invidious/src/invidious.cr

287 lines
7.0 KiB
Crystal
Raw Normal View History

2018-01-28 18:32:40 +01:00
# "Invidious" (which indexes popular video sites)
# Copyright (C) 2018 Omar Roth
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as published
# by the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU Affero General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
2017-11-23 08:48:55 +01:00
require "kemal"
require "option_parser"
require "pg"
2018-01-16 21:02:35 +01:00
require "xml"
2018-01-21 01:19:12 +01:00
require "./helpers"
2017-11-29 22:33:46 +01:00
threads = 10
Kemal.config.extra_options do |parser|
parser.banner = "Usage: invidious [arguments]"
2018-02-06 02:08:36 +01:00
parser.on("-t THREADS", "--threads=THREADS", "Number of threads for crawling (default: 10)") do |number|
begin
threads = number.to_i32
rescue ex
puts "THREADS must be integer"
exit
end
end
end
2018-02-11 23:48:27 +01:00
Kemal::CLI.new
PG_DB = DB.open "postgres://kemal:kemal@localhost:5432/invidious"
2018-01-17 04:42:48 +01:00
URL = URI.parse("https://www.youtube.com")
CONTEXT = OpenSSL::SSL::Context::Client.new
CONTEXT.verify_mode = OpenSSL::SSL::VerifyMode::NONE
CONTEXT.add_options(
OpenSSL::SSL::Options::ALL |
OpenSSL::SSL::Options::NO_SSL_V2 |
OpenSSL::SSL::Options::NO_SSL_V3
)
pool = Deque.new((threads * 1.2 + 1).to_i) do
2018-02-06 02:07:49 +01:00
make_client(URL, CONTEXT)
2018-01-17 04:42:48 +01:00
end
2018-01-28 03:09:27 +01:00
# Refresh pool by crawling YT
threads.times do
2018-01-28 03:09:27 +01:00
spawn do
io = STDOUT
ids = Deque(String).new
random = Random.new
2018-02-05 02:40:06 +01:00
client = get_client(pool)
2018-01-21 18:04:58 +01:00
2018-01-28 03:09:27 +01:00
search(random.base64(3), client) do |id|
ids << id
2018-01-22 00:49:27 +01:00
end
2018-02-09 03:19:44 +01:00
pool << client
2018-01-28 03:09:27 +01:00
loop do
2018-02-09 03:19:44 +01:00
client = get_client(pool)
2018-01-28 03:09:27 +01:00
if ids.empty?
search(random.base64(3), client) do |id|
ids << id
end
end
2018-01-21 18:04:58 +01:00
2018-01-28 03:09:27 +01:00
if rand(300) < 1
2018-02-06 02:07:49 +01:00
pool << make_client(URL, CONTEXT)
2018-02-09 03:19:44 +01:00
client = get_client(pool)
2018-01-28 03:09:27 +01:00
end
2018-01-17 04:42:48 +01:00
2018-01-28 03:09:27 +01:00
begin
id = ids[0]
2018-02-05 02:40:06 +01:00
video = get_video(id, client, PG_DB)
2018-01-28 03:09:27 +01:00
rescue ex
io << id << " : " << ex << "\n"
2018-02-06 02:07:49 +01:00
pool << make_client(URL, CONTEXT)
2018-01-28 03:09:27 +01:00
next
ensure
ids.delete(id)
end
2018-01-21 01:19:55 +01:00
2018-01-28 03:09:27 +01:00
rvs = [] of Hash(String, String)
if video.info.has_key?("rvs")
video.info["rvs"].split(",").each do |rv|
rvs << HTTP::Params.parse(rv).to_h
end
2018-01-19 04:46:29 +01:00
end
2018-01-21 01:19:55 +01:00
2018-01-28 03:09:27 +01:00
rvs.each do |rv|
