invidious-mod-JP/src/invidious/channels.cr

1085 lines
34 KiB
Crystal
Raw Normal View History

2019-03-29 21:30:02 +00:00
struct InvidiousChannel
db_mapping({
id: String,
author: String,
updated: Time,
deleted: Bool,
2019-03-04 02:40:24 +00:00
subscribed: Time?,
2018-08-04 20:30:44 +00:00
})
end
2019-03-29 21:30:02 +00:00
struct ChannelVideo
def to_json(locale, config, kemal_config, json : JSON::Builder)
json.object do
json.field "type", "shortVideo"
json.field "title", self.title
json.field "videoId", self.id
json.field "videoThumbnails" do
generate_thumbnails(json, self.id, config, Kemal.config)
end
json.field "lengthSeconds", self.length_seconds
json.field "author", self.author
json.field "authorId", self.ucid
json.field "authorUrl", "/channel/#{self.ucid}"
json.field "published", self.published.to_unix
json.field "publishedText", translate(locale, "`x` ago", recode_date(self.published, locale))
2019-05-30 20:09:39 +00:00
json.field "viewCount", self.views
end
end
def to_json(locale, config, kemal_config, json : JSON::Builder | Nil = nil)
if json
to_json(locale, config, kemal_config, json)
else
JSON.build do |json|
to_json(locale, config, kemal_config, json)
end
end
end
2019-06-07 17:39:12 +00:00
def to_xml(locale, host_url, xml : XML::Builder)
xml.element("entry") do
xml.element("id") { xml.text "yt:video:#{self.id}" }
xml.element("yt:videoId") { xml.text self.id }
xml.element("yt:channelId") { xml.text self.ucid }
xml.element("title") { xml.text self.title }
xml.element("link", rel: "alternate", href: "#{host_url}/watch?v=#{self.id}")
xml.element("author") do
xml.element("name") { xml.text self.author }
xml.element("uri") { xml.text "#{host_url}/channel/#{self.ucid}" }
end
xml.element("content", type: "xhtml") do
xml.element("div", xmlns: "http://www.w3.org/1999/xhtml") do
xml.element("a", href: "#{host_url}/watch?v=#{self.id}") do
xml.element("img", src: "#{host_url}/vi/#{self.id}/mqdefault.jpg")
end
end
end
xml.element("published") { xml.text self.published.to_s("%Y-%m-%dT%H:%M:%S%:z") }
xml.element("updated") { xml.text self.updated.to_s("%Y-%m-%dT%H:%M:%S%:z") }
xml.element("media:group") do
xml.element("media:title") { xml.text self.title }
xml.element("media:thumbnail", url: "#{host_url}/vi/#{self.id}/mqdefault.jpg",
width: "320", height: "180")
end
end
end
def to_xml(locale, config, kemal_config, xml : XML::Builder | Nil = nil)
if xml
to_xml(locale, config, kemal_config, xml)
else
XML.build do |xml|
to_xml(locale, config, kemal_config, xml)
end
end
end
db_mapping({
id: String,
title: String,
published: Time,
updated: Time,
ucid: String,
author: String,
length_seconds: {type: Int32, default: 0},
live_now: {type: Bool, default: false},
premiere_timestamp: {type: Time?, default: nil},
2019-05-30 20:09:39 +00:00
views: {type: Int64?, default: nil},
2018-08-04 20:30:44 +00:00
})
end
struct AboutRelatedChannel
db_mapping({
ucid: String,
author: String,
author_url: String,
author_thumbnail: String,
})
end
# TODO: Refactor into either SearchChannel or InvidiousChannel
struct AboutChannel
db_mapping({
ucid: String,
author: String,
auto_generated: Bool,
author_url: String,
author_thumbnail: String,
banner: String?,
description_html: String,
paid: Bool,
total_views: Int64,
sub_count: Int64,
joined: Time,
is_family_friendly: Bool,
allowed_regions: Array(String),
related_channels: Array(AboutRelatedChannel),
2019-07-09 14:31:04 +00:00
tabs: Array(String),
})
end
def get_batch_channels(channels, db, refresh = false, pull_all_videos = true, max_threads = 10)
2019-04-04 19:49:32 +00:00
finished_channel = Channel(String | Nil).new
2019-04-04 19:49:32 +00:00
spawn do
active_threads = 0
active_channel = Channel(Nil).new
channels.each do |ucid|
if active_threads >= max_threads
active_channel.receive
active_threads -= 1
end
2019-04-04 19:49:32 +00:00
active_threads += 1
spawn do
begin
get_channel(ucid, db, refresh, pull_all_videos)
finished_channel.send(ucid)
rescue ex
finished_channel.send(nil)
ensure
active_channel.send(nil)
end
end
end
end
2019-04-04 19:49:32 +00:00
final = [] of String
channels.size.times do
if ucid = finished_channel.receive
2019-04-04 19:49:32 +00:00
final << ucid
end
end
return final
end
2018-12-15 18:05:52 +00:00
def get_channel(id, db, refresh = true, pull_all_videos = true)
2019-06-08 18:31:41 +00:00
if channel = db.query_one?("SELECT * FROM channels WHERE id = $1", id, as: InvidiousChannel)
2019-06-08 00:56:41 +00:00
if refresh && Time.utc - channel.updated > 10.minutes
channel = fetch_channel(id, db, pull_all_videos: pull_all_videos)
2018-08-04 20:30:44 +00:00
channel_array = channel.to_a
args = arg_array(channel_array)
db.exec("INSERT INTO channels VALUES (#{args}) \
2018-12-15 18:02:57 +00:00
ON CONFLICT (id) DO UPDATE SET author = $2, updated = $3", channel_array)
2018-08-04 20:30:44 +00:00
end
else
channel = fetch_channel(id, db, pull_all_videos: pull_all_videos)
2018-10-30 14:20:51 +00:00
channel_array = channel.to_a
args = arg_array(channel_array)
db.exec("INSERT INTO channels VALUES (#{args})", channel_array)
2018-08-04 20:30:44 +00:00
end
return channel
end
def fetch_channel(ucid, db, pull_all_videos = true, locale = nil)
client = make_client(YT_URL)
2018-08-04 20:30:44 +00:00
rss = client.get("/feeds/videos.xml?channel_id=#{ucid}").body
rss = XML.parse_html(rss)
author = rss.xpath_node(%q(//feed/title))
if !author
2018-12-20 21:32:09 +00:00
raise translate(locale, "Deleted or invalid channel")
2018-08-04 20:30:44 +00:00
end
author = author.content
2018-09-17 02:44:24 +00:00
# Auto-generated channels
# https://support.google.com/youtube/answer/2579942
if author.ends_with?(" - Topic") ||
{"Popular on YouTube", "Music", "Sports", "Gaming"}.includes? author
auto_generated = true
end
page = 1
2018-10-30 14:20:51 +00:00
url = produce_channel_videos_url(ucid, page, auto_generated: auto_generated)
response = client.get(url)
json = JSON.parse(response.body)
2018-10-30 14:20:51 +00:00
if json["content_html"]? && !json["content_html"].as_s.empty?
document = XML.parse_html(json["content_html"].as_s)
nodeset = document.xpath_nodes(%q(//li[contains(@class, "feed-item-container")]))
if auto_generated
videos = extract_videos(nodeset)
else
videos = extract_videos(nodeset, ucid, author)
2018-10-30 14:20:51 +00:00
end
end
2018-10-30 14:20:51 +00:00
videos ||= [] of ChannelVideo
2018-10-30 14:20:51 +00:00
rss.xpath_nodes("//feed/entry").each do |entry|
video_id = entry.xpath_node("videoid").not_nil!.content
title = entry.xpath_node("title").not_nil!.content
published = Time.parse_rfc3339(entry.xpath_node("published").not_nil!.content)
updated = Time.parse_rfc3339(entry.xpath_node("updated").not_nil!.content)
author = entry.xpath_node("author/name").not_nil!.content
ucid = entry.xpath_node("channelid").not_nil!.content
2019-05-30 20:09:39 +00:00
views = entry.xpath_node("group/community/statistics").try &.["views"]?.try &.to_i64?
views ||= 0_i64
2018-08-04 20:30:44 +00:00
channel_video = videos.select { |video| video.id == video_id }[0]?
length_seconds = channel_video.try &.length_seconds
length_seconds ||= 0
2018-10-30 14:20:51 +00:00
live_now = channel_video.try &.live_now
live_now ||= false
2019-03-22 15:32:42 +00:00
premiere_timestamp = channel_video.try &.premiere_timestamp
video = ChannelVideo.new(
id: video_id,
title: title,
published: published,
2019-06-08 00:56:41 +00:00
updated: Time.utc,
ucid: ucid,
author: author,
length_seconds: length_seconds,
live_now: live_now,
2019-05-30 20:09:39 +00:00
premiere_timestamp: premiere_timestamp,
views: views,
)
2019-04-10 22:58:42 +00:00
emails = db.query_all("UPDATE users SET notifications = notifications || $1 \
2019-05-26 16:28:54 +00:00
WHERE updated < $2 AND $3 = ANY(subscriptions) AND $1 <> ALL(notifications) RETURNING email",
video.id, video.published, ucid, as: String)
2018-08-04 20:30:44 +00:00
video_array = video.to_a
args = arg_array(video_array)
2018-10-30 14:20:51 +00:00
2019-05-26 16:28:54 +00:00
# We don't include the 'premiere_timestamp' here because channel pages don't include them,
# meaning the above timestamp is always null
db.exec("INSERT INTO channel_videos VALUES (#{args}) \
2018-10-30 15:03:03 +00:00
ON CONFLICT (id) DO UPDATE SET title = $2, published = $3, \
updated = $4, ucid = $5, author = $6, length_seconds = $7, \
2019-05-30 20:09:39 +00:00
live_now = $8, views = $10", video_array)
2019-05-26 16:28:54 +00:00
# Update all users affected by insert
if emails.empty?
values = "'{}'"
else
values = "VALUES #{emails.map { |id| %(('#{id}')) }.join(",")}"
2019-05-26 16:28:54 +00:00
end
2019-06-01 16:19:01 +00:00
db.exec("UPDATE users SET feed_needs_update = true WHERE email = ANY(#{values})")
end
if pull_all_videos
page += 1
2018-09-17 01:32:39 +00:00
ids = [] of String
2018-08-04 20:30:44 +00:00
loop do
2018-09-17 02:44:24 +00:00
url = produce_channel_videos_url(ucid, page, auto_generated: auto_generated)
response = client.get(url)
json = JSON.parse(response.body)
if json["content_html"]? && !json["content_html"].as_s.empty?
document = XML.parse_html(json["content_html"].as_s)
nodeset = document.xpath_nodes(%q(//li[contains(@class, "feed-item-container")]))
else
break
end
nodeset = nodeset.not_nil!
2018-09-17 02:44:24 +00:00
if auto_generated
videos = extract_videos(nodeset)
else
videos = extract_videos(nodeset, ucid, author)
2018-09-17 02:44:24 +00:00
end
count = nodeset.size
videos = videos.map { |video| ChannelVideo.new(
2019-04-10 22:58:42 +00:00
id: video.id,
title: video.title,
published: video.published,
2019-06-08 00:56:41 +00:00
updated: Time.utc,
2019-04-10 22:58:42 +00:00
ucid: video.ucid,
author: video.author,
length_seconds: video.length_seconds,
live_now: video.live_now,
2019-05-30 20:09:39 +00:00
premiere_timestamp: video.premiere_timestamp,
views: video.views
) }
2018-09-17 01:32:39 +00:00
videos.each do |video|
ids << video.id
2019-05-26 16:28:54 +00:00
# We are notified of Red videos elsewhere (PubSub), which includes a correct published date,
# so since they don't provide a published date here we can safely ignore them.
2019-06-08 00:56:41 +00:00
if Time.utc - video.published > 1.minute
emails = db.query_all("UPDATE users SET notifications = notifications || $1 \
2019-05-26 16:28:54 +00:00
WHERE updated < $2 AND $3 = ANY(subscriptions) AND $1 <> ALL(notifications) RETURNING email",
video.id, video.published, video.ucid, as: String)
2018-09-17 01:32:39 +00:00
2018-10-30 15:03:03 +00:00
video_array = video.to_a
args = arg_array(video_array)
2018-10-30 14:20:51 +00:00
# We don't update the 'premire_timestamp' here because channel pages don't include them
db.exec("INSERT INTO channel_videos VALUES (#{args}) \
2019-05-30 20:09:39 +00:00
ON CONFLICT (id) DO UPDATE SET title = $2, published = $3, \
updated = $4, ucid = $5, author = $6, length_seconds = $7, \
live_now = $8, views = $10", video_array)
2019-05-26 16:28:54 +00:00
# Update all users affected by insert
if emails.empty?
values = "'{}'"
else
values = "VALUES #{emails.map { |id| %(('#{id}')) }.join(",")}"
2019-05-26 16:28:54 +00:00
end
2019-06-01 16:19:01 +00:00
db.exec("UPDATE users SET feed_needs_update = true WHERE email = ANY(#{values})")
end
2018-08-04 20:30:44 +00:00
end
if count < 25
2018-08-04 20:30:44 +00:00
break
end
page += 1
end
# When a video is deleted from a channel, we find and remove it here
2018-09-17 01:32:39 +00:00
db.exec("DELETE FROM channel_videos * WHERE NOT id = ANY ('{#{ids.map { |id| %("#{id}") }.join(",")}}') AND ucid = $1", ucid)
2018-08-04 20:30:44 +00:00
end
2019-06-08 00:56:41 +00:00
channel = InvidiousChannel.new(ucid, author, Time.utc, false, nil)
2018-08-04 20:30:44 +00:00
return channel
end
2018-09-04 13:52:30 +00:00
def fetch_channel_playlists(ucid, author, auto_generated, continuation, sort_by)
client = make_client(YT_URL)
if continuation
url = produce_channel_playlists_url(ucid, continuation, sort_by, auto_generated)
response = client.get(url)
json = JSON.parse(response.body)
if json["load_more_widget_html"].as_s.empty?
continuation = nil
else
continuation = XML.parse_html(json["load_more_widget_html"].as_s)
continuation = continuation.xpath_node(%q(//button[@data-uix-load-more-href]))
if continuation
continuation = extract_channel_playlists_cursor(continuation["data-uix-load-more-href"], auto_generated)
end
end
html = XML.parse_html(json["content_html"].as_s)
nodeset = html.xpath_nodes(%q(//li[contains(@class, "feed-item-container")]))
else
url = "/channel/#{ucid}/playlists?disable_polymer=1&flow=list"
if auto_generated
url += "&view=50"
else
url += "&view=1"
end
case sort_by
when "last", "last_added"
#
when "oldest", "oldest_created"
url += "&sort=da"
when "newest", "newest_created"
url += "&sort=dd"
end
response = client.get(url)
html = XML.parse_html(response.body)
continuation = html.xpath_node(%q(//button[@data-uix-load-more-href]))
if continuation
continuation = extract_channel_playlists_cursor(continuation["data-uix-load-more-href"], auto_generated)
end
nodeset = html.xpath_nodes(%q(//ul[@id="browse-items-primary"]/li[contains(@class, "feed-item-container")]))
end
if auto_generated
items = extract_shelf_items(nodeset, ucid, author)
else
items = extract_items(nodeset, ucid, author)
end
return items, continuation
end
2018-11-14 01:04:25 +00:00
def produce_channel_videos_url(ucid, page = 1, auto_generated = nil, sort_by = "newest")
2018-09-05 02:04:40 +00:00
if auto_generated
2018-11-04 15:37:12 +00:00
seed = Time.unix(1525757349)
2018-09-04 13:52:30 +00:00
2019-06-08 00:56:41 +00:00
until seed >= Time.utc
2018-09-05 02:04:40 +00:00
seed += 1.month
end
timestamp = seed - (page - 1).months
2018-11-04 15:37:12 +00:00
page = "#{timestamp.to_unix}"
switch = 0x36
2018-09-05 02:04:40 +00:00
else
page = "#{page}"
switch = 0x00
2018-09-05 02:04:40 +00:00
end
2019-07-21 01:18:08 +00:00
data = IO::Memory.new
data.write_byte 0x12
data.write_byte 0x06
data.print "videos"
data.write Bytes[0x30, 0x02]
data.write Bytes[0x38, 0x01]
data.write Bytes[0x60, 0x01]
data.write Bytes[0x6a, 0x00]
data.write Bytes[0xb8, 0x01, 0x00]
data.write Bytes[0x20, switch]
data.write_byte 0x7a
VarInt.to_io(data, page.bytesize)
data.print page
2018-09-04 13:52:30 +00:00
2018-11-14 01:04:25 +00:00
case sort_by
when "newest"
# Empty tags can be omitted
# meta.write(Bytes[0x18,0x00])
2018-11-14 01:04:25 +00:00
when "popular"
2019-07-21 01:18:08 +00:00
data.write Bytes[0x18, 0x01]
2018-11-14 01:04:25 +00:00
when "oldest"
2019-07-21 01:18:08 +00:00
data.write Bytes[0x18, 0x02]
2018-11-14 01:04:25 +00:00
end
2019-07-21 01:18:08 +00:00
data = Base64.urlsafe_encode(data)
cursor = URI.escape(data)
data = IO::Memory.new
2018-09-04 13:52:30 +00:00
2019-07-21 01:18:08 +00:00
data.write_byte 0x12
VarInt.to_io(data, ucid.bytesize)
data.print ucid
2019-07-21 01:18:08 +00:00
data.write_byte 0x1a
VarInt.to_io(data, cursor.bytesize)
data.print cursor
2019-07-21 01:18:08 +00:00
data.rewind
2018-09-04 13:52:30 +00:00
2019-07-21 01:18:08 +00:00
buffer = IO::Memory.new
buffer.write Bytes[0xe2, 0xa9, 0x85, 0xb2, 0x02]
VarInt.to_io(buffer, data.bytesize)
2018-09-04 13:52:30 +00:00
2019-07-21 01:18:08 +00:00
IO.copy data, buffer
2018-09-04 13:52:30 +00:00
2019-07-21 01:18:08 +00:00
continuation = Base64.urlsafe_encode(buffer)
continuation = URI.escape(continuation)
url = "/browse_ajax?continuation=#{continuation}&gl=US&hl=en"
2018-09-04 13:52:30 +00:00
return url
end
2018-09-21 14:40:04 +00:00
2019-02-15 23:28:54 +00:00
def produce_channel_playlists_url(ucid, cursor, sort = "newest", auto_generated = false)
if !auto_generated
cursor = Base64.urlsafe_encode(cursor, false)
end
2019-07-21 01:18:08 +00:00
data = IO::Memory.new
2019-02-15 23:28:54 +00:00
if auto_generated
2019-07-21 01:18:08 +00:00
data.write Bytes[0x08, 0x0a]
2019-02-15 23:28:54 +00:00
end
2019-07-21 01:18:08 +00:00
data.write Bytes[0x12, 0x09]
data.print "playlists"
2019-02-15 23:28:54 +00:00
if auto_generated
2019-07-21 01:18:08 +00:00
data.write Bytes[0x20, 0x32]
2019-02-15 23:28:54 +00:00
else
# TODO: Look at 0x01, 0x00
case sort
when "oldest", "oldest_created"
2019-07-21 01:18:08 +00:00
data.write Bytes[0x18, 0x02]
2019-02-15 23:28:54 +00:00
when "newest", "newest_created"
2019-07-21 01:18:08 +00:00
data.write Bytes[0x18, 0x03]
2019-02-15 23:28:54 +00:00
when "last", "last_added"
2019-07-21 01:18:08 +00:00
data.write Bytes[0x18, 0x04]
2019-02-15 23:28:54 +00:00
end
2019-07-21 01:18:08 +00:00
data.write Bytes[0x20, 0x01]
2019-02-15 23:28:54 +00:00
end
2019-07-21 01:18:08 +00:00
data.write Bytes[0x30, 0x02]
data.write Bytes[0x38, 0x01]
data.write Bytes[0x60, 0x01]
data.write Bytes[0x6a, 0x00]
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
data.write_byte 0x7a
VarInt.to_io(data, cursor.bytesize)
data.print cursor
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
data.write Bytes[0xb8, 0x01, 0x00]
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
data.rewind
data = Base64.urlsafe_encode(data)
continuation = URI.escape(data)
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
data = IO::Memory.new
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
data.write_byte 0x12
VarInt.to_io(data, ucid.bytesize)
data.print ucid
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
data.write_byte 0x1a
VarInt.to_io(data, continuation.bytesize)
data.print continuation
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
data.rewind
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
buffer = IO::Memory.new
buffer.write Bytes[0xe2, 0xa9, 0x85, 0xb2, 0x02]
VarInt.to_io(buffer, data.bytesize)
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
IO.copy data, buffer
continuation = Base64.urlsafe_encode(buffer)
continuation = URI.escape(continuation)
url = "/browse_ajax?continuation=#{continuation}&gl=US&hl=en"
2019-02-15 23:28:54 +00:00
return url
end
def extract_channel_playlists_cursor(url, auto_generated)
2019-07-21 01:18:08 +00:00
continuation = HTTP::Params.parse(URI.parse(url).query.not_nil!)["continuation"]
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
continuation = URI.unescape(continuation)
data = IO::Memory.new(Base64.decode(continuation))
2019-02-15 23:28:54 +00:00
# 0xe2 0xa9 0x85 0xb2 0x02
2019-07-21 01:18:08 +00:00
data.pos += 5
continuation = Bytes.new(data.read_bytes(VarInt))
data.read continuation
data = IO::Memory.new(continuation)
data.read_byte # => 0x12
ucid = Bytes.new(data.read_bytes(VarInt))
data.read ucid
data.read_byte # => 0x1a
inner_continuation = Bytes.new(data.read_bytes(VarInt))
data.read inner_continuation
continuation = String.new(inner_continuation)
continuation = URI.unescape(continuation)
data = IO::Memory.new(Base64.decode(continuation))
2019-02-15 23:28:54 +00:00
# 0x12 0x09 playlists
2019-07-21 01:18:08 +00:00
data.pos += 11
2019-02-15 23:28:54 +00:00
2019-07-21 01:18:08 +00:00
until data.peek[0] == 0x7a
key = data.read_bytes(VarInt)
value = data.read_bytes(VarInt)
2019-02-15 23:28:54 +00:00
end
2019-07-21 01:18:08 +00:00
data.pos += 1 # => 0x7a
cursor = Bytes.new(data.read_bytes(VarInt))
data.read cursor
2019-02-15 23:28:54 +00:00
cursor = String.new(cursor)
if !auto_generated
cursor = URI.unescape(cursor)
cursor = Base64.decode_string(cursor)
end
return cursor
end
2019-07-02 23:53:19 +00:00
# TODO: Add "sort_by"
2019-07-09 14:31:04 +00:00
def fetch_channel_community(ucid, continuation, locale, config, kemal_config, format, thin_mode)
2019-07-02 23:53:19 +00:00
client = make_client(YT_URL)
headers = HTTP::Headers.new
headers["User-Agent"] = "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/68.0.3440.106 Safari/537.36"
response = client.get("/channel/#{ucid}/community?gl=US&hl=en", headers)
if response.status_code == 404
response = client.get("/user/#{ucid}/community?gl=US&hl=en", headers)
end
if response.status_code == 404
error_message = translate(locale, "This channel does not exist.")
raise error_message
end
2019-07-09 14:31:04 +00:00
ucid = response.body.match(/https:\/\/www.youtube.com\/channel\/(?<ucid>UC[a-zA-Z0-9_-]{22})/).not_nil!["ucid"]
2019-07-02 23:53:19 +00:00
if !continuation || continuation.empty?
2019-07-11 12:27:42 +00:00
initial_data = extract_initial_data(response.body)
body = initial_data["contents"]?.try &.["twoColumnBrowseResultsRenderer"]["tabs"].as_a.select { |tab| tab["tabRenderer"]?.try &.["selected"].as_bool.== true }[0]?
2019-07-02 23:53:19 +00:00
if !body
raise "Could not extract community tab."
end
body = body["tabRenderer"]["content"]["sectionListRenderer"]["contents"][0]["itemSectionRenderer"]
else
2019-07-09 14:31:04 +00:00
continuation = produce_channel_community_continuation(ucid, continuation)
2019-07-02 23:53:19 +00:00
headers["cookie"] = response.cookies.add_request_headers(headers)["cookie"]
headers["content-type"] = "application/x-www-form-urlencoded"
headers["x-client-data"] = "CIi2yQEIpbbJAQipncoBCNedygEIqKPKAQ=="
headers["x-spf-previous"] = ""
headers["x-spf-referer"] = ""
headers["x-youtube-client-name"] = "1"
headers["x-youtube-client-version"] = "2.20180719"
session_token = response.body.match(/"XSRF_TOKEN":"(?<session_token>[A-Za-z0-9\_\-\=]+)"/).try &.["session_token"]? || ""
post_req = {
session_token: session_token,
}
response = client.post("/comment_service_ajax?action_get_comments=1&ctoken=#{continuation}&continuation=#{continuation}&hl=en&gl=US", headers, form: post_req)
body = JSON.parse(response.body)
body = body["response"]["continuationContents"]["itemSectionContinuation"]? ||
body["response"]["continuationContents"]["backstageCommentsContinuation"]?
if !body
raise "Could not extract continuation."
end
end
continuation = body["continuations"]?.try &.[0]["nextContinuationData"]["continuation"].as_s
posts = body["contents"].as_a
if message = posts[0]["messageRenderer"]?
error_message = (message["text"]["simpleText"]? ||
message["text"]["runs"]?.try &.[0]?.try &.["text"]?)
.try &.as_s || ""
raise error_message
end
2019-07-09 14:31:04 +00:00
response = JSON.build do |json|
2019-07-02 23:53:19 +00:00
json.object do
json.field "authorId", ucid
json.field "comments" do
json.array do
posts.each do |post|
comments = post["backstagePostThreadRenderer"]?.try &.["comments"]? ||
post["backstageCommentsContinuation"]?
post = post["backstagePostThreadRenderer"]?.try &.["post"]["backstagePostRenderer"]? ||
post["commentThreadRenderer"]?.try &.["comment"]["commentRenderer"]?
if !post
next
end
if !post["contentText"]?
content_html = ""
else
content_html = post["contentText"]["simpleText"]?.try &.as_s.rchop('\ufeff').try { |block| HTML.escape(block) }.to_s ||
content_to_comment_html(post["contentText"]["runs"].as_a).try &.to_s || ""
end
author = post["authorText"]?.try &.["simpleText"]? || ""
json.object do
json.field "author", author
json.field "authorThumbnails" do
json.array do
qualities = {32, 48, 76, 100, 176, 512}
author_thumbnail = post["authorThumbnail"]["thumbnails"].as_a[0]["url"].as_s
qualities.each do |quality|
json.object do
json.field "url", author_thumbnail.gsub(/s\d+-/, "s#{quality}-")
2019-07-02 23:53:19 +00:00
json.field "width", quality
json.field "height", quality
end
end
end
end
if post["authorEndpoint"]?
json.field "authorId", post["authorEndpoint"]["browseEndpoint"]["browseId"]
json.field "authorUrl", post["authorEndpoint"]["commandMetadata"]["webCommandMetadata"]["url"].as_s
else
json.field "authorId", ""
json.field "authorUrl", ""
end
published_text = post["publishedTimeText"]["runs"][0]["text"].as_s
published = decode_date(published_text.rchop(" (edited)"))
if published_text.includes?(" (edited)")
json.field "isEdited", true
else
json.field "isEdited", false
end
like_count = post["actionButtons"]["commentActionButtonsRenderer"]["likeButton"]["toggleButtonRenderer"]["accessibilityData"]["accessibilityData"]["label"]
.try &.as_s.gsub(/\D/, "").to_i? || 0
json.field "content", html_to_content(content_html)
json.field "contentHtml", content_html
json.field "published", published.to_unix
json.field "publishedText", translate(locale, "`x` ago", recode_date(published, locale))
json.field "likeCount", like_count
json.field "commentId", post["postId"]? || post["commentId"]? || ""
2019-07-09 14:31:04 +00:00
json.field "authorIsChannelOwner", post["authorEndpoint"]["browseEndpoint"]["browseId"] == ucid
2019-07-02 23:53:19 +00:00
if attachment = post["backstageAttachment"]?
json.field "attachment" do
json.object do
case attachment.as_h
when .has_key?("videoRenderer")
attachment = attachment["videoRenderer"]
json.field "type", "video"
if !attachment["videoId"]?
error_message = (attachment["title"]["simpleText"]? ||
attachment["title"]["runs"]?.try &.[0]?.try &.["text"]?)
json.field "error", error_message
else
video_id = attachment["videoId"].as_s
json.field "title", attachment["title"]["simpleText"].as_s
json.field "videoId", video_id
json.field "videoThumbnails" do
generate_thumbnails(json, video_id, config, kemal_config)
end
json.field "lengthSeconds", decode_length_seconds(attachment["lengthText"]["simpleText"].as_s)
author_info = attachment["ownerText"]["runs"][0].as_h
json.field "author", author_info["text"].as_s
json.field "authorId", author_info["navigationEndpoint"]["browseEndpoint"]["browseId"]
json.field "authorUrl", author_info["navigationEndpoint"]["commandMetadata"]["webCommandMetadata"]["url"]
# TODO: json.field "authorThumbnails", "channelThumbnailSupportedRenderers"
# TODO: json.field "authorVerified", "ownerBadges"
published = decode_date(attachment["publishedTimeText"]["simpleText"].as_s)
json.field "published", published.to_unix
json.field "publishedText", translate(locale, "`x` ago", recode_date(published, locale))
view_count = attachment["viewCountText"]["simpleText"].as_s.gsub(/\D/, "").to_i64? || 0_i64
json.field "viewCount", view_count
json.field "viewCountText", translate(locale, "`x` views", number_to_short_text(view_count))
end
when .has_key?("backstageImageRenderer")
attachment = attachment["backstageImageRenderer"]
json.field "type", "image"
json.field "imageThumbnails" do
json.array do
thumbnail = attachment["image"]["thumbnails"][0].as_h
width = thumbnail["width"].as_i
height = thumbnail["height"].as_i
aspect_ratio = (width.to_f / height.to_f)
2019-08-01 00:16:09 +00:00
url = thumbnail["url"].as_s.gsub(/=w\d+-h\d+(-p)?(-nd)?(-df)?(-rwa)?/, "=s640")
2019-07-02 23:53:19 +00:00
qualities = {320, 560, 640, 1280, 2000}
qualities.each do |quality|
json.object do
2019-08-01 00:16:09 +00:00
json.field "url", url.gsub(/=s\d+/, "=s#{quality}")
2019-07-02 23:53:19 +00:00
json.field "width", quality
json.field "height", (quality / aspect_ratio).ceil.to_i
end
end
end
end
# TODO
# when .has_key?("pollRenderer")
# attachment = attachment["pollRenderer"]
# json.field "type", "poll"
else
json.field "type", "unknown"
json.field "error", "Unrecognized attachment type."
2019-07-02 23:53:19 +00:00
end
end
end
end
if comments && (reply_count = (comments["backstageCommentsRenderer"]["moreText"]["simpleText"]? ||
comments["backstageCommentsRenderer"]["moreText"]["runs"]?.try &.[0]?.try &.["text"]?)
.try &.as_s.gsub(/\D/, "").to_i?)
continuation = comments["backstageCommentsRenderer"]["continuations"]?.try &.as_a[0]["nextContinuationData"]["continuation"].as_s
continuation ||= ""
json.field "replies" do
json.object do
json.field "replyCount", reply_count
2019-07-09 14:31:04 +00:00
json.field "continuation", extract_channel_community_cursor(continuation)
2019-07-02 23:53:19 +00:00
end
end
end
end
end
end
end
if body["continuations"]?
2019-07-09 14:31:04 +00:00
continuation = body["continuations"][0]["nextContinuationData"]["continuation"].as_s
json.field "continuation", extract_channel_community_cursor(continuation)
end
end
end
if format == "html"
response = JSON.parse(response)
content_html = template_youtube_comments(response, locale, thin_mode)
response = JSON.build do |json|
json.object do
json.field "contentHtml", content_html
2019-07-02 23:53:19 +00:00
end
end
end
2019-07-09 14:31:04 +00:00
return response
end
def produce_channel_community_continuation(ucid, cursor)
cursor = URI.escape(cursor)
2019-07-21 01:18:08 +00:00
data = IO::Memory.new
data.write_byte 0x12
VarInt.to_io(data, ucid.bytesize)
data.print ucid
2019-07-09 14:31:04 +00:00
2019-07-21 01:18:08 +00:00
data.write_byte 0x1a
VarInt.to_io(data, cursor.bytesize)
data.print cursor
2019-07-09 14:31:04 +00:00
2019-07-21 01:18:08 +00:00
data.rewind
2019-07-09 14:31:04 +00:00
2019-07-21 01:18:08 +00:00
buffer = IO::Memory.new
buffer.write Bytes[0xe2, 0xa9, 0x85, 0xb2, 0x02]
VarInt.to_io(buffer, data.size)
IO.copy data, buffer
continuation = Base64.urlsafe_encode(buffer)
2019-07-09 14:31:04 +00:00
continuation = URI.escape(continuation)
return continuation
end
def extract_channel_community_cursor(continuation)
continuation = URI.unescape(continuation)
2019-07-21 01:18:08 +00:00
data = IO::Memory.new(Base64.decode(continuation))
2019-07-09 14:31:04 +00:00
# 0xe2 0xa9 0x85 0xb2 0x02
2019-07-21 01:18:08 +00:00
data.pos += 5
continuation = Bytes.new(data.read_bytes(VarInt))
data.read continuation
data = IO::Memory.new(continuation)
data.read_byte # => 0x12
ucid = Bytes.new(data.read_bytes(VarInt))
data.read ucid
data.read_byte # => 0x1a
until data.peek[0] == 'E'.ord
data.read_byte
2019-07-09 14:31:04 +00:00
end
2019-07-21 01:18:08 +00:00
return URI.unescape(data.gets_to_end)
2019-07-02 23:53:19 +00:00
end
2018-12-20 21:32:09 +00:00
def get_about_info(ucid, locale)
2018-09-21 14:40:04 +00:00
client = make_client(YT_URL)
about = client.get("/channel/#{ucid}/about?disable_polymer=1&gl=US&hl=en")
2018-10-24 02:04:15 +00:00
if about.status_code == 404
about = client.get("/user/#{ucid}/about?disable_polymer=1&gl=US&hl=en")
end
2018-09-21 14:40:04 +00:00
about = XML.parse_html(about.body)
2018-10-24 02:04:15 +00:00
if about.xpath_node(%q(//div[contains(@class, "channel-empty-message")]))
2018-12-20 21:32:09 +00:00
error_message = translate(locale, "This channel does not exist.")
2018-10-24 02:04:15 +00:00
raise error_message
2018-09-21 14:40:04 +00:00
end
if about.xpath_node(%q(//span[contains(@class,"qualified-channel-title-text")]/a)).try &.content.empty?
error_message = about.xpath_node(%q(//div[@class="yt-alert-content"])).try &.content.strip
2018-12-20 21:32:09 +00:00
error_message ||= translate(locale, "Could not get channel info.")
raise error_message
2018-09-21 14:40:04 +00:00
end
sub_count = about.xpath_node(%q(//span[contains(text(), "subscribers")]))
if sub_count
sub_count = sub_count.content.delete(", subscribers").to_i?
end
sub_count ||= 0
2018-10-22 02:44:20 +00:00
author = about.xpath_node(%q(//span[contains(@class,"qualified-channel-title-text")]/a)).not_nil!.content
author_url = about.xpath_node(%q(//span[contains(@class,"qualified-channel-title-text")]/a)).not_nil!["href"]
author_thumbnail = about.xpath_node(%q(//img[@class="channel-header-profile-image"])).not_nil!["src"]
2019-02-15 23:28:54 +00:00
ucid = about.xpath_node(%q(//meta[@itemprop="channelId"])).not_nil!["content"]
2018-09-21 14:40:04 +00:00
banner = about.xpath_node(%q(//div[@id="gh-banner"]/style)).not_nil!.content
banner = "https:" + banner.match(/background-image: url\((?<url>[^)]+)\)/).not_nil!["url"]
if banner.includes? "channels/c4/default_banner"
banner = nil
end
description_html = about.xpath_node(%q(//div[contains(@class,"about-description")])).try &.to_s ||
%(<div class="about-description branded-page-box-padding"><pre></pre></div>)
paid = about.xpath_node(%q(//meta[@itemprop="paid"])).not_nil!["content"] == "True"
is_family_friendly = about.xpath_node(%q(//meta[@itemprop="isFamilyFriendly"])).not_nil!["content"] == "True"
allowed_regions = about.xpath_node(%q(//meta[@itemprop="regionsAllowed"])).not_nil!["content"].split(",")
related_channels = about.xpath_nodes(%q(//div[contains(@class, "branded-page-related-channels")]/ul/li))
related_channels = related_channels.map do |node|
related_id = node["data-external-id"]?
related_id ||= ""
anchor = node.xpath_node(%q(.//h3[contains(@class, "yt-lockup-title")]/a))
related_title = anchor.try &.["title"]
related_title ||= ""
related_author_url = anchor.try &.["href"]
related_author_url ||= ""
related_author_thumbnail = node.xpath_node(%q(.//img)).try &.["data-thumb"]
related_author_thumbnail ||= ""
AboutRelatedChannel.new(
ucid: related_id,
author: related_title,
author_url: related_author_url,
author_thumbnail: related_author_thumbnail,
)
end
total_views = 0_i64
sub_count = 0_i64
joined = Time.unix(0)
metadata = about.xpath_nodes(%q(//span[@class="about-stat"]))
metadata.each do |item|
case item.content
when .includes? "views"
total_views = item.content.gsub(/\D/, "").to_i64
when .includes? "subscribers"
sub_count = item.content.delete("subscribers").gsub(/\D/, "").to_i64
when .includes? "Joined"
joined = Time.parse(item.content.lchop("Joined "), "%b %-d, %Y", Time::Location.local)
end
end
2018-09-21 14:40:04 +00:00
# Auto-generated channels
# https://support.google.com/youtube/answer/2579942
auto_generated = false
if about.xpath_node(%q(//ul[@class="about-custom-links"]/li/a[@title="Auto-generated by YouTube"])) ||
about.xpath_node(%q(//span[@class="qualified-channel-title-badge"]/span[@title="Auto-generated by YouTube"]))
auto_generated = true
end
2019-07-09 14:31:04 +00:00
tabs = about.xpath_nodes(%q(//ul[@id="channel-navigation-menu"]/li/a/span)).map { |node| node.content.downcase }
return AboutChannel.new(
ucid: ucid,
author: author,
auto_generated: auto_generated,
author_url: author_url,
author_thumbnail: author_thumbnail,
banner: banner,
description_html: description_html,
paid: paid,
total_views: total_views,
sub_count: sub_count,
joined: joined,
is_family_friendly: is_family_friendly,
allowed_regions: allowed_regions,
2019-07-09 14:31:04 +00:00
related_channels: related_channels,
tabs: tabs
)
2018-09-21 14:40:04 +00:00
end
def get_60_videos(ucid, author, page, auto_generated, sort_by = "newest")
count = 0
videos = [] of SearchVideo
client = make_client(YT_URL)
2.times do |i|
2018-11-14 01:04:25 +00:00
url = produce_channel_videos_url(ucid, page * 2 + (i - 1), auto_generated: auto_generated, sort_by: sort_by)
response = client.get(url)
json = JSON.parse(response.body)
if json["content_html"]? && !json["content_html"].as_s.empty?
document = XML.parse_html(json["content_html"].as_s)
nodeset = document.xpath_nodes(%q(//li[contains(@class, "feed-item-container")]))
if !json["load_more_widget_html"]?.try &.as_s.empty?
count += 30
end
if auto_generated
videos += extract_videos(nodeset)
else
videos += extract_videos(nodeset, ucid, author)
end
else
break
end
end
return videos, count
end
2019-02-19 23:00:06 +00:00
def get_latest_videos(ucid)
client = make_client(YT_URL)
videos = [] of SearchVideo
url = produce_channel_videos_url(ucid, 0)
response = client.get(url)
json = JSON.parse(response.body)
if json["content_html"]? && !json["content_html"].as_s.empty?
document = XML.parse_html(json["content_html"].as_s)
nodeset = document.xpath_nodes(%q(//li[contains(@class, "feed-item-container")]))
videos = extract_videos(nodeset, ucid)
end
return videos
end