invidious-copy-2023-06-08/src/invidious/helpers/helpers.cr

377 lines
11 KiB
Crystal
Raw Normal View History

2018-08-04 20:30:44 +00:00
class Config
YAML.mapping({
crawl_threads: Int32,
channel_threads: Int32,
video_threads: Int32,
db: NamedTuple(
user: String,
password: String,
host: String,
port: Int32,
dbname: String,
),
2018-08-09 01:12:17 +00:00
dl_api_key: String?,
https_only: Bool?,
hmac_key: String?,
full_refresh: Bool,
2018-10-03 15:36:30 +00:00
geo_bypass: Bool,
2018-08-04 20:30:44 +00:00
})
end
class FilteredCompressHandler < Kemal::Handler
2018-09-17 23:39:28 +00:00
exclude ["/videoplayback", "/videoplayback/*", "/vi/*", "/api/*", "/ggpht/*"]
2018-08-04 20:30:44 +00:00
def call(env)
return call_next env if exclude_match? env
{% if flag?(:without_zlib) %}
call_next env
{% else %}
request_headers = env.request.headers
if request_headers.includes_word?("Accept-Encoding", "gzip")
env.response.headers["Content-Encoding"] = "gzip"
env.response.output = Gzip::Writer.new(env.response.output, sync_close: true)
elsif request_headers.includes_word?("Accept-Encoding", "deflate")
env.response.headers["Content-Encoding"] = "deflate"
env.response.output = Flate::Writer.new(env.response.output, sync_close: true)
end
call_next env
{% end %}
end
end
class DenyFrame < Kemal::Handler
exclude ["/embed/*"]
def call(env)
return call_next env if exclude_match? env
env.response.headers["X-Frame-Options"] = "sameorigin"
call_next env
end
end
2018-08-04 20:30:44 +00:00
def rank_videos(db, n, filter, url)
top = [] of {Float64, String}
db.query("SELECT id, wilson_score, published FROM videos WHERE views > 5000 ORDER BY published DESC LIMIT 1000") do |rs|
rs.each do
id = rs.read(String)
wilson_score = rs.read(Float64)
published = rs.read(Time)
# Exponential decay, older videos tend to rank lower
temperature = wilson_score * Math.exp(-0.000005*((Time.now - published).total_minutes))
top << {temperature, id}
end
end
top.sort!
# Make hottest come first
top.reverse!
top = top.map { |a, b| b }
if filter
language_list = [] of String
top.each do |id|
if language_list.size == n
break
else
client = make_client(url)
begin
video = get_video(id, db)
rescue ex
next
end
if video.language
language = video.language
else
description = XML.parse(video.description)
content = [video.title, description.content].join(" ")
content = content[0, 10000]
results = DetectLanguage.detect(content)
language = results[0].language
db.exec("UPDATE videos SET language = $1 WHERE id = $2", language, id)
end
if language == "en"
language_list << id
end
end
end
return language_list
else
return top[0..n - 1]
end
end
def login_req(login_form, f_req)
data = {
"pstMsg" => "1",
"checkConnection" => "youtube",
"checkedDomains" => "youtube",
"hl" => "en",
"deviceinfo" => %q([null,null,null,[],null,"US",null,null,[],"GlifWebSignIn",null,[null,null,[]]]),
"f.req" => f_req,
"flowName" => "GlifWebSignIn",
"flowEntry" => "ServiceLogin",
}
data = login_form.merge(data)
return HTTP::Params.encode(data)
end
def generate_captcha(key)
minute = Random::Secure.rand(12)
minute_angle = minute * 30
minute = minute * 5
hour = Random::Secure.rand(12)
hour_angle = hour * 30 + minute_angle.to_f / 12
if hour == 0
hour = 12
end
clock_svg = <<-END_SVG
<svg viewBox="0 0 100 100" width="200px">
<circle cx="50" cy="50" r="45" fill="#eee" stroke="black" stroke-width="2"></circle>
<text x="69" y="20.091" text-anchor="middle" fill="black" font-family="Arial" font-size="10px"> 1</text>
<text x="82.909" y="34" text-anchor="middle" fill="black" font-family="Arial" font-size="10px"> 2</text>
<text x="88" y="53" text-anchor="middle" fill="black" font-family="Arial" font-size="10px"> 3</text>
<text x="82.909" y="72" text-anchor="middle" fill="black" font-family="Arial" font-size="10px"> 4</text>
<text x="69" y="85.909" text-anchor="middle" fill="black" font-family="Arial" font-size="10px"> 5</text>
<text x="50" y="91" text-anchor="middle" fill="black" font-family="Arial" font-size="10px"> 6</text>