if rv.has_key?("id") && !PG_DB.query_one?("SELECT EXISTS (SELECT true FROM videos WHERE id = $1)", rv["id"], as: Bool)
ids.delete(id)
2018-01-22 00:49:27 +01:00
ids << rv["id"]
2018-01-28 03:09:27 +01:00
if ids.size == 150
2018-01-22 00:49:27 +01:00
ids.shift
2018-01-21 18:07:32 +01:00
end
end
2018-01-21 18:04:58 +01:00
end
2018-01-21 01:19:55 +01:00
2018-02-09 03:19:44 +01:00
pool << client
2018-01-28 03:09:27 +01:00
end
2018-01-17 04:42:48 +01:00
end
2018-01-08 00:18:24 +01:00
end
2018-02-08 05:04:47 +01:00
top_videos = [] of Video
2017-11-23 08:48:55 +01:00
2018-02-08 05:04:47 +01:00
spawn do
loop do
2018-02-09 03:19:44 +01:00
top = rank_videos(PG_DB, 40)
2018-02-08 05:04:47 +01:00
client = get_client(pool)
2018-02-06 00:56:40 +01:00
2018-02-08 05:04:47 +01:00
args = [] of String
2018-02-12 05:05:22 +01:00
(1..top.size).each { |i| args << "($#{i})," }
2018-02-08 05:04:47 +01:00
args = args.join("")
2018-02-12 05:05:22 +01:00
args = args.chomp(",")
2018-02-06 00:56:40 +01:00
2018-02-09 03:19:44 +01:00
videos = [] of Video
2018-02-08 05:04:47 +01:00
PG_DB.query("SELECT * FROM videos d INNER JOIN (VALUES #{args}) v(id) USING (id)", top) do |rs|
rs.each do
video = rs.read(Video)
2018-02-09 03:19:44 +01:00
videos << video
2018-02-08 05:04:47 +01:00
end
2018-02-06 00:56:40 +01:00
end
2018-02-08 05:04:47 +01:00
2018-02-09 03:19:44 +01:00
top_videos = videos
2018-02-08 05:04:47 +01:00
2018-02-09 03:19:44 +01:00
pool << client
2018-02-08 05:04:47 +01:00
end
end
macro templated(filename)
render "src/views/#{{{filename}}}.ecr", "src/views/layout.ecr"
2018-02-06 00:56:40 +01:00
end
2018-02-08 05:04:47 +01:00
get "/" do |env|
templated "index"
end
get "/watch" do |env|
id = env.params.query["v"]
2018-01-16 03:30:57 +01:00
2018-02-12 00:01:32 +01:00
if env.params.query["listen"]? && env.params.query["listen"] == "true"
listen = true
2018-01-16 03:30:57 +01:00
env.params.query.delete_all("listen")
2018-02-12 00:01:32 +01:00
else
listen = false
end
2018-02-05 02:40:06 +01:00
client = get_client(pool)
begin
2018-01-28 03:09:27 +01:00
video = get_video(id, client, PG_DB)
rescue ex
error_message = ex.message
next templated "error"
end
fmt_stream = [] of HTTP::Params
video.info["url_encoded_fmt_stream_map"].split(",") do |string|
fmt_stream << HTTP::Params.parse(string)
end
2018-01-21 18:07:32 +01:00
if fmt_stream[0]?
if fmt_stream[0]["s"]?
fmt_stream.each do |fmt|
fmt["url"] = "#{fmt["url"]}&signature=#{decrypt_signature(fmt["s"])}"
end
end
end
# We want lowest quality first
fmt_stream.reverse!
2018-01-21 18:07:32 +01:00
2018-01-04 03:06:16 +01:00
adaptive_fmts = [] of HTTP::Params
if video.info.has_key?("adaptive_fmts")
video.info["adaptive_fmts"].split(",") do |string|
adaptive_fmts << HTTP::Params.parse(string)
end
2018-01-04 03:06:16 +01:00
end
2018-01-28 03:09:27 +01:00
if adaptive_fmts[0]?
if adaptive_fmts[0]["s"]?
adaptive_fmts.each do |fmt|
fmt["url"] = "#{fmt["url"]}&signature=#{decrypt_signature(fmt["s"])}"
end
end
end
rvs = [] of Hash(String, String)
if video.info.has_key?("rvs")
video.info["rvs"].split(",").each do |rv|
rvs << HTTP::Params.parse(rv).to_h
end
end
2018-01-21 18:07:32 +01:00
player_response = JSON.parse(video.info["player_response"])
2018-01-21 18:07:32 +01:00
description = video.html.xpath_node(%q(//p[@id="eow-description"]))
2018-01-15 04:16:09 +01:00
description = description ? description.to_xml : "Could not load description"
rating = video.info["avg_rating"].to_f64
2018-01-28 03:09:27 +01:00
engagement = ((video.dislikes.to_f + video.likes.to_f)/video.views * 100)
2018-02-05 02:42:13 +01:00
if video.likes > 0 || video.dislikes > 0
calculated_rating = (video.likes.to_f/(video.likes.to_f + video.dislikes.to_f) * 4 + 1)
2018-02-05 02:42:13 +01:00
else
calculated_rating = 0.0
end
2018-02-06 01:52:21 +01:00
pool << client
2017-11-23 08:48:55 +01:00
templated "watch"
end
2017-12-30 22:21:43 +01:00
get "/search" do |env|
2018-01-07 18:42:24 +01:00
query = env.params.query["q"]
2018-01-15 04:16:09 +01:00
page = env.params.query["page"]? && env.params.query["page"].to_i? ? env.params.query["page"].to_i : 1
2018-01-07 18:42:24 +01:00
2018-02-05 02:40:06 +01:00
client = get_client(pool)
2018-01-08 00:18:24 +01:00
2018-01-16 05:11:51 +01:00
html = client.get("https://www.youtube.com/results?q=#{URI.escape(query)}&page=#{page}&sp=EgIQAVAU").body
2018-01-15 04:16:09 +01:00
html = XML.parse_html(html)
2018-01-15 04:16:09 +01:00
videos = Array(Hash(String, String)).new
2017-12-30 22:21:43 +01:00
2018-01-16 05:11:51 +01:00
html.xpath_nodes(%q(//ol[@class="item-section"]/li)).each do |item|
root = item.xpath_node(%q(div[contains(@class,"yt-lockup-video")]/div))
if root
video = {} of String => String
2017-12-30 22:21:43 +01:00
2018-01-16 05:11:51 +01:00
link = root.xpath_node(%q(div[contains(@class,"yt-lockup-thumbnail")]/a/@href))
2018-01-15 04:16:09 +01:00
if link
video["link"] = link.content
2018-01-16 05:11:51 +01:00
else
video["link"] = "#"
2018-01-15 04:16:09 +01:00
end
2017-12-30 22:21:43 +01:00
2018-01-16 05:11:51 +01:00
title = root.xpath_node(%q(div[@class="yt-lockup-content"]/h3/a))
2018-01-15 04:16:09 +01:00
if title
video["title"] = title.content
2018-01-16 05:11:51 +01:00
else
video["title"] = "Something went wrong"
2018-01-15 04:16:09 +01:00
end
2018-01-16 05:11:51 +01:00
thumbnail = root.xpath_node(%q(div[contains(@class,"yt-lockup-thumbnail")]/a/div/span/img/@src))
if thumbnail && !thumbnail.content.ends_with?(".gif")
video["thumbnail"] = thumbnail.content
else
thumbnail = root.xpath_node(%q(div[contains(@class,"yt-lockup-thumbnail")]/a/div/span/img/@data-thumb))
if thumbnail
video["thumbnail"] = thumbnail.content
else
video["thumbnail"] = "http://via.placeholder.com/246x138"
end
end
videos << video
end
end
2018-02-05 02:40:06 +01:00
pool << client
2017-12-30 22:21:43 +01:00
templated "search"
end
2018-02-10 16:15:23 +01:00
error 404 do |env|
error_message = "404 Page not found"
templated "error"
2017-12-30 22:21:43 +01:00
end
error 500 do |env|
2018-02-10 16:15:23 +01:00
error_message = "500 Server error"
templated "error"
2017-12-30 22:21:43 +01:00
end
2017-11-23 08:48:55 +01:00
public_folder "assets"
Kemal.run