share/lua/website/youtube.lua

   1
   2 -- libquvi-scripts
   3 -- Copyright (C) 2010-2012  Toni Gundogdu <legatvs@gmail.com>
   4 --
   5 -- This file is part of libquvi-scripts <http://quvi.sourceforge.net/>.
   6 --
   7 -- This library is free software; you can redistribute it and/or
   8 -- modify it under the terms of the GNU Lesser General Public
   9 -- License as published by the Free Software Foundation; either
  10 -- version 2.1 of the License, or (at your option) any later version.
  11 --
  12 -- This library is distributed in the hope that it will be useful,
  13 -- but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 -- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  15 -- Lesser General Public License for more details.
  16 --
  17 -- You should have received a copy of the GNU Lesser General Public
  18 -- License along with this library; if not, write to the Free Software
  19 -- Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
  20 -- 02110-1301  USA
  21 --
  22
  23 local YouTube = {} -- Utility functions unique to this script
  24
  25 -- <http://en.wikipedia.org/wiki/YouTube#Quality_and_codecs>
  26
  27 -- Identify the script.
  28 function ident(self)
  29     package.path = self.script_dir .. '/?.lua'
  30     local C      = require 'quvi/const'
  31     local r      = {}
  32     r.domain     = "youtube%.com"
  33     r.formats    = "default|best"
  34     r.categories  = C.proto_http
  35     self.page_url = YouTube.normalize(self.page_url)
  36     local U       = require 'quvi/util'
  37     r.handles     = U.handles(self.page_url,
  38                         {r.domain}, {"/watch"}, {"v=[%w-_]+"})
  39     return r
  40 end
  41
  42 -- Query available formats.
  43 function query_formats(self)
  44     local config,U  = YouTube.get_config(self)
  45     local formats   = YouTube.iter_formats(config, U)
  46
  47     local t = {}
  48     for _,v in pairs(formats) do
  49         table.insert(t, YouTube.to_s(v))
  50     end
  51
  52     table.sort(t)
  53     self.formats = table.concat(t, "|")
  54
  55     return self
  56 end
  57
  58 -- Parse URL.
  59 function parse(self)
  60     self.host_id = "youtube"
  61
  62     local p_url = YouTube.normalize(self.page_url)
  63     self.start_time = p_url:match('#a?t=(.+)') or ''
  64
  65     return YouTube.get_video_info(self)
  66 end
  67
  68 --
  69 -- Utility functions
  70 --
  71
  72 function YouTube.normalize(s)
  73     if not s then return s end
  74     local U = require 'quvi/url'
  75     local t = U.parse(s)
  76     if not t.host then return s end
  77     t.host = t.host:gsub('youtu%.be', 'youtube.com')
  78     t.host = t.host:gsub('-nocookie', '')
  79     if t.path then
  80         local p = {'/embed/([-_%w]+)', '/%w/([-_%w]+)', '/([-_%w]+)'}
  81         for _,v in pairs(p) do
  82             local m = t.path:match(v)
  83             if m and #m == 11 then
  84                 t.query = 'v=' .. m
  85                 t.path  = '/watch'
  86             end
  87         end
  88     end
  89     return U.build(t)
  90 end
  91
  92 function YouTube.get_config(self)
  93     local sch = self.page_url:match('^(%w+)://')
  94                   or error("no match: scheme")
  95
  96     local p_url = YouTube.normalize(self.page_url)
  97
  98     self.id = p_url:match("v=([%w-_]+)")
  99                 or error("no match: media ID")
 100
 101     local s_fmt = "%s://www.youtube.com/get_video_info?&video_id=%s"
 102                     .. "&el=detailpage&ps=default&eurl=&gl=US&hl=en"
 103
 104     local c_url = string.format(s_fmt, sch, self.id)
 105
 106     local U = require 'quvi/util'
 107     local c = U.decode(quvi.fetch(c_url, {fetch_type='config'}))
 108
 109     if c['reason'] then
 110         local reason = U.unescape(c['reason'])
 111         local code = c['errorcode']
 112         error(string.format("%s (code=%s)", reason, code))
 113     end
 114
 115     return c, U
 116 end
 117
 118 function YouTube.iter_formats(config, U)
 119     local fmt_stream_map = config['url_encoded_fmt_stream_map']
 120                         or error("no match: url_encoded_fmt_stream_map")
 121
 122     fmt_stream_map = U.unescape(fmt_stream_map) .. ","
 123
 124     local urls = {}
 125     for f in fmt_stream_map:gmatch('([^,]*),') do
 126         local d = U.decode(f)
 127         if d['itag'] and d['url'] then
 128             local uurl = U.unescape(d['url'])
 129             if d['sig'] then
 130                 uurl = uurl .. "&signature=" .. U.unescape(d['sig'])
 131             end
 132             urls[U.unescape(d['itag'])] = uurl
 133         end
 134     end
 135
 136     local fmt_map = config['fmt_list'] or error("no match: fmt_list")
 137     fmt_map = U.unescape(fmt_map)
 138
 139     local r = {}