<text x="31" y="85.909" text-anchor="middle" fill="black" font-family="Arial" font-size="10px"> 7</text>
<text x="17.091" y="72" text-anchor="middle" fill="black" font-family="Arial" font-size="10px"> 8</text>
<text x="12" y="53" text-anchor="middle" fill="black" font-family="Arial" font-size="10px"> 9</text>
<text x="17.091" y="34" text-anchor="middle" fill="black" font-family="Arial" font-size="10px">10</text>
<text x="31" y="20.091" text-anchor="middle" fill="black" font-family="Arial" font-size="10px">11</text>
<text x="50" y="15" text-anchor="middle" fill="black" font-family="Arial" font-size="10px">12</text>
<circle cx="50" cy="50" r="3" fill="black"></circle>
<line id="minute" transform="rotate(#{minute_angle}, 50, 50)" x1="50" y1="50" x2="50" y2="16" fill="black" stroke="black" stroke-width="2"></line>
<line id="hour" transform="rotate(#{hour_angle}, 50, 50)" x1="50" y1="50" x2="50" y2="24" fill="black" stroke="black" stroke-width="2"></line>
</svg>
END_SVG
challenge = ""
convert = Process.run(%(convert -density 1200 -resize 400x400 -background none svg:- png:-), shell: true,
input: IO::Memory.new(clock_svg), output: Process::Redirect::Pipe) do |proc|
challenge = proc.output.gets_to_end
challenge = Base64.strict_encode(challenge)
challenge = "data:image/png;base64,#{challenge}"
end
answer = "#{hour}:#{minute.to_s.rjust(2, '0')}"
token = OpenSSL::HMAC.digest(:sha256, key, answer)
2018-08-15 17:40:42 +00:00
token = Base64.urlsafe_encode(token)
2018-08-04 20:30:44 +00:00
return {challenge: challenge, token: token}
end
2018-08-10 13:38:31 +00:00
2018-09-04 13:52:30 +00:00
def html_to_content(description_html)
2018-08-10 13:38:31 +00:00
if !description_html
description = ""
description_html = ""
else
description_html = description_html.to_s
description = description_html.gsub("<br>", "\n")
description = description.gsub("<br/>", "\n")
2018-09-06 21:50:12 +00:00
if description.empty?
description = ""
else
description = XML.parse_html(description).content.strip("\n ")
end
2018-08-10 13:38:31 +00:00
end
2018-09-04 13:52:30 +00:00
return description_html, description
2018-08-10 13:38:31 +00:00
end
def extract_videos(nodeset, ucid = nil)
videos = extract_items(nodeset, ucid)
videos.select! { |item| !item.is_a?(SearchChannel | SearchPlaylist) }
videos.map { |video| video.as(SearchVideo) }
end
def extract_items(nodeset, ucid = nil)
# TODO: Make this a 'common', so it makes more sense to be used here
items = [] of SearchItem
nodeset.each do |node|
anchor = node.xpath_node(%q(.//h3[contains(@class,"yt-lockup-title")]/a))
if !anchor
next
end
if anchor["href"].starts_with? "https://www.googleadservices.com"
next
end
anchor = node.xpath_node(%q(.//div[contains(@class, "yt-lockup-byline")]/a))
if !anchor
author = ""
author_id = ""
else
2018-09-22 15:49:42 +00:00
author = anchor.content.strip
author_id = anchor["href"].split("/")[-1]
end
anchor = node.xpath_node(%q(.//h3[contains(@class, "yt-lockup-title")]/a))
if !anchor
2018-08-16 14:05:48 +00:00
next
2018-08-21 00:25:12 +00:00
end
title = anchor.content.strip
id = anchor["href"]
description_html = node.xpath_node(%q(.//div[contains(@class, "yt-lockup-description")]))
description_html, description = html_to_content(description_html)
2018-09-19 20:24:19 +00:00
tile = node.xpath_node(%q(.//div[contains(@class, "yt-lockup-tile")]))
if !tile
next
end
case tile["class"]
when .includes? "yt-lockup-playlist"
plid = HTTP::Params.parse(URI.parse(id).query.not_nil!)["list"]
2018-08-21 00:25:12 +00:00
anchor = node.xpath_node(%q(.//div[contains(@class, "yt-lockup-meta")]/a))
2018-09-29 04:12:35 +00:00
if !anchor
anchor = node.xpath_node(%q(.//ul[@class="yt-lockup-meta-info"]/li/a))
end
2018-09-29 04:12:35 +00:00
video_count = node.xpath_node(%q(.//span[@class="formatted-video-count-label"]/b))
if video_count
video_count = video_count.content
if video_count == "50+"
author = "YouTube"
author_id = "UC-9-kyTW8ZkZNDHQJ6FgpwQ"
video_count = video_count.rchop("+")
end
video_count = video_count.to_i?
