2018-01-28 02:09:27 +00:00
|
|
|
class Video
|
|
|
|
module HTTPParamConverter
|
|
|
|
def self.from_rs(rs)
|
|
|
|
HTTP::Params.parse(rs.read(String))
|
|
|
|
end
|
|
|
|
end
|
|
|
|
|
|
|
|
module XMLConverter
|
|
|
|
def self.from_rs(rs)
|
|
|
|
XML.parse_html(rs.read(String))
|
|
|
|
end
|
|
|
|
end
|
|
|
|
|
2018-02-03 03:44:10 +00:00
|
|
|
def initialize(id, info, html, updated, title, views, likes, dislikes, wilson_score, published)
|
2018-01-28 02:09:27 +00:00
|
|
|
@id = id
|
|
|
|
@info = info
|
|
|
|
@html = html
|
|
|
|
@updated = updated
|
|
|
|
@title = title
|
|
|
|
@views = views
|
|
|
|
@likes = likes
|
|
|
|
@dislikes = dislikes
|
|
|
|
@wilson_score = wilson_score
|
2018-02-03 03:44:10 +00:00
|
|
|
@published = published
|
2018-01-28 02:09:27 +00:00
|
|
|
end
|
|
|
|
|
|
|
|
def to_a
|
2018-02-03 03:44:10 +00:00
|
|
|
return [@id, @info, @html, @updated, @title, @views, @likes, @dislikes, @wilson_score, @published]
|
2018-01-28 02:09:27 +00:00
|
|
|
end
|
|
|
|
|
|
|
|
DB.mapping({
|
|
|
|
id: String,
|
|
|
|
info: {
|
|
|
|
type: HTTP::Params,
|
|
|
|
default: HTTP::Params.parse(""),
|
|
|
|
converter: Video::HTTPParamConverter,
|
|
|
|
},
|
|
|
|
html: {
|
|
|
|
type: XML::Node,
|
|
|
|
default: XML.parse_html(""),
|
|
|
|
converter: Video::XMLConverter,
|
|
|
|
},
|
|
|
|
updated: Time,
|
|
|
|
title: String,
|
|
|
|
views: Int64,
|
|
|
|
likes: Int32,
|
|
|
|
dislikes: Int32,
|
|
|
|
wilson_score: Float64,
|
2018-02-03 03:44:10 +00:00
|
|
|
published: Time,
|
2018-01-28 02:09:27 +00:00
|
|
|
})
|
|
|
|
end
|
|
|
|
|
2018-01-21 00:19:12 +00:00
|
|
|
# See http://www.evanmiller.org/how-not-to-sort-by-average-rating.html
|
|
|
|
def ci_lower_bound(pos, n)
|
|
|
|
if n == 0
|
2018-01-28 02:09:27 +00:00
|
|
|
return 0.0
|
2018-01-21 00:19:12 +00:00
|
|
|
end
|
|
|
|
|
|
|
|
# z value here represents a confidence level of 0.95
|
|
|
|
z = 1.96
|
|
|
|
phat = 1.0*pos/n
|
|
|
|
|
|
|
|
return (phat + z*z/(2*n) - z * Math.sqrt((phat*(1 - phat) + z*z/(4*n))/n))/(1 + z*z/n)
|
|
|
|
end
|
|
|
|
|
|
|
|
def elapsed_text(elapsed)
|
|
|
|
millis = elapsed.total_milliseconds
|
|
|
|
return "#{millis.round(2)}ms" if millis >= 1
|
|
|
|
|
|
|
|
"#{(millis * 1000).round(2)}µs"
|
|
|
|
end
|
|
|
|
|
2018-01-28 02:09:27 +00:00
|
|
|
def get_client(pool)
|
|
|
|
while pool.empty?
|
2018-01-21 00:19:12 +00:00
|
|
|
sleep rand(0..10).milliseconds
|
|
|
|
end
|
|
|
|
|
2018-01-28 02:09:27 +00:00
|
|
|
return pool.shift
|
2018-01-21 00:19:12 +00:00
|
|
|
end
|
|
|
|
|
2018-01-28 02:09:27 +00:00
|
|
|
def fetch_video(id, client)
|
|
|
|
begin
|
|
|
|
info = client.get("/get_video_info?video_id=#{id}&el=detailpage&ps=default&eurl=&gl=US&hl=en").body
|
|
|
|
html = client.get("/watch?v=#{id}").body
|
|
|
|
end
|
2018-01-21 00:19:12 +00:00
|
|
|
|
|
|
|
html = XML.parse_html(html)
|
2018-01-28 02:09:27 +00:00
|
|
|
info = HTTP::Params.parse(info)
|
2018-01-21 00:19:12 +00:00
|
|
|
|
|
|
|
if info["reason"]?
