Merge branch 'master' into api-only

This commit is contained in:
Omar Roth 2018-11-25 18:07:01 -06:00
commit 564fba87d2
7 changed files with 111 additions and 108 deletions

View File

@ -16,7 +16,9 @@ dependencies:
commit: afd17fc commit: afd17fc
pg: pg:
github: will/crystal-pg github: will/crystal-pg
sqlite3:
github: crystal-lang/crystal-sqlite3
crystal: 0.26.1 crystal: 0.27.0
license: AGPLv3 license: AGPLv3

View File

@ -16,10 +16,12 @@
require "crypto/bcrypt/password" require "crypto/bcrypt/password"
require "detect_language" require "detect_language"
require "digest/md5"
require "kemal" require "kemal"
require "openssl/hmac" require "openssl/hmac"
require "option_parser" require "option_parser"
require "pg" require "pg"
require "sqlite3"
require "xml" require "xml"
require "yaml" require "yaml"
require "zip" require "zip"
@ -524,14 +526,19 @@ get "/api/v1/videos/:id" do |env|
end end
get "/api/v1/trending" do |env| get "/api/v1/trending" do |env|
client = make_client(YT_URL) region = env.params.query["region"]?
trending = client.get("/feed/trending?disable_polymer=1").body trending_type = env.params.query["type"]?
begin
trending = fetch_trending(trending_type, proxies, region)
rescue ex
error_message = {"error" => ex.message}.to_json
halt env, status_code: 500, response: error_message
end
trending = XML.parse_html(trending)
videos = JSON.build do |json| videos = JSON.build do |json|
json.array do json.array do
nodeset = trending.xpath_nodes(%q(//ul/li[@class="expanded-shelf-content-item-wrapper"])) trending.each do |video|
extract_videos(nodeset).each do |video|
json.object do json.object do
json.field "title", video.title json.field "title", video.title
json.field "videoId", video.id json.field "videoId", video.id
@ -550,6 +557,9 @@ get "/api/v1/trending" do |env|
json.field "publishedText", "#{recode_date(video.published)} ago" json.field "publishedText", "#{recode_date(video.published)} ago"
json.field "description", video.description json.field "description", video.description
json.field "descriptionHtml", video.description_html json.field "descriptionHtml", video.description_html
json.field "liveNow", video.live_now
json.field "paid", video.paid
json.field "premium", video.premium
end end
end end
end end
@ -1367,45 +1377,24 @@ get "/videoplayback" do |env|
host = "https://r#{fvip}---#{mn}.googlevideo.com" host = "https://r#{fvip}---#{mn}.googlevideo.com"
url = "/videoplayback?#{query_params.to_s}" url = "/videoplayback?#{query_params.to_s}"
if query_params["region"]? region = query_params["region"]?
client = make_client(URI.parse(host)) client = make_client(URI.parse(host), proxies, region)
response = HTTP::Client::Response.new(status_code: 403) response = client.head(url)
if !proxies[query_params["region"]]?
halt env, status_code: 403
end
proxies[query_params["region"]].each do |proxy|
begin
client = HTTPClient.new(URI.parse(host))
client.read_timeout = 10.seconds
client.connect_timeout = 10.seconds
proxy = HTTPProxy.new(proxy_host: proxy[:ip], proxy_port: proxy[:port])
client.set_proxy(proxy)
response = client.head(url)
if response.status_code == 200
# For whatever reason the proxy needs to be set again
client.set_proxy(proxy)
break
end
rescue ex
end
end
else
client = make_client(URI.parse(host))
response = client.head(url)
end
if response.status_code != 200
halt env, status_code: 403
end
if response.headers["Location"]? if response.headers["Location"]?
url = URI.parse(response.headers["Location"]) url = URI.parse(response.headers["Location"])
env.response.headers["Access-Control-Allow-Origin"] = "*" env.response.headers["Access-Control-Allow-Origin"] = "*"
next env.redirect url.full_path
url = url.full_path
if region
url += "&region=#{region}"
end
next env.redirect url
end
if response.status_code >= 400
halt env, status_code: 403
end end
headers = env.request.headers headers = env.request.headers
@ -1414,6 +1403,7 @@ get "/videoplayback" do |env|
headers.delete("User-Agent") headers.delete("User-Agent")
headers.delete("Referer") headers.delete("Referer")
client = make_client(URI.parse(host), proxies, region)
client.get(url, headers) do |response| client.get(url, headers) do |response|
env.response.status_code = response.status_code env.response.status_code = response.status_code

View File

@ -70,34 +70,18 @@ def fetch_youtube_comments(id, continuation, proxies, format)
if body.match(/<meta itemprop="regionsAllowed" content="">/) if body.match(/<meta itemprop="regionsAllowed" content="">/)
bypass_channel = Channel({String, HTTPClient, HTTP::Headers} | Nil).new bypass_channel = Channel({String, HTTPClient, HTTP::Headers} | Nil).new
proxies.each do |region, list| proxies.each do |proxy_region, list|
spawn do spawn do
proxy_html = %(<meta itemprop="regionsAllowed" content="">) proxy_client = make_client(YT_URL, proxies, proxy_region)
list.each do |proxy| response = proxy_client.get("/watch?v=#{id}&gl=US&hl=en&disable_polymer=1&has_verified=1&bpctr=9999999999")
begin proxy_headers = HTTP::Headers.new
proxy_client = HTTPClient.new(YT_URL) proxy_headers["Cookie"] = response.cookies.add_request_headers(headers)["cookie"]
proxy_client.read_timeout = 10.seconds proxy_html = response.body
proxy_client.connect_timeout = 10.seconds
proxy = HTTPProxy.new(proxy_host: proxy[:ip], proxy_port: proxy[:port]) if !proxy_html.match(/<meta itemprop="regionsAllowed" content="">/)
proxy_client.set_proxy(proxy) bypass_channel.send({proxy_html, proxy_client, proxy_headers})
else
response = proxy_client.get("/watch?v=#{id}&gl=US&hl=en&disable_polymer=1&has_verified=1&bpctr=9999999999")
proxy_headers = HTTP::Headers.new
proxy_headers["cookie"] = response.cookies.add_request_headers(headers)["cookie"]
proxy_html = response.body
if !proxy_html.match(/<meta itemprop="regionsAllowed" content="">/)
bypass_channel.send({proxy_html, proxy_client, proxy_headers})
break
end
rescue ex
end
end
# If none of the proxies we tried returned a valid response
if proxy_html.match(/<meta itemprop="regionsAllowed" content="">/)
bypass_channel.send(nil) bypass_channel.send(nil)
end end
end end
@ -106,12 +90,12 @@ def fetch_youtube_comments(id, continuation, proxies, format)
proxies.size.times do proxies.size.times do
response = bypass_channel.receive response = bypass_channel.receive
if response if response
session_token = response[0].match(/'XSRF_TOKEN': "(?<session_token>[A-Za-z0-9\_\-\=]+)"/).not_nil!["session_token"] html, client, headers = response
itct = response[0].match(/itct=(?<itct>[^"]+)"/).not_nil!["itct"]
ctoken = response[0].match(/'COMMENTS_TOKEN': "(?<ctoken>[^"]+)"/) session_token = html.match(/'XSRF_TOKEN': "(?<session_token>[A-Za-z0-9\_\-\=]+)"/).not_nil!["session_token"]
itct = html.match(/itct=(?<itct>[^"]+)"/).not_nil!["itct"]
ctoken = html.match(/'COMMENTS_TOKEN': "(?<ctoken>[^"]+)"/)
client = response[1]
headers = response[2]
break break
end end
end end

View File

@ -30,7 +30,7 @@ def make_client(url, proxies = {} of String => Array({ip: String, port: Int32}),
client.connect_timeout = 10.seconds client.connect_timeout = 10.seconds
if region if region
proxies[region]?.try &.each do |proxy| proxies[region]?.try &.sample(40).each do |proxy|
begin begin
proxy = HTTPProxy.new(proxy_host: proxy[:ip], proxy_port: proxy[:port]) proxy = HTTPProxy.new(proxy_host: proxy[:ip], proxy_port: proxy[:port])
client.set_proxy(proxy) client.set_proxy(proxy)

41
src/invidious/trending.cr Normal file
View File

@ -0,0 +1,41 @@
def fetch_trending(trending_type, proxies, region)
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"
region ||= "US"
region = region.upcase
trending = ""
if trending_type
trending_type = trending_type.downcase.capitalize
response = client.get("/feed/trending?gl=#{region}&hl=en", headers).body
yt_data = response.match(/window\["ytInitialData"\] = (?<data>.*);/)
if yt_data
yt_data = JSON.parse(yt_data["data"].rchop(";"))
else
raise "Could not pull trending pages."
end
tabs = yt_data["contents"]["twoColumnBrowseResultsRenderer"]["tabs"][0]["tabRenderer"]["content"]["sectionListRenderer"]["subMenu"]["channelListSubMenuRenderer"]["contents"].as_a
url = tabs.select { |tab| tab["channelListSubMenuAvatarRenderer"]["title"]["simpleText"] == trending_type }[0]?
if url
url = url["channelListSubMenuAvatarRenderer"]["navigationEndpoint"]["commandMetadata"]["webCommandMetadata"]["url"].as_s
url += "&disable_polymer=1&gl=#{region}&hl=en"
trending = client.get(url).body
else
trending = client.get("/feed/trending?gl=#{region}&hl=en&disable_polymer=1").body
end
else
trending = client.get("/feed/trending?gl=#{region}&hl=en&disable_polymer=1").body
end
trending = XML.parse_html(trending)
nodeset = trending.xpath_nodes(%q(//ul/li[@class="expanded-shelf-content-item-wrapper"]))
trending = extract_videos(nodeset)
return trending
end

View File

@ -578,47 +578,26 @@ def fetch_video(id, proxies, region)
info = info_channel.receive info = info_channel.receive
if info["reason"]? && info["reason"].includes? "your country" if info["reason"]? && info["reason"].includes? "your country"
bypass_channel = Channel(HTTPProxy | Nil).new bypass_channel = Channel({HTTPClient, String} | Nil).new
proxies.each do |region, list| proxies.each do |proxy_region, list|
spawn do spawn do
info = HTTP::Params.new({ client = make_client(YT_URL, proxies, proxy_region)
"reason" => [info["reason"]],
})
list.each do |proxy| info = HTTP::Params.parse(client.get("/get_video_info?video_id=#{id}&ps=default&eurl=&gl=US&hl=en&disable_polymer=1").body)
begin if !info["reason"]?
client = HTTPClient.new(YT_URL) bypass_channel.send({client, proxy_region})
client.read_timeout = 10.seconds else
client.connect_timeout = 10.seconds
proxy = HTTPProxy.new(proxy_host: proxy[:ip], proxy_port: proxy[:port])
client.set_proxy(proxy)
info = HTTP::Params.parse(client.get("/get_video_info?video_id=#{id}&ps=default&eurl=&gl=US&hl=en&disable_polymer=1").body)
if !info["reason"]?
bypass_channel.send(proxy)
break
end
rescue ex
end
end
# If none of the proxies we tried returned a valid response
if info["reason"]?
bypass_channel.send(nil) bypass_channel.send(nil)
end end
end end
end end
proxies.size.times do proxies.size.times do
proxy = bypass_channel.receive response = bypass_channel.receive
if proxy if response
begin begin
client = HTTPClient.new(YT_URL) client, proxy_region = response
client.read_timeout = 10.seconds
client.connect_timeout = 10.seconds
client.set_proxy(proxy)
html = XML.parse_html(client.get("/watch?v=#{id}&gl=US&hl=en&disable_polymer=1&has_verified=1&bpctr=9999999999").body) html = XML.parse_html(client.get("/watch?v=#{id}&gl=US&hl=en&disable_polymer=1&has_verified=1&bpctr=9999999999").body)
info = HTTP::Params.parse(client.get("/get_video_info?video_id=#{id}&el=detailpage&ps=default&eurl=&gl=US&hl=en&disable_polymer=1").body) info = HTTP::Params.parse(client.get("/get_video_info?video_id=#{id}&el=detailpage&ps=default&eurl=&gl=US&hl=en&disable_polymer=1").body)
@ -627,11 +606,7 @@ def fetch_video(id, proxies, region)
info = HTTP::Params.parse(client.get("/get_video_info?video_id=#{id}&ps=default&eurl=&gl=US&hl=en&disable_polymer=1").body) info = HTTP::Params.parse(client.get("/get_video_info?video_id=#{id}&ps=default&eurl=&gl=US&hl=en&disable_polymer=1").body)
end end
proxy = {ip: proxy.proxy_host, port: proxy.proxy_port} info["region"] = proxy_region
region_proxies = proxies.select { |region, list| list.includes? proxy }
if !region_proxies.empty?
info["region"] = region_proxies.keys[0]
end
break break
rescue ex rescue ex

View File

@ -0,0 +1,11 @@
<% content_for "header" do %>
<title>Trending - Invidious</title>
<% end %>
<div class="pure-g">
<% trending.each_slice(4) do |slice| %>
<% slice.each do |item| %>
<%= rendered "components/item" %>
<% end %>
<% end %>
</div>