share/lua/website/youtube.lua

   1
   2 -- libquvi-scripts
   3 -- Copyright (C) 2010-2011  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     local page_url = YouTube.normalize(self.page_url)
  62
  63     local _,_,s = page_url:find('#a?t=(.+)')
  64     self.start_time = s or ''
  65
  66     return YouTube.get_video_info(self)
  67 end
  68
  69 --
  70 -- Utility functions
  71 --
  72
  73 function YouTube.normalize(s)
  74     if not s then return s end
  75     local U = require 'quvi/url'
  76     local t = U.parse(s)
  77     if not t.host then return s end
  78     t.host = t.host:gsub('youtu%.be', 'youtube.com')
  79     t.host = t.host:gsub('-nocookie', '')
  80     if t.path then
  81         local p = {'/embed/([-_%w]+)', '/%w/([-_%w]+)', '/([-_%w]+)'}
  82         for _,v in pairs(p) do
  83             local m = t.path:match(v)
  84             if m and not t.query then
  85                 t.query = 'v=' .. m
  86                 t.path  = '/watch'
  87             end
  88         end
  89     end
  90     return U.build(t)
  91 end
  92
  93 function YouTube.get_config(self)
  94     local _,_,s  = self.page_url:find('^(%w+)://')
  95     local scheme = s or error("no match: scheme")
  96
  97     local page_url = YouTube.normalize(self.page_url)
  98
  99     local _,_,s = page_url:find("v=([%w-_]+)")
 100     self.id = s or error("no match: media id")
 101
 102     local s_fmt = "%s://www.youtube.com/get_video_info?&video_id=%s"
 103                     .. "&el=detailpage&ps=default&eurl=&gl=US&hl=en"
 104
 105     local config_url = string.format(s_fmt, scheme, self.id)
 106
 107     local U      = require 'quvi/util'
 108     local config = U.decode(quvi.fetch(config_url, {fetch_type='config'}))
 109
 110     if config['reason'] then
 111         local reason = U.unescape(config['reason'])
 112         local code   = config['errorcode']
 113         error(string.format("%s (code=%s)", reason, code))
 114     end
 115
 116     return config,U
 117 end
 118
 119 function YouTube.iter_formats(config, U)
 120     local fmt_stream_map = config['url_encoded_fmt_stream_map']
 121                         or error("no match: url_encoded_fmt_stream_map")
 122
 123     fmt_stream_map = U.unescape(fmt_stream_map) .. ","
 124
 125     local urls = {}
 126     for f in fmt_stream_map:gfind('([^,]*),') do
 127         local d = U.decode(f)
 128         if d['itag'] and d['url'] then
 129             urls[U.unescape(d['itag'])] = U.unescape(d['url'])
 130         end
 131     end
 132
 133     local fmt_map = config['fmt_list'] or error("no match: fmt_list")
 134     fmt_map = U.unescape(fmt_map)
 135
 136     local r = {}
 137     for f,w,h in fmt_map:gfind('(%d+)/(%d+)x(%d+)') do
 138 --        print(f,w,h)
 139         table.insert(r, {fmt_id=tonumber(f),    url=urls[f],
 140                           width=tonumber(w), height=tonumber(h)})
 141     end
 142
 143     return r
 144 end
 145
 146 function YouTube.get_video_info(self)
 147     local config,U = YouTube.get_config(self)
 148
 149     self.title = config['title'] or error('no match: media title')
 150     self.title = U.unescape(self.title)
 151
 152     self.thumbnail_url = config['thumbnail_url'] or ''
 153     if #self.thumbnail_url > 0 then
 154         self.thumbnail_url = U.unescape(self.thumbnail_url)
 155     end
 156
 157     self.duration = (config['length_seconds'] or 0)*1000 -- to msec
 158
 159     self.requested_format =
 160         YouTube.convert_deprecated_id(self.requested_format)
 161
 162     local formats = YouTube.iter_formats(config, U)
 163     local url     = U.choose_format(self, formats,