end
video_count ||= 0
videos = [] of SearchPlaylistVideo
2018-09-22 15:49:42 +00:00
node.xpath_nodes(%q(.//*[contains(@class, "yt-lockup-playlist-items")]/li)).each do |video|
anchor = video.xpath_node(%q(.//a))
if anchor
2018-09-22 15:49:42 +00:00
video_title = anchor.content.strip
id = HTTP::Params.parse(URI.parse(anchor["href"]).query.not_nil!)["v"]
end
video_title ||= ""
id ||= ""
2018-09-19 20:24:19 +00:00
anchor = video.xpath_node(%q(.//span/span))
if anchor
length_seconds = decode_length_seconds(anchor.content)
end
length_seconds ||= 0
videos << SearchPlaylistVideo.new(
video_title,
id,
length_seconds
)
end
items << SearchPlaylist.new(
title,
plid,
author,
author_id,
video_count,
videos
)
when .includes? "yt-lockup-channel"
2018-09-22 15:49:42 +00:00
author = title.strip
ucid = id.split("/")[-1]
author_thumbnail = node.xpath_node(%q(.//div/span/img)).try &.["data-thumb"]?
author_thumbnail ||= node.xpath_node(%q(.//div/span/img)).try &.["src"]
author_thumbnail ||= ""
subscriber_count = node.xpath_node(%q(.//span[contains(@class, "yt-subscriber-count")])).try &.["title"].delete(",").to_i?
subscriber_count ||= 0
video_count = node.xpath_node(%q(.//ul[@class="yt-lockup-meta-info"]/li)).try &.content.split(" ")[0].delete(",").to_i?
video_count ||= 0
items << SearchChannel.new(
author,
ucid,
author_thumbnail,
subscriber_count,
video_count,
description,
description_html
)
else
id = id.lchop("/watch?v=")
metadata = node.xpath_nodes(%q(.//div[contains(@class,"yt-lockup-meta")]/ul/li))
begin
published = decode_date(metadata[0].content.lchop("Streamed ").lchop("Starts "))
rescue ex
end
begin
published ||= Time.epoch(metadata[0].xpath_node(%q(.//span)).not_nil!["data-timestamp"].to_i64)
rescue ex
end
published ||= Time.now
begin
view_count = metadata[0].content.rchop(" watching").delete(",").try &.to_i64?
rescue ex
end
begin
view_count ||= metadata.try &.[1].content.delete("No views,").try &.to_i64?
rescue ex
end
view_count ||= 0_i64
length_seconds = node.xpath_node(%q(.//span[@class="video-time"]))
if length_seconds
length_seconds = decode_length_seconds(length_seconds.content)
else
length_seconds = -1
end
live_now = node.xpath_node(%q(.//span[contains(@class, "yt-badge-live")]))
if live_now
live_now = true
else
live_now = false
end
items << SearchVideo.new(
title,
id,
author,
author_id,
published,
view_count,
description,
description_html,
length_seconds,
live_now
)
end
end
return items
end