src/invidious/helpers/youtube_api.cr
+++ b/src/invidious/helpers/youtube_api.cr
@@ -4,8 +4,18 @@
 
 # Hard-coded constants required by the API
 HARDCODED_API_KEY     = "AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8"
-HARDCODED_CLIENT_VERS = "2.20210318.08.00"
+HARDCODED_CLIENT_VERS = "2.20210330.08.00"
 
+####################################################################
+# request_youtube_api_browse(continuation)
+#
+# Requests the youtubei/vi/browse endpoint with the required headers
+# to get JSON in en-US (english US).
+#
+# The requested data is a continuation token (ctoken). Depending on
+# this token's contents, the returned data can be comments, playlist
+# videos, search results, channel community tab, ...
+#
 def request_youtube_api_browse(continuation)
   # JSON Request data, required by the API
   data = {
@@ -20,12 +30,23 @@ def request_youtube_api_browse(continuation)
     "continuation": continuation,
   }
 
-  # Send the POST request and return result
+  # Send the POST request and parse result
   response = YT_POOL.client &.post(
     "/youtubei/v1/browse?key=#{HARDCODED_API_KEY}",
-    headers: HTTP::Headers{"content-type" => "application/json"},
+    headers: HTTP::Headers{"content-type" => "application/json; charset=UTF-8"},
     body: data.to_json
   )
 
-  return response.body
+  initial_data = JSON.parse(response.body).as_h
+
+  # Error handling
+  if initial_data.has_key?("error")
+    code = initial_data["error"]["code"]
+    message = initial_data["error"]["message"].to_s.sub(/(\\n)+\^$/, "")
+
+    raise InfoException.new("Could not extract JSON. Youtube API returned \
+      error #{code} with message:<br>\"#{message}\"")
+  end
+
+  return initial_data
 end
diff --git a/src/invidious/playlists.cr b/src/invidious/playlists.cr
index 073a9986..150f1c15 100644
--- a/src/invidious/playlists.cr
+++ b/src/invidious/playlists.cr
@@ -451,7 +451,7 @@ def get_playlist_videos(db, playlist, offset, locale = nil, continuation = nil)
       offset = (offset / 100).to_i64 * 100_i64
 
       ctoken = produce_playlist_continuation(playlist.id, offset)
-      initial_data = JSON.parse(request_youtube_api_browse(ctoken)).as_h
+      initial_data = request_youtube_api_browse(ctoken)
     else
       response = YT_POOL.client &.get("/playlist?list=#{playlist.id}&gl=US&hl=en")
       initial_data = extract_initial_data(response.body)
diff --git a/src/invidious/search.cr b/src/invidious/search.cr
index 4b216613..7c9c389e 100644
--- a/src/invidious/search.cr
+++ b/src/invidious/search.cr
@@ -246,8 +246,7 @@ def channel_search(query, page, channel)
   continuation = produce_channel_search_continuation(ucid, query, page)
   response_json = request_youtube_api_browse(continuation)
 
-  result = JSON.parse(response_json)
-  continuationItems = result["onResponseReceivedActions"]?
+  continuationItems = response_json["onResponseReceivedActions"]?
     .try &.[0]["appendContinuationItemsAction"]["continuationItems"]
 
   return 0, [] of SearchItem if !continuationItems