Started working on the custom speech api

lib/watson.js

@@ -7,55 +7,55 @@
 // !Dependencies
 var fs = require('fs'),
 	request = require('request'),
-	merge = require('merge-descriptors');
+	merge = require('merge-descriptors'),
+	path = require('path'),
+	readStreamBuffer = require('stream-buffers').ReadableStreamBuffer;
 
 // !Set up the client options and defaults
 var Watson = exports.Watson = function(options) {
-    this.client_id = options.client_id || null ;
-    this.client_secret = options.client_secret || null;
-    this.access_token = options.access_token || null;
-    this.scope = options.scope || "SPEECH";
-    this.context = options.context || "Generic";
-    this.xarg = options.xarg || "";
-    this.access_token_url = options.access_token_url || "https://api.att.com/oauth/token"
-    this.api_domain = options.api_domain || "api.att.com";
+	this.client_id = options.client_id || null ;
+	this.client_secret = options.client_secret || null;
+	this.access_token = options.access_token || null;
+	this.scope = options.scope || 'SPEECH';
+	this.access_token_url = options.access_token_url || 'https://api.att.com/oauth/token'
+	this.api_domain = options.api_domain || 'api.att.com';
 };
 
 // Get Access Token Method
 Watson.prototype.getAccessToken = function(callback) {
 	// Set the params for the OAuth 2.0 Request
-    var request_params = {
+	var request_params = {
 		client_id: this.client_id,
 		client_secret: this.client_secret,
-		grant_type: "client_credentials",
+		grant_type: 'client_credentials',
 		scope: this.scope
-    };
-    
-    // Create the Param String
-    var paramlist  = [];
-    for (pk in request_params) {
-		paramlist.push(pk + "=" + request_params[pk]);
-    };
-    var body_string = paramlist.join("&");
-    
-    // !Details of the OAuth 2.0 Request
-    var request_details = {  
-		method: "POST",
+	};
+	
+	// Create the Param String
+	var paramlist  = [];
+	for (pk in request_params) {
+		paramlist.push(pk + '=' + request_params[pk]);
+	};
+	var body_string = paramlist.join('&');
+	
+	// !Details of the OAuth 2.0 Request
+	var request_details = {  
+		method: 'POST',
 		headers: {'content-type' : 'application/x-www-form-urlencoded'},
 		uri: this.access_token_url,
 		body: body_string
-    };
-    
-    // !Make the request
-    request(request_details, function(error, response, body) {
-    	if(error) {
+	};
+	
+	// !Make the request
+	request(request_details, function(error, response, body) {
+		if(error) {
 			error = new Error('Failed to get access token.');
-    	}
+		}
 
-    	this.access_token = JSON.parse(body)['access_token'];
+		this.access_token = JSON.parse(body)['access_token'];
 
-    	// Pass any errors and the Access Token back to the app
-    	callback(error, this.access_token );
+		// Pass any errors and the Access Token back to the app
+		callback(error, this.access_token );
 	});
 };
 
@@ -66,8 +66,7 @@ Watson.prototype.speechToText = function(speechFile, reqParams, callback) {
 		'Accept': 'application/json',
 		'Authorization': 'Bearer ' + this.access_token,
 		'Content-Type': 'audio/wav',
-		'X-SpeechContext': this.context,			// Possible Values: Generic, UVerseEPG, BusinessSearch, Websearch, SMS, Voicemail, QuestionAndAnswer
-		'X-Arg': this.xarg							// Occasionally used for custom grammar sets. Unnecessary by default.
+		'X-SpeechContext': 'Generic'			// Possible Values: Generic, UVerseEPG, BusinessSearch, Websearch, SMS, Voicemail, QuestionAndAnswer
 	};
 
 	// get a call back
@@ -89,16 +88,32 @@ Watson.prototype.speechToText = function(speechFile, reqParams, callback) {
 	var request_details = {
 		method: 'POST',
 		headers: reqHeaders,
-		uri: "https://" + this.api_domain + "/speech/v3/speechToText"
+		uri: 'https://' + this.api_domain + '/speech/v3/speechToText'
 	};
 
+	var stream = null,
+		dataIsBuffer = Buffer.isBuffer( speechFile );
+
+	if( dataIsBuffer ) {
+
+		stream = new readStreamBuffer({
+			frequency: 1,  // in milliseconds.
+			chunkSize: 2048 // in bytes.
+		});
+	} else {
+
+		stream = fs.createReadStream(speechFile);
+	}
+
 	// !Pipe the Speech file from the Node.js server to the AT&T API server
-	fs.createReadStream(speechFile).pipe(request(
+	stream.pipe(request(
 				request_details,
 				function(error, response, body) {
 					// Try to read the `body` as JSON
 					try {
 						var parsed = JSON.parse(body);
+
+						console.log( 'START END', Date.now() );
 					} catch(e) {
 					// If it fails, it's an invalid API Reply
 						error = new Error('API Reply is not a valid JSON string.');
@@ -110,4 +125,78 @@ Watson.prototype.speechToText = function(speechFile, reqParams, callback) {
 				}
 		)
 	);
+
+	if( dataIsBuffer ) {
+
+		stream.put( speechFile );
+
+		console.log( 'START SEND', Date.now() );
+	}
+};
+
+Watson.prototype.speechToTextCustom = function(speechFile, grammarFile, reqParams, callback) {
+
+	var reqHeaders = {
+		'Accept': 'application/json',
+		'Authorization': 'Bearer ' + this.access_token,
+		'Content-Type': 'multipart/x-srgs-audio',
+		'X-SpeechContext': 'GrammarList'			// Possible Values: GenericHints, GrammarList
+	};
+
+	// get a call back
+	callback = arguments[ arguments.length - 1 ];
+
+	// make sure we have an access_token
+	if( this.access_token === null )
+		throw new Error( 'Call the getAccessToken function or pass in "access_token" in the constructor options before calling speechToText' );
+
+	// ensure that we have a callback. This function is sort of pointless without a callback
+	if( typeof callback != 'function' )
+		throw new Error( 'The last argument you pass in must be a callback function' );
+
+	// add the params to the request headers
+	// possible parameter values can be seen here under request parameters: http://developer.att.com/apis/speech/docs/v3
+	merge( reqHeaders, reqParams );
+
+	// !Details of the API Call Request
+	var request_details = {
+		method: 'POST',
+		headers: reqHeaders,
+		uri: 'https://' + this.api_domain + '/speech/v3/speechToTextCustom',
+		multipart: [
+			{
+				'content-disposition': 'form-data;name="x-grammar";filename="example.srgs"',
+				'content-type': 'application/srgs+xml',
+				body: fs.readFileSync( grammarFile, { encoding: 'utf8' } )
+			},
+			{ 
+				'content-disposition': 'form-data; name="x-voice"; filename="' + path.basename( speechFile ) + '"',
+				'content-type': 'audio/wav',
+				body: fs.readFileSync( speechFile )
+			}
+		]
+	};
+
+	// !Pipe the Speech file from the Node.js server to the AT&T API server
+	request(
+		request_details,
+		function(error, response, body) {
+			// Try to read the `body` as JSON
+			try {
+				var parsed = JSON.parse(body);
+			} catch(e) {
+			// If it fails, it's an invalid API Reply
+				error = new Error('API Reply is not a valid JSON string.');
+				error.reply = body;
+			} finally {
+			// Toss the error and/or parsed response back to the app
+				callback(error, parsed);
+			}
+		}
+	);
 };
+
+function getStreamFromFile() {
+
+
+}

package.json

@@ -11,7 +11,8 @@
   "dependencies": {
     "oauth": ">= 0.9.8",
     "request": ">= 1.9.0",
-    "merge-descriptors": "0.0.2"
+    "merge-descriptors": "0.0.2",
+    "stream-buffers": "~0.2.5"
   },
   "engines": {
     "node": ">= 0.4.0"