|
2018-02-03 04:04:34 +00:00
|
|
|
info = client.get("/get_video_info?video_id=#{id}&ps=default&eurl=&gl=US&hl=en").body
|
|
|
|
info = HTTP::Params.parse(info)
|
|
|
|
if info["reason"]?
|
|
|
|
raise info["reason"]
|
|
|
|
end
|
2018-01-21 00:19:12 +00:00
|
|
|
end
|
|
|
|
|
2018-01-28 02:09:27 +00:00
|
|
|
title = info["title"]
|
|
|
|
|
|
|
|
views = info["view_count"].to_i64
|
2018-01-21 00:19:12 +00:00
|
|
|
|
2018-01-28 02:09:27 +00:00
|
|
|
likes = html.xpath_node(%q(//button[@title="I like this"]/span))
|
2018-02-05 01:42:13 +00:00
|
|
|
likes = likes ? likes.content.delete(",").to_i : 0
|
2018-01-28 02:09:27 +00:00
|
|
|
|
|
|
|
dislikes = html.xpath_node(%q(//button[@title="I dislike this"]/span))
|
|
|
|
dislikes = dislikes ? dislikes.content.delete(",").to_i : 0
|
|
|
|
|
|
|
|
wilson_score = ci_lower_bound(likes, likes + dislikes)
|
|
|
|
|
2018-02-05 01:42:13 +00:00
|
|
|
published = html.xpath_node(%q(//strong[contains(@class,"watch-time-text")]))
|
2018-02-03 03:44:10 +00:00
|
|
|
if published
|
|
|
|
published = published.content
|
2018-02-05 01:42:13 +00:00
|
|
|
else
|
|
|
|
raise "Could not find date published"
|
|
|
|
end
|
|
|
|
|
|
|
|
published = published.lchop("Published ")
|
|
|
|
published = published.lchop("Streamed live ")
|
|
|
|
published = published.lchop("Started streaming ")
|
|
|
|
published = published.lchop("on ")
|
|
|
|
published = published.lchop("Scheduled for ")
|
2018-02-05 23:57:03 +00:00
|
|
|
if !published.includes?("ago")
|
|
|
|
published = Time.parse(published, "%b %-d, %Y")
|
|
|
|
else
|
|
|
|
# Time matches format "20 hours ago", "40 minutes ago"...
|
2018-02-05 01:42:13 +00:00
|
|
|
delta = published.split(" ")[0].to_i
|
|
|
|
case published
|
|
|
|
when .includes? "minute"
|
|
|
|
published = Time.now - delta.minutes
|
|
|
|
when .includes? "hour"
|
|
|
|
published = Time.now - delta.hours
|
2018-02-05 23:57:03 +00:00
|
|
|
else
|
2018-02-05 01:42:13 +00:00
|
|
|
raise "Could not parse #{published}"
|
2018-02-05 23:57:03 +00:00
|
|
|
end
|
2018-02-03 03:44:10 +00:00
|
|
|
end
|
|
|
|
|
|
|
|
video = Video.new(id, info, html, Time.now, title, views, likes, dislikes, wilson_score, published)
|
2018-01-21 00:19:12 +00:00
|
|
|
|
|
|
|
return video
|
|
|
|
end
|
|
|
|
|
2018-01-28 02:09:27 +00:00
|
|
|
def get_video(id, client, db, refresh = true)
|
|
|
|
if db.query_one?("SELECT EXISTS (SELECT true FROM videos WHERE id = $1)", id, as: Bool)
|
|
|
|
video = db.query_one("SELECT * FROM videos WHERE id = $1", id, as: Video)
|
2018-01-21 00:19:12 +00:00
|
|
|
|
2018-01-28 02:09:27 +00:00
|
|
|
# If record was last updated over an hour ago, refresh (expire param in response lasts for 6 hours)
|
|
|
|
if refresh && Time.now - video.updated > 1.hours
|
|
|
|
video = fetch_video(id, client)
|
|
|
|
db.exec("UPDATE videos SET info = $2, html = $3, updated = $4,\
|
2018-02-03 03:44:10 +00:00
|
|
|
title = $5, views = $6, likes = $7, dislikes = $8, wilson_score = $9, published = $10 WHERE id = $1", video.to_a)
|
2018-01-21 00:19:12 +00:00
|
|
|
end
|
|
|
|
else
|
2018-01-28 02:09:27 +00:00
|
|
|
video = fetch_video(id, client)
|
2018-02-03 03:44:10 +00:00
|
|
|
db.exec("INSERT INTO videos VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10)", video.to_a)
|
2018-01-21 00:19:12 +00:00
|
|
|
end
|
|
|
|
|
|
|
|
return video
|
|
|
|
end
|
2018-01-21 23:49:27 +00:00
|
|
|
|
2018-01-28 02:09:27 +00:00
|
|
|
def search(query, client)
|
|
|
|
begin
|
|
|
|
html = client.get("https://www.youtube.com/results?q=#{query}&sp=EgIQAVAU").body
|
|
|
|
end
|
2018-01-21 23:49:27 +00:00
|
|
|
|
|
|
|
html = XML.parse_html(html)
|
|
|
|
|
|
|
|
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
|
|
|
|
link = root.xpath_node(%q(div[contains(@class,"yt-lockup-thumbnail")]/a/@href))
|
|
|
|
if link
|
|
|
|
yield link.content.split("=")[1]
|
|
|
|
end
|
|
|
|
end
|
|
|
|
end
|
|
|
|
end
|
2018-02-03 20:41:59 +00:00
|
|
|
|
2018-02-22 19:01:37 +00:00
|
|
|
def splice(a, b)
|
2018-02-03 20:41:59 +00:00
|
|
|
c = a[0]
|
2018-02-12 04:06:29 +00:00
|
|
|
a[0] = a[b % a.size]
|
|
|
|
a[b % a.size] = c
|
|
|
|
return a
|
|
|
|
end
|
|
|
|
|
2018-02-22 19:01:37 +00:00
|
|
|
def decrypt_signature(a)
|
2018-02-13 16:44:11 +00:00
|
|
|
a = a.split("")
|
2018-02-12 04:06:29 +00:00
|
|
|
|
2018-02-22 19:01:37 +00:00
|
|
|
a.delete_at(0..1)
|
|
|
|
a = splice(a, 2)
|
|
|
|
a = splice(a, 51)
|
|
|
|
a = splice(a, 9)
|
|
|
|
a.delete_at(0..1)
|
|
|
|
a.reverse!
|
|
|
|
a = splice(a, 15)
|
|
|
|
a.delete_at(0..2)
|
2018-02-05 23:56:40 +00:00
|
|
|
|
2018-02-13 16:44:11 +00:00
|
|
|
return a.join("")
|
|
|
|
end
|
|
|
|
|
2018-02-05 23:56:40 +00:00
|
|
|
def rank_videos(db, n)
|
|
|
|
top = [] of {Float64, String}
|
|
|
|
|
|
|
|
db.query("SELECT id, wilson_score, published FROM videos WHERE views > 5000 ORDER BY published DESC LIMIT 10000") 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
|
2018-02-10 16:06:37 +00:00
|
|
|
temperature = wilson_score * Math.exp(-0.000005*((Time.now - published).total_minutes))
|
2018-02-05 23:56:40 +00:00
|
|
|
top << {temperature, id}
|
|
|
|
end
|
|
|
|
end
|
|
|
|
|
|
|
|
top.sort!
|
|
|
|
|
|
|
|
# Make hottest come first
|
|
|
|
top.reverse!
|
|
|
|
top = top.map { |a, b| b }
|
|
|
|
|
|
|
|
# Return top
|
2018-02-13 19:14:21 +00:00
|
|
|
return top[0..n - 1]
|
2018-02-05 23:56:40 +00:00
|
|
|
end
|
2018-02-06 01:07:49 +00:00
|
|
|
|
|
|
|
def make_client(url, context)
|
|
|
|
client = HTTP::Client.new(URL, CONTEXT)
|
|
|
|
client.read_timeout = 10.seconds
|
|
|
|
client.connect_timeout = 10.seconds
|
|
|
|
return client
|
|
|
|
end
|