 140     for f,w,h in fmt_map:gmatch('(%d+)/(%d+)x(%d+)') do
 141 --        print(f,w,h)
 142         table.insert(r, {fmt_id=tonumber(f),    url=urls[f],
 143                           width=tonumber(w), height=tonumber(h)})
 144     end
 145
 146     return r
 147 end
 148
 149 function YouTube.get_video_info(self)
 150     local config,U = YouTube.get_config(self)
 151
 152     self.title = config['title'] or error('no match: media title')
 153     self.title = U.unescape(self.title)
 154
 155     self.thumbnail_url = config['thumbnail_url'] or ''
 156     if #self.thumbnail_url > 0 then
 157         self.thumbnail_url = U.unescape(self.thumbnail_url)
 158     end
 159
 160     self.duration = (config['length_seconds'] or 0)*1000 -- to msec
 161
 162     self.requested_format =
 163         YouTube.convert_deprecated_id(self.requested_format)
 164
 165     local formats = YouTube.iter_formats(config, U)
 166     local format  = U.choose_format(self, formats,
 167                                     YouTube.choose_best,
 168                                     YouTube.choose_default,
 169                                     YouTube.to_s)
 170                         or error("unable to choose format")
 171     local url     = format.url or error("no match: media url")
 172
 173     if url and #self.start_time > 0 then
 174         local min, sec = self.start_time:match("^(%d+)m(%d+)s$")
 175         min = tonumber(min) or 0
 176         sec = tonumber(sec) or 0
 177         local msec = (min * 60000) + (sec * 1000)
 178         if msec > 0 then
 179             url = url .. "&begin=" .. msec
 180         end
 181     end
 182
 183     self.url = {url}
 184     return self
 185 end
 186
 187 function YouTube.choose_best(formats) -- Highest quality available
 188     local r = {width=0, height=0, url=nil}
 189     local U = require 'quvi/util'
 190     for _,v in pairs(formats) do
 191         if U.is_higher_quality(v,r) then
 192             r = v
 193         end
 194     end
 195 --    for k,v in pairs(r) do print(k,v) end
 196     return r
 197 end
 198
 199 function YouTube.choose_default(formats)
 200     local r = formats[1] -- Either whatever YouTube returns as the first.
 201     for _,v in pairs(formats) do
 202         if v.height == 480 then -- Or, whichever is of 480p and found first.
 203             r = v
 204             break
 205         end
 206     end
 207     return r
 208 end
 209
 210 YouTube.conv_table = { -- Deprecated.
 211     -- flv
 212     flv_240p =  '5',
 213     flv_360p = '34',
 214     flv_480p = '35',
 215     -- mp4
 216      mp4_360p = '18',
 217      mp4_720p = '22',
 218     mp4_1080p = '37',
 219     mp4_3072p = '38'
 220 }
 221
 222 function YouTube.convert_deprecated_id(r_fmt)
 223     if YouTube.conv_table[r_fmt] then
 224         local s = string.format("fmt%02d_", YouTube.conv_table[r_fmt])
 225         r_fmt = r_fmt:gsub("^(%w+)_", s)
 226     end
 227     return r_fmt
 228 end
 229
 230 function YouTube.to_s(t)
 231     return string.format("fmt%02d_%sp", t.fmt_id, t.height)
 232 end
 233
 234 --[[
 235 local a = {
 236   {u='http://youtu.be/3WSQH__H1XE',             -- u=page url
 237    e='http://youtube.com/watch?v=3WSQH__H1XE'}, -- e=expected url
 238   {u='http://youtu.be/v/3WSQH__H1XE?hl=en',
 239    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 240   {u='http://youtu.be/watch?v=3WSQH__H1XE',
 241    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 242   {u='http://youtu.be/embed/3WSQH__H1XE',
 243    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 244   {u='http://youtu.be/v/3WSQH__H1XE',
 245    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 246   {u='http://youtu.be/e/3WSQH__H1XE',
 247    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 248   {u='http://youtube.com/watch?v=3WSQH__H1XE',
 249    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 250   {u='http://youtube.com/embed/3WSQH__H1XE',
 251    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 252   {u='http://jp.youtube.com/watch?v=3WSQH__H1XE',
 253    e='http://jp.youtube.com/watch?v=3WSQH__H1XE'},
 254   {u='http://jp.youtube-nocookie.com/e/3WSQH__H1XE',
 255    e='http://jp.youtube.com/watch?v=3WSQH__H1XE'},
 256   {u='http://jp.youtube.com/embed/3WSQH__H1XE',
 257    e='http://jp.youtube.com/watch?v=3WSQH__H1XE'},
 258   {u='http://youtube.com/3WSQH__H1XE', -- invalid page url
 259    e='http://youtube.com/watch?v=3WSQH__H1XE'}
 260 }
 261 local e = 0
 262 for i,v in pairs(a) do
 263   local s = YouTube.normalize(v.u)
 264   if s ~= v.e then
 265     print('\n   input: ' .. v.u .. " (#" .. i .. ")")
 266     print('expected: '   .. v.e)
 267     print('     got: '   .. s)
 268     e = e + 1
 269   end
 270 end
 271 print((e == 0) and 'Tests OK' or ('\nerrors: ' .. e))
 272 ]]--
 273
 274 -- vim: set ts=4 sw=4 tw=72 expandtab: