xf实时语音转写

pom.xml

@@ -353,6 +353,24 @@
 			<version>3.4.0</version>
 		</dependency>
 
+		<!-- 讯飞-星火大模型相关	-->
+		<dependency>
+			<groupId>cn.xfyun</groupId>
+			<artifactId>websdk-java-spark</artifactId>
+			<version>2.1.3</version>
+		</dependency>
+		<!-- 讯飞-语音相关能力	-->
+		<dependency>
+			<groupId>cn.xfyun</groupId>
+			<artifactId>websdk-java-speech</artifactId>
+			<version>3.0.3</version>
+		</dependency>
+		<!-- 讯飞-自然语言处理	-->
+		<dependency>
+			<groupId>cn.xfyun</groupId>
+			<artifactId>websdk-java-nlp</artifactId>
+			<version>2.0.7</version>
+		</dependency>
 
 	</dependencies>
 

src/main/java/com/newyetai/ai/controller/AiController.java

@@ -1,6 +1,7 @@
 package com.newyetai.ai.controller;
 
 import com.newyetai.common.AlitaResult;
+import com.newyetai.sdk.xf.dto.RtasrReqDTO;
 import com.newyetai.sdk.xf.model.ChatReq;
 import com.newyetai.sdk.xf.model.TranslateReq;
 import com.newyetai.sdk.xf.util.XfUtils;
@@ -25,9 +26,15 @@ public class AiController {
 
   @PostMapping("translate")
   @ApiOperation("ai翻译")
-  public AlitaResult<String> chat(@Validated @RequestBody TranslateReq req) {
+  public AlitaResult<String> translate(@Validated @RequestBody TranslateReq req) {
     return AlitaResult.success(XfUtils.translate(req));
   }
 
+  @PostMapping("rtasr")
+  @ApiOperation("ai实时语音转写")
+  public AlitaResult<String> rtasr(@Validated @RequestBody RtasrReqDTO req) {
+    return AlitaResult.success(XfUtils.rtasr(req));
+  }
+
 
 }

src/main/java/com/newyetai/sdk/xf/config/XfProperties.java

@@ -36,6 +36,16 @@ public class XfProperties implements Serializable{
   private String ttsHostUrl;
   /** 机器翻译niutrans **/
   private String otsHostUrl;
+  /** 实时语音转写 **/
+  private XfAppProperties rtasr;
+
+  @Data
+  @NoArgsConstructor
+  @AllArgsConstructor
+  @Builder
+  public static class XfAppProperties implements Serializable{
+    private String apiKey;
+  }
 
 
 }

src/main/java/com/newyetai/sdk/xf/dto/RtasrReqDTO.java

@@ -0,0 +1,24 @@
+package com.newyetai.sdk.xf.dto;
+
+import java.io.Serializable;
+import javax.validation.constraints.NotBlank;
+import lombok.AllArgsConstructor;
+import lombok.Builder;
+import lombok.Builder.Default;
+import lombok.Data;
+import lombok.NoArgsConstructor;
+
+@Data
+@AllArgsConstructor
+@NoArgsConstructor
+@Builder
+public class RtasrReqDTO implements Serializable {
+
+  @NotBlank(message = "语音文件不能为空")
+  private String fileCode;
+  @Default
+  private String from = "cn_uyghur";
+
+
+
+}

src/main/java/com/newyetai/sdk/xf/model/ChatReq.java

@@ -22,7 +22,18 @@ public class ChatReq implements Serializable{
   @Singular
   @NotNull(message = "消息内容不能为空")
   private List<MessageDTO> messages;
-  
+  @Default
+  private String temperature = "0.3";
+  @Default
+  private String max_tokens = "150";
+  @Default
+  private String repetition_penalty = "1.25";
+  @Default
+  private String top_p = "0.65";
+  @Default
+  private String presence_penalty = "-0.6";
+  @Default
+  private String frequency_penalty = "0.2";
 
 
 

src/main/java/com/newyetai/sdk/xf/service/RtasrClientApp.java

@@ -0,0 +1,155 @@
+package com.newyetai.sdk.xf.service;
+
+import cn.hutool.json.JSON;
+import cn.hutool.json.JSONArray;
+import cn.hutool.json.JSONObject;
+import cn.hutool.json.JSONUtil;
+import cn.xfyun.api.RtasrClient;
+import cn.xfyun.model.response.rtasr.RtasrResponse;
+import cn.xfyun.service.rta.AbstractRtasrWebSocketListener;
+import com.newyetai.common.AlitaResult;
+import com.newyetai.common.util.JsonUtils;
+import com.newyetai.sdk.xf.config.XfProperties;
+import java.io.InputStream;
+import java.util.concurrent.CountDownLatch;
+import okhttp3.Response;
+import okhttp3.WebSocket;
+import org.apache.commons.codec.binary.StringUtils;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+/**
+ * ( Real-time ASR ) 实时语音转写 1、APPID、APIKey信息获取：<a
+ * href="https://console.xfyun.cn/services/rta">...</a> 2、文档地址：<a
+ * href="https://www.xfyun.cn/doc/asr/rtasr/API.html">...</a>
+ *
+ * @author kaili23
+ */
+public class RtasrClientApp extends AbstractRtasrWebSocketListener{
+
+  private static final Logger logger = LoggerFactory.getLogger(RtasrClientApp.class);
+
+  /**
+   * 记录完整结果
+   */
+  private StringBuffer finalResult = new StringBuffer();
+
+  private AlitaResult<String> result;
+
+  /**
+   * 线程同步
+   */
+  private CountDownLatch latch;
+
+  RtasrClientApp(CountDownLatch latch){
+    this.latch = latch;
+  }
+
+  public AlitaResult<String> getResult() {
+    return result;
+  }
+
+  @Override
+  public void onSuccess(WebSocket webSocket, String text) {
+    RtasrResponse response = JsonUtils.toObject(text, RtasrResponse.class);
+    // finalResult为服务端已转写的完整语句，后面拼接服务端本次返回的当前语句的转写中间结果。
+    String tempResult = handleAndReturnContent(response.getData());
+    logger.info("实时转写结果：{}", finalResult + tempResult);
+    result = AlitaResult.success(finalResult.toString());
+  }
+
+  @Override
+  public void onFail(WebSocket webSocket, Throwable t, Response response) {
+    result = AlitaResult.bizError(t.getMessage());
+    latch.countDown();
+    System.exit(0);
+  }
+
+  @Override
+  public void onBusinessFail(WebSocket webSocket, String text) {
+    logger.error("业务异常，返回信息为：{}", text);
+    result = AlitaResult.bizError(text);
+    latch.countDown();
+    System.exit(0);
+  }
+
+  @Override
+  public void onClosing(WebSocket webSocket, int code, String reason) {
+    latch.countDown();
+    System.exit(0);
+  }
+
+  @Override
+  public void onClosed() {
+    latch.countDown();
+    System.exit(0);
+  }
+
+  public static AlitaResult<String> execute(InputStream is, String from, XfProperties xfProperties) {
+    try {
+      CountDownLatch latch = new CountDownLatch(1);
+      RtasrClientApp app = new RtasrClientApp(latch);
+      RtasrClient client = new RtasrClient.Builder()
+          .signature(xfProperties.getAppId(), xfProperties.getRtasr().getApiKey())
+          .lang(from)
+          // 设置此参数需需开通翻译功能，请在控制台申请并开通
+//          .targetLang(to)
+          .build();
+      client.send(is, app);
+      latch.await();
+      return app.getResult();
+    } catch (Exception e) {
+      if (e instanceof InterruptedException) {
+        Thread.currentThread().interrupt();
+      }
+      logger.error("语音转写失败", e);
+      return AlitaResult.error("语音转写失败：{}", e.getMessage());
+    }
+  }
+
+  /**
+   * 解析转写流式响应，更新实时结果，返回此次的中间结果
+   */
+  public String handleAndReturnContent(String message) {
+    StringBuilder tempResult = new StringBuilder();
+
+    try {
+      // 解析本次服务端返回的文本内容
+
+      JSONObject messageObj = JSONUtil.parseObj(message);
+      JSONObject cn = messageObj.getJSONObject("cn");
+      JSONObject st = cn.getJSONObject("st");
+      JSONArray rtArr = st.getJSONArray("rt");
+      for (int i = 0; i < rtArr.size(); i++) {
+        JSONObject rtArrObj = rtArr.getJSONObject(i);
+        JSONArray wsArr = rtArrObj.getJSONArray("ws");
+        for (int j = 0; j < wsArr.size(); j++) {
+          JSONObject wsArrObj = wsArr.getJSONObject(j);
+          JSONArray cwArr = wsArrObj.getJSONArray("cw");
+          for (int k = 0; k < cwArr.size(); k++) {
+            JSONObject cwArrObj = cwArr.getJSONObject(k);
+            String wStr = cwArrObj.getStr("w");
+            tempResult.append(wStr);
+          }
+        }
+      }
+
+      // 根据中间结果类型更新最终结果
+      String type = st.getStr("type");
+      if (StringUtils.equals("1", type)) {
+        // 此时服务端返回的是当前语句的实时转写内容，不保存到最终结果中，返回到调用处进行拼接展示。
+        return tempResult.toString();
+      } else if (StringUtils.equals("0", type)) {
+        // 此时服务端返回的是当前语句的完整转写内容，保存到最终结果中，并返回空字符串。
+        finalResult.append(tempResult);
+        return "";
+      } else {
+        logger.warn("未知的转写响应类型：{}", type);
+        return tempResult.toString();
+      }
+    } catch (Exception e) {
+      return message;
+    }
+  }
+
+}

src/main/java/com/newyetai/sdk/xf/util/XfUtils.java

@@ -1,6 +1,8 @@
 package com.newyetai.sdk.xf.util;
 
+import com.newyetai.sdk.xf.dto.RtasrReqDTO;
 import com.newyetai.sdk.xf.model.TranslateReq;
+import com.newyetai.sdk.xf.service.RtasrClientApp;
 import java.io.File;
 import java.io.IOException;
 import java.io.InputStream;
@@ -99,7 +101,7 @@ public class XfUtils {
   
   /**
    * 语音听写
-   * @param filePath
+   * @param fileCode
    * @return
    * @author sr on 2025-01-09 
    */
@@ -117,7 +119,6 @@ public class XfUtils {
   
   /**
    * 语音合成
-   * @param filePath
    * @return
    * @author sr on 2025-01-09 
    */
@@ -230,16 +231,42 @@ public class XfUtils {
     return result.getData();
   }
 
+  /**
+   * 对话
+   * @param req
+   * @return
+   */
   public static String chat(ChatReq req) {
     AlitaResult<String> result = XfSparkProvider.httpChat(req);
     Assert.isTrue(result.getSuccess(), result.getMsg());
     return result.getData();
   }
 
+  /**
+   * 翻译
+   * @param req
+   * @return
+   */
   public static String translate(TranslateReq req) {
     AlitaResult<String> result = XfSparkProvider.translate(req);
     Assert.isTrue(result.getSuccess(), result.getMsg());
     return result.getData();
   }
 
+  /**
+   * 实时语音转写
+   * @param reqDTO
+   * @return
+   */
+  public static String rtasr(RtasrReqDTO reqDTO) {
+    try (InputStream is = fileRecordService.getFile(reqDTO.getFileCode())){
+      AlitaResult<String> result =  RtasrClientApp.execute(is, reqDTO.getFrom(), xfProperties);
+      Assert.isTrue(result.getSuccess(), result.getMsg());
+      return result.getData();
+    } catch (Exception e) {
+      throw new BusinessException("语音听写异常:" + e.getMessage());
+    }
+
+  }
+
 }

src/main/resources/application.properties

@@ -165,6 +165,8 @@ xf.iatHostUrl=https://iat-api.xfyun.cn/v2/iat
 xf.ttsHostUrl=https://tts-api.xfyun.cn/v2/tts
 #机器翻译niutrans
 xf.otsHostUrl=https://ntrans.xfyun.cn/v2/ots
+#实时语音转写
+xf.rtasr.APIKey=2e33d9128a73be21b294594e5377ca53
 
 #腾讯地图
 tencent.map.key=TS7BZ-GOVCB-CUZUL-JUUAV-WTYT3-Z2B4Z

src/test/java/com/newyetai/SparkTest.java

@@ -1,5 +1,8 @@
 package com.newyetai;
 
+import com.newyetai.sdk.xf.config.XfProperties.XfAppProperties;
+import com.newyetai.sdk.xf.model.TranslateReq;
+import com.newyetai.sdk.xf.service.RtasrClientApp;
 import java.io.File;
 import java.io.FileInputStream;
 import java.io.FileNotFoundException;
@@ -47,9 +50,11 @@ public class SparkTest {
   static XfProperties xfProperties = XfProperties.builder()
       .iatHostUrl("https://iat-api.xfyun.cn/v2/iat")
       .ttsHostUrl("https://tts-api.xfyun.cn/v2/tts")
-      .appId("a2847f13")
-      .apiSecret("")
-      .apiKey("")
+      .otsHostUrl("https://ntrans.xfyun.cn/v2/ots")
+      .appId("83b8ddca")
+      .apiSecret("ZWE2MjkzY2VmZDkwM2EwZGJhNGRhOWJm")
+      .apiKey("e092f93e28ebe6d27f756bb2d884dcff")
+      .rtasr(XfAppProperties.builder().apiKey("2e33d9128a73be21b294594e5377ca53").build())
       .build();
   
   @Test
@@ -218,4 +223,20 @@ public class SparkTest {
     }
   }
 
+  @Test
+  public void ots(){
+//    AlitaResult<?> result = XfSparkProvider.translate(TranslateReq.builder().text("欢迎#你啊,.;@#$你好").build(), xfProperties);
+    AlitaResult<?> result = XfSparkProvider.translate(TranslateReq.builder().text("تىيانەي ئەقىللىق خىزمەت مۇلازىمىتى بىرلەشمە سۇپىسىغا خوش ھەمرىيىڭىز.").from("uy").to("cn").build(), xfProperties);
+    log.info("{}", result);
+  }
+
+  @Test
+  public void rtasr() throws IOException {
+    String file = "src/main/resources/iat/ug.pcm";
+    try (InputStream is = new FileInputStream(file)) {
+      AlitaResult<String> result = RtasrClientApp.execute(is, "cn_uyghur",  xfProperties);
+      log.info("{}", result);
+    }
+  }
+
 }