 164                                     YouTube.choose_best,
 165                                     YouTube.choose_default,
 166                                     YouTube.to_s).url
 167                         or error("no match: media url")
 168
 169     if url and #self.start_time > 0 then
 170         local min, sec = self.start_time:match("^(%d+)m(%d+)s$")
 171         min = tonumber(min) or 0
 172         sec = tonumber(sec) or 0
 173         local msec = (min * 60000) + (sec * 1000)
 174         if msec > 0 then
 175             url = url .. "&begin=" .. msec
 176         end
 177     end
 178
 179     self.url = {url}
 180
 181     return self
 182 end
 183
 184 function YouTube.choose_best(formats) -- Highest quality available
 185     local r = {width=0, height=0, url=nil}
 186     local U = require 'quvi/util'
 187     for _,v in pairs(formats) do
 188         if U.is_higher_quality(v,r) then
 189             r = v
 190         end
 191     end
 192 --    for k,v in pairs(r) do print(k,v) end
 193     return r
 194 end
 195
 196 function YouTube.choose_default(formats) -- Lowest quality available
 197     local r = {width=0xffff, height=0xffff, url=nil}
 198     local U = require 'quvi/util'
 199     for _,v in pairs(formats) do
 200         if U.is_lower_quality(v,r) then
 201             r = v
 202         end
 203     end
 204 --    for k,v in pairs(r) do print(k,v) end
 205     return r
 206 end
 207
 208 YouTube.conv_table = { -- Deprecated.
 209     -- flv
 210     flv_240p =  '5',
 211     flv_360p = '34',
 212     flv_480p = '35',
 213     -- mp4
 214      mp4_360p = '18',
 215      mp4_720p = '22',
 216     mp4_1080p = '37',
 217     mp4_3072p = '38'
 218 }
 219
 220 function YouTube.convert_deprecated_id(r_fmt)
 221     if YouTube.conv_table[r_fmt] then
 222         local s = string.format("fmt%02d_", YouTube.conv_table[r_fmt])
 223         r_fmt = r_fmt:gsub("^(%w+)_", s)
 224     end
 225     return r_fmt
 226 end
 227
 228 function YouTube.to_s(t)
 229     return string.format("fmt%02d_%sp", t.fmt_id, t.height)
 230 end
 231
 232 --[[
 233 local a = {
 234   {u='http://youtu.be/3WSQH__H1XE',             -- u=page url
 235    e='http://youtube.com/watch?v=3WSQH__H1XE'}, -- e=expected url
 236   {u='http://youtu.be/watch?v=3WSQH__H1XE',
 237    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 238   {u='http://youtu.be/embed/3WSQH__H1XE',
 239    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 240   {u='http://youtu.be/v/3WSQH__H1XE',
 241    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 242   {u='http://youtu.be/e/3WSQH__H1XE',
 243    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 244   {u='http://youtube.com/watch?v=3WSQH__H1XE',
 245    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 246   {u='http://youtube.com/embed/3WSQH__H1XE',
 247    e='http://youtube.com/watch?v=3WSQH__H1XE'},
 248   {u='http://jp.youtube.com/watch?v=3WSQH__H1XE',
 249    e='http://jp.youtube.com/watch?v=3WSQH__H1XE'},
 250   {u='http://jp.youtube-nocookie.com/e/3WSQH__H1XE',
 251    e='http://jp.youtube.com/watch?v=3WSQH__H1XE'},
 252   {u='http://jp.youtube.com/embed/3WSQH__H1XE',
 253    e='http://jp.youtube.com/watch?v=3WSQH__H1XE'},
 254   {u='http://youtube.com/3WSQH__H1XE', -- invalid page url
 255    e='http://youtube.com/watch?v=3WSQH__H1XE'}
 256 }
 257 local e = 0
 258 for i,v in pairs(a) do
 259   local s = YouTube.normalize(v.u)
 260   if s ~= v.e then
 261     print('\n   input: ' .. v.u .. " (#" .. i .. ")")
 262     print('expected: '   .. v.e)
 263     print('     got: '   .. s)
 264     e = e + 1
 265   end
 266 end
 267 print('\nerrors: ' .. e)
 268 ]]--
 269
 270 -- vim: set ts=4 sw=4 tw=72 expandtab: