add hotwords for h5 and java (#876)

* add hotwords for h5 and java

* add timestamp for h5 and java

funasr/runtime/html5/static/index.html

@@ -52,7 +52,11 @@
 				</div>
 				<br>
 		
+				热词设置(空格隔开,如"阿里巴巴 达摩院 阿里云")：
+				<br>
 				
+				<input  type="text"  id="varHot"  style=" width: 100%;height:100%" /> 
+				<br>
 			 
 				语音识别结果显示：
 				<br>

funasr/runtime/html5/static/main.js

@@ -172,6 +172,17 @@ function on_recoder_mode_change()
 	 
 			}
 }
+function getHotwords(){
+  var obj = document.getElementById("varHot");
+  
+  if(typeof(obj) == 'undefined' || obj==null || obj.value.length<=0){
+	return "";
+  }
+  let val = obj.value.toString();
+  console.log("hotwords="+val);
+  return val;
+
+}
 function getAsrMode(){
 
             var item = null;
@@ -193,7 +204,34 @@ function getAsrMode(){
 		   return item;
 }
 		   
+function handleWithTimestamp(tmptext,tmptime)
+{
+	console.log( "tmptext: " + tmptext);
+	console.log( "tmptime: " + tmptime);
+    if(tmptime==null || tmptime=="undefined" || tmptext.length<=0)
+	{
+		return tmptext;
+	}
+	tmptext=tmptext.replace(/。/g, ","); // in case there are a lot of "。"
+	var words=tmptext.split(",");
+	var jsontime=JSON.parse(tmptime.replace(/\]\]\[\[/g, "],[")); // in case there are a lot segments by VAD
+	var char_index=0;
+	var text_withtime="";
+	for(var i=0;i<words.length;i++)
+	{   
+	if(words[i]=="undefined"  || words[i].length<=0)
+	{
+		continue;
+	}
+        console.log("words===",words[i]);
+		console.log( "words: " + words[i]+",time="+jsontime[char_index][0]/1000);
+		text_withtime=text_withtime+jsontime[char_index][0]/1000+":"+words[i]+"\n";
+		char_index=char_index+words[i].length;
+	}
+	return text_withtime;
+	
 
+}
 // 语音识别结果; 对jsonMsg数据解析,将识别结果附加到编辑框中
 function getJsonMessage( jsonMsg ) {
 	//console.log(jsonMsg);
@@ -201,9 +239,11 @@ function getJsonMessage( jsonMsg ) {
 	var rectxt=""+JSON.parse(jsonMsg.data)['text'];
 	var asrmodel=JSON.parse(jsonMsg.data)['mode'];
 	var is_final=JSON.parse(jsonMsg.data)['is_final'];
-	if(asrmodel=="2pass-offline")
+	var timestamp=JSON.parse(jsonMsg.data)['timestamp'];
+	if(asrmodel=="2pass-offline" || asrmodel=="offline")
 	{
-		offline_text=offline_text+rectxt; //.replace(/ +/g,"");
+		
+		offline_text=offline_text+handleWithTimestamp(rectxt,timestamp); //rectxt; //.replace(/ +/g,"");
 		rec_text=offline_text;
 	}
 	else

funasr/runtime/html5/static/wsconnecter.js

@@ -78,6 +78,11 @@ function WebSocketConnectMethod( config ) { //定义socket连接方法类
 		{
 			request.wav_format=file_ext;
 		}
+		var hotwords=getHotwords();
+		if(hotwords.length>0)
+		{
+			request.hotwords=hotwords;
+		}
 		console.log(request);
 		speechSokt.send( JSON.stringify(request) );
 		console.log("连接成功");

funasr/runtime/java/FunasrWsClient.java

@@ -79,6 +79,10 @@ public class FunasrWsClient extends WebSocketClient {
       obj.put("chunk_size", array);
       obj.put("chunk_interval", new Integer(chunkInterval));
       obj.put("wav_name", wavName);
+	  if(FunasrWsClient.hotwords.trim().length()>0)
+	  {
+		  obj.put("hotwords", FunasrWsClient.hotwords.trim());
+	  }
 	  if(suffix.equals("wav")){
 	      suffix="pcm";
 	  }
@@ -185,6 +189,10 @@ public class FunasrWsClient extends WebSocketClient {
     try {
       jsonObject = (JSONObject) jsonParser.parse(message);
       logger.info("text: " + jsonObject.get("text"));
+	  if(jsonObject.containsKey("timestamp"))
+	  {
+		  logger.info("timestamp: " + jsonObject.get("timestamp"));
+	  }
     } catch (org.json.simple.parser.ParseException e) {
       e.printStackTrace();
     }
@@ -222,6 +230,7 @@ public class FunasrWsClient extends WebSocketClient {
   static String strChunkSize = "5,10,5";
   static int chunkInterval = 10;
   static int sendChunkSize = 1920;
+  static String hotwords="";
 
   String wavName = "javatest";
 
@@ -270,6 +279,12 @@ public class FunasrWsClient extends WebSocketClient {
         .setDefault("offline")
         .type(String.class)
         .required(false);
+    parser
+        .addArgument("--hotwords")
+        .help("hotwords, splited by space")
+        .setDefault("")
+        .type(String.class)
+        .required(false);
     String srvIp = "";
     String srvPort = "";
     String wavPath = "";
@@ -277,7 +292,7 @@ public class FunasrWsClient extends WebSocketClient {
     String chunk_size = "";
     int chunk_interval = 10;
     String strmode = "offline";
-
+    String hot="";
     try {
       Namespace ns = parser.parseArgs(args);
       srvIp = ns.get("host");
@@ -287,6 +302,7 @@ public class FunasrWsClient extends WebSocketClient {
       chunk_size = ns.get("chunk_size");
       chunk_interval = ns.get("chunk_interval");
       strmode = ns.get("mode");
+	  hot=ns.get("hotwords");
       System.out.println(srvPort);
 
     } catch (ArgumentParserException ex) {
@@ -298,6 +314,7 @@ public class FunasrWsClient extends WebSocketClient {
     FunasrWsClient.chunkInterval = chunk_interval;
     FunasrWsClient.wavPath = wavPath;
     FunasrWsClient.mode = strmode;
+	FunasrWsClient.hotwords=hot;
     System.out.println(
         "serIp="
             + srvIp