当前位置： 首页 > article >正文

Netty对接阿里云语音识别和录音识别

article 2024/9/22 6:30:58

阿里云实时语音识别：https://help.aliyun.com/document_detail/84430.html?spm=a2c4g.324262.0.0.564f73e9O6yq25

阿里云录音识别：https://help.aliyun.com/document_detail/90727.html?spm=a2c4g.90726.0.0.662d73e9qr8DqE

语音识别的流程为：前端和后端构建websocket连接，然后传二进制音频流给后端，后端拿到音频流，后阿里云构建websocket连接，转发音频流，阿里云收到后进行翻译，再返回给后端，后端再返回给前端

录音识别流程为：前端上传一段录音到阿里云oss上，返回录音的url，然后调用阿里云的录音识别拿到录音并解析，将结果返回给后端，后端再将结果返回给前端

实现

pom.xml

<dependencies>
    <!-- https://mvnrepository.com/artifact/io.netty/netty-all -->
    <dependency>
        <groupId>io.netty</groupId>
        <artifactId>netty-all</artifactId>
        <version>4.1.76.Final</version>
    </dependency>
    <dependency>
        <groupId>cn.hutool</groupId>
        <artifactId>hutool-all</artifactId>
        <version>5.8.18</version>
    </dependency>

<!--        <dependency>-->
<!--            <groupId>com.aliyun</groupId>-->
<!--            <artifactId>aliyun-java-sdk-core</artifactId>-->
<!--            <version>3.7.1</version>-->
<!--        </dependency>-->
    <!-- https://mvnrepository.com/artifact/com.alibaba/fastjson -->
    <dependency>
        <groupId>com.alibaba</groupId>
        <artifactId>fastjson</artifactId>
        <version>1.2.72</version>
    </dependency>

    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-web</artifactId>
    </dependency>

    <dependency>
        <groupId>org.projectlombok</groupId>
        <artifactId>lombok</artifactId>
        <optional>true</optional>
    </dependency>
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-test</artifactId>
        <scope>test</scope>
    </dependency>
    <!--阿里云语音识别-->
    <dependency>
        <groupId>com.alibaba.nls</groupId>
        <artifactId>nls-sdk-transcriber</artifactId>
        <version>2.2.1</version>
    </dependency>

    <!--swagger3-->
    <dependency>
        <groupId>io.springfox</groupId>
        <artifactId>springfox-boot-starter</artifactId>
        <version>3.0.0</version>
    </dependency>

    <!-- 阿里云OSS -->
    <dependency>
        <groupId>com.aliyun.oss</groupId>
        <artifactId>aliyun-sdk-oss</artifactId>
        <version>3.15.1</version>
    </dependency>
    <dependency>
        <groupId>org.apache.commons</groupId>
        <artifactId>commons-lang3</artifactId>
        <version>3.12.0</version>
    </dependency>

    <!-- guava -->
    <dependency>
        <groupId>com.google.guava</groupId>
        <artifactId>guava</artifactId>
        <version>31.0.1-jre</version>
    </dependency>


<!--        <dependency>-->
<!--            <groupId>org.springblade</groupId>-->
<!--            <artifactId>blade-core-boot</artifactId>-->
<!--            <version>3.0.1.RELEASE</version>-->
<!--            <exclusions>-->
<!--                <exclusion>-->
<!--                    <groupId>org.springblade</groupId>-->
<!--                    <artifactId>blade-core-cloud</artifactId>-->
<!--                </exclusion>-->
<!--            </exclusions>-->
<!--        </dependency>-->

    <!--redis客户端-->
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-data-redis</artifactId>
        <exclusions>
            <exclusion>
                <groupId>io.lettuce</groupId>
                <artifactId>lettuce-core</artifactId>
            </exclusion>
        </exclusions>
    </dependency>
    <!-- lettuce连接池 -->
    <dependency>
        <groupId>io.lettuce</groupId>
        <artifactId>lettuce-core</artifactId>
        <version>6.1.6.RELEASE</version>
    </dependency>
    <dependency>
        <groupId>org.apache.commons</groupId>
        <artifactId>commons-pool2</artifactId>
    </dependency>

</dependencies>

ApplicationService 启动时构建阿里云的nls连接，并获取token放入redis中

@Slf4j
@Service
public class ApplicationService implements ApplicationListener<ContextRefreshedEvent> {
   

    @Autowired
    RedisTemplate<String, String> redisTemplate;

    public static ApplicationService application = null;

    @Value("${aliyun.nls.accessKeyId}")
    private String accessKeyId;
    @Value("${aliyun.nls.accessKeySecret}")
    private String accessKeySecret;
    @Value("${aliyun.