Media retrieval refactor; first pass

This commit is contained in:
Calvin Montgomery 2013-07-31 23:26:11 -04:00
parent 7477c64ae2
commit 90437e5ac2

View file

@ -14,6 +14,21 @@ var https = require("https");
var Logger = require("./logger.js");
var Media = require("./media.js").Media;
function urlRetrieve(transport, options, callback) {
var req = transport.request(options, function (res) {
var buffer = "";
res.setEncoding("utf-8");
res.on("data", function (chunk) {
buffer += chunk;
});
res.on("end", function () {
callback(res.statusCode, buffer);
});
});
req.end();
}
// Helper function for making an HTTP request and getting the result
// as JSON
function getJSONInternal(transport, options, callback) {
@ -63,6 +78,256 @@ function getJSONHTTPS(options, callback) {
getJSONInternal(https, options, callback);
}
var Getters = {
/* youtube.com */
yt: function (id, callback) {
var options = {
host: "gdata.youtube.com",
port: 443,
path: "/feeds/api/videos/" + id + "?v=2&alt=json",
method: "GET",
dataType: "jsonp",
timeout: 1000
};
urlRetrieve(https, options, function (status, data) {
if(status !== 200) {
callback(true, null);
return;
}
try {
data = JSON.parse(data);
var seconds = data.entry.media$group.yt$duration.seconds;
var title = data.entry.title.$t;
var media = new Media(id, title, seconds, "yt");
callback(false, media);
} catch(e) {
// Gdata version 2 has the rather silly habit of
// returning error codes in XML when I explicitly asked
// for JSON
var m = buffer.match(/<internalReason>([^<]+)<\/internalReason>/);
if(m === null)
m = buffer.match(/<code>([^<]+)<\/code>/);
var err = true;
if(m) {
if(m[1] === "too_many_recent_calls") {
err = "YouTube is throttling the server right "+
"now for making too many requests. "+
"Please try again in a moment.";
} else {
err = m[1];
}
}
callback(err, null);
}
});
},
// TODO Add youtube playlists
/* vimeo.com */
vi: function (id, callback) {
var options = {
host: "vimeo.com",
port: 443,
path: "/api/v2/video/" + id + ".json",
method: "GET",
dataType: "jsonp",
timeout: 1000
};
urlRetrieve(https, options, function (status, data) {
if(status !== 200) {
callback(true, null);
return;
}
try {
data = JSON.parse(data);
data = data[0];
var seconds = data.duration;
var title = data.title;
var media = new Media(id, title, seconds, "vi");
callback(false, media);
} catch(e) {
var err = true;
if(buffer.match(/not found/))
err = "Video not found";
callback(err, null);
}
});
},
/* dailymotion.com */
dm: function (id, callback) {
// Dailymotion's API is an example of an API done right
// - Supports SSL
// - I can ask for exactly which fields I want
// - URL is simple
// - Field names are sensible
// Other media providers take notes, please
var options = {
host: "api.dailymotion.com",
port: 443,
path: "/video/" + id + "?fields=duration,title",
method: "GET",
dataType: "jsonp",
timeout: 1000
};
urlRetrieve(https, options, function (status, data) {
if(status !== 200) {
callback(true, null);
return;
}
try {
data = JSON.parse(data);
var title = data.title;
var seconds = data.duration;
var media = new Media(id, title, seconds, "dm");
callback(false, media);
} catch(e) {
// TODO See what kinds of errors DM returns
var err = true;
callback(err, null);
}
});
},
/* soundcloud.com */
sc: function (id, callback) {
// Soundcloud's API is badly designed and badly documented
// In order to lookup track data from a URL, I have to first
// make a call to /resolve to get the track id, then make a second
// call to /tracks/{track.id} to actally get useful data
// This is a waste of bandwidth and a pain in the ass
const SC_CLIENT = "2e0c82ab5a020f3a7509318146128abd";
var options = {
host: "api.soundcloud.com",
port: 443,
path: "/resolve.json?url=" + id + "&client_id=" + SC_CLIENT,
method: "GET",
dataType: "jsonp",
timeout: 1000
};
urlRetrieve(https, options, function (status, data) {
if(status !== 302) {
callback(true, null);
return;
}
var track = null;
try {
data = JSON.parse(data);
track = data.location;
} catch(e) {
callback(true, null);
return;
}
var options2 = {
host: "api.soundcloud.com",
port: 443,
path: track,
method: "GET",
dataType: "jsonp",
timeout: 1000
};
// I want to get off async's wild ride
urlRetrieve(https, options2, function (status, data) {
if(status !== 200) {
callback(true, null);
return;
}
try {
data = JSON.parse(data);
// Duration is in ms, but I want s
var seconds = data.duration / 1000;
var title = data.title;
var media = new Media(id, title, seconds, "sc");
callback(false, media);
} catch(e) {
callback(true, null);
}
});
});
},
/* livestream.com */
li: function (id, callback) {
var title = "Livestream.com - " + id;
var media = new Media(id, title, "--:--", "li");
callback(false, media);
},
/* twitch.tv */
tw: function (id, callback) {
var title = "Twitch.tv - " + id;
var media = new Media(id, title, "--:--", "tw");
callback(false, media);
},
/* justin.tv */
jt: function (id, callback) {
var title = "Justin.tv - " + id;
var media = new Media(id, title, "--:--", "jt");
callback(false, media);
},
/* ustream.tv */
us: function (id, callback) {
var options = {
host: "www.ustream.tv",
port: 80,
path: "/" + id,
method: "GET",
timeout: 1000
};
urlRetrieve(http, options, function (status, data) {
if(status !== 200) {
callback(true, null);
return;
}
// Regexing the ID out of the HTML because
// Ustream's API is so horribly documented
// I literally could not figure out how to retrieve
// this information.
//
// [](/eatadick)
var m = data.match(/cid":([0-9]+)/);
if(m) {
var title = "Ustream.tv - " + id;
var media = new Media(m[1], title, "--:--", "us");
callback(false, media);
}
callback(true, null);
});
},
/* JWPlayer */
jw: function (id, callback) {
var title = "JWPlayer - " + id;
var media = new Media(id, title, "--:--", "jw");
callback(false, media);
}
}
exports.Getters = Getters;
// Look up YouTube metadata
// Fairly straightforward
exports.getYTInfo = function(id, callback) {