Use new API to fetch videos from channels

This mirrors the process used by subscriptions.gir.st. The old API is
tried first, and if it fails then the new one is used.
This commit is contained in:
afuous 2020-08-29 07:52:30 -07:00
parent 13f58d602f
commit b43866eeda
2 changed files with 197 additions and 151 deletions

View File

@ -396,7 +396,7 @@ def fetch_channel_playlists(ucid, author, auto_generated, continuation, sort_by)
return items, continuation return items, continuation
end end
def produce_channel_videos_url(ucid, page = 1, auto_generated = nil, sort_by = "newest") def produce_channel_videos_url(ucid, page = 1, auto_generated = nil, sort_by = "newest", v2 = false)
object = { object = {
"80226972:embedded" => { "80226972:embedded" => {
"2:string" => ucid, "2:string" => ucid,
@ -411,6 +411,7 @@ def produce_channel_videos_url(ucid, page = 1, auto_generated = nil, sort_by = "
}, },
} }
if !v2
if auto_generated if auto_generated
seed = Time.unix(1525757349) seed = Time.unix(1525757349)
until seed >= Time.utc until seed >= Time.utc
@ -424,6 +425,20 @@ def produce_channel_videos_url(ucid, page = 1, auto_generated = nil, sort_by = "
object["80226972:embedded"]["3:base64"].as(Hash)["4:varint"] = 0_i64 object["80226972:embedded"]["3:base64"].as(Hash)["4:varint"] = 0_i64
object["80226972:embedded"]["3:base64"].as(Hash)["15:string"] = "#{page}" object["80226972:embedded"]["3:base64"].as(Hash)["15:string"] = "#{page}"
end end
else
object["80226972:embedded"]["3:base64"].as(Hash)["4:varint"] = 0_i64
object["80226972:embedded"]["3:base64"].as(Hash)["61:string"] = Base64.urlsafe_encode(Protodec::Any.from_json(Protodec::Any.cast_json({
"1:embedded" => {
"1:varint" => 6307666885028338688_i64,
"2:embedded" => {
"1:string" => Base64.urlsafe_encode(Protodec::Any.from_json(Protodec::Any.cast_json({
"1:varint" => 30_i64 * (page - 1),
}))),
},
},
})))
end
case sort_by case sort_by
when "newest" when "newest"
@ -904,12 +919,25 @@ end
def get_60_videos(ucid, author, page, auto_generated, sort_by = "newest") def get_60_videos(ucid, author, page, auto_generated, sort_by = "newest")
videos = [] of SearchVideo videos = [] of SearchVideo
2.times do |i| needs_v2 = false
url = produce_channel_videos_url(ucid, page * 2 + (i - 1), auto_generated: auto_generated, sort_by: sort_by)
i = 0
while i < 2
url = produce_channel_videos_url(ucid, page * 2 + (i - 1), auto_generated: auto_generated, sort_by: sort_by, v2: needs_v2)
response = YT_POOL.client &.get(url) response = YT_POOL.client &.get(url)
initial_data = JSON.parse(response.body).as_a.find &.["response"]? initial_data = JSON.parse(response.body).as_a.find &.["response"]?
break if !initial_data break if !initial_data
v1_error = !needs_v2 && initial_data
.try &.["response"]?.try &.["alerts"]?
.try &.as_a.any? { |alert|
alert.try &.["alertRenderer"]?.try &.["type"]?.try { |t| t == "ERROR" }
}
if v1_error
needs_v2 = true
else
videos.concat extract_videos(initial_data.as_h, author, ucid) videos.concat extract_videos(initial_data.as_h, author, ucid)
i += 1
end
end end
return videos.size, videos return videos.size, videos

View File

@ -164,20 +164,8 @@ def extract_videos(initial_data : Hash(String, JSON::Any), author_fallback : Str
extract_items(initial_data, author_fallback, author_id_fallback).select(&.is_a?(SearchVideo)).map(&.as(SearchVideo)) extract_items(initial_data, author_fallback, author_id_fallback).select(&.is_a?(SearchVideo)).map(&.as(SearchVideo))
end end
def extract_items(initial_data : Hash(String, JSON::Any), author_fallback : String? = nil, author_id_fallback : String? = nil) def extract_item(item : JSON::Any, author_fallback : String? = nil, author_id_fallback : String? = nil)
items = [] of SearchItem if i = (item["videoRenderer"]? || item["gridVideoRenderer"]?)
initial_data.try { |t| t["contents"]? || t["response"]? }
.try { |t| t["twoColumnBrowseResultsRenderer"]?.try &.["tabs"].as_a.select(&.["tabRenderer"]?.try &.["selected"].as_bool)[0]?.try &.["tabRenderer"]["content"] ||
t["twoColumnSearchResultsRenderer"]?.try &.["primaryContents"] ||
t["continuationContents"]? }
.try { |t| t["sectionListRenderer"]? || t["sectionListContinuation"]? }
.try &.["contents"].as_a
.each { |c| c.try &.["itemSectionRenderer"]?.try &.["contents"].as_a
.try { |t| t[0]?.try &.["shelfRenderer"]?.try &.["content"]["expandedShelfContentsRenderer"]?.try &.["items"].as_a ||
t[0]?.try &.["gridRenderer"]?.try &.["items"].as_a || t }
.each { |item|
if i = item["videoRenderer"]?
video_id = i["videoId"].as_s video_id = i["videoId"].as_s
title = i["title"].try { |t| t["simpleText"]?.try &.as_s || t["runs"]?.try &.as_a.map(&.["text"].as_s).join("") } || "" title = i["title"].try { |t| t["simpleText"]?.try &.as_s || t["runs"]?.try &.as_a.map(&.["text"].as_s).join("") } || ""
@ -212,7 +200,7 @@ def extract_items(initial_data : Hash(String, JSON::Any), author_fallback : Stri
end end
end end
items << SearchVideo.new({ SearchVideo.new({
title: title, title: title,
id: video_id, id: video_id,
author: author, author: author,
@ -238,7 +226,7 @@ def extract_items(initial_data : Hash(String, JSON::Any), author_fallback : Stri
video_count = i["videoCountText"]?.try &.["runs"].as_a[0]?.try &.["text"].as_s.gsub(/\D/, "").to_i || 0 video_count = i["videoCountText"]?.try &.["runs"].as_a[0]?.try &.["text"].as_s.gsub(/\D/, "").to_i || 0
description_html = i["descriptionSnippet"]?.try { |t| parse_content(t) } || "" description_html = i["descriptionSnippet"]?.try { |t| parse_content(t) } || ""
items << SearchChannel.new({ SearchChannel.new({
author: author, author: author,
ucid: author_id, ucid: author_id,
author_thumbnail: author_thumbnail, author_thumbnail: author_thumbnail,
@ -254,7 +242,7 @@ def extract_items(initial_data : Hash(String, JSON::Any), author_fallback : Stri
video_count = i["videoCountText"]["runs"].as_a[0]?.try &.["text"].as_s.gsub(/\D/, "").to_i || 0 video_count = i["videoCountText"]["runs"].as_a[0]?.try &.["text"].as_s.gsub(/\D/, "").to_i || 0
playlist_thumbnail = i["thumbnail"]["thumbnails"][0]?.try &.["url"]?.try &.as_s || "" playlist_thumbnail = i["thumbnail"]["thumbnails"][0]?.try &.["url"]?.try &.as_s || ""
items << SearchPlaylist.new({ SearchPlaylist.new({
title: title, title: title,
id: plid, id: plid,
author: author_fallback || "", author: author_fallback || "",
@ -288,7 +276,7 @@ def extract_items(initial_data : Hash(String, JSON::Any), author_fallback : Stri
# TODO: i["publishedTimeText"]? # TODO: i["publishedTimeText"]?
items << SearchPlaylist.new({ SearchPlaylist.new({
title: title, title: title,
id: plid, id: plid,
author: author, author: author,
@ -305,7 +293,37 @@ def extract_items(initial_data : Hash(String, JSON::Any), author_fallback : Stri
elsif i = item["horizontalCardListRenderer"]? elsif i = item["horizontalCardListRenderer"]?
elsif i = item["searchPyvRenderer"]? # Ad elsif i = item["searchPyvRenderer"]? # Ad
end end
end
def extract_items(initial_data : Hash(String, JSON::Any), author_fallback : String? = nil, author_id_fallback : String? = nil)
items = [] of SearchItem
channel_v2_response = initial_data
.try &.["response"]?
.try &.["continuationContents"]?
.try &.["gridContinuation"]?
.try &.["items"]?
if channel_v2_response
channel_v2_response.try &.as_a.each { |item|
extract_item(item, author_fallback, author_id_fallback)
.try { |t| items << t }
}
else
initial_data.try { |t| t["contents"]? || t["response"]? }
.try { |t| t["twoColumnBrowseResultsRenderer"]?.try &.["tabs"].as_a.select(&.["tabRenderer"]?.try &.["selected"].as_bool)[0]?.try &.["tabRenderer"]["content"] ||
t["twoColumnSearchResultsRenderer"]?.try &.["primaryContents"] ||
t["continuationContents"]? }
.try { |t| t["sectionListRenderer"]? || t["sectionListContinuation"]? }
.try &.["contents"].as_a
.each { |c| c.try &.["itemSectionRenderer"]?.try &.["contents"].as_a
.try { |t| t[0]?.try &.["shelfRenderer"]?.try &.["content"]["expandedShelfContentsRenderer"]?.try &.["items"].as_a ||
t[0]?.try &.["gridRenderer"]?.try &.["items"].as_a || t }
.each { |item|
extract_item(item, author_fallback, author_id_fallback)
.try { |t| items << t }
} } } }
end
items items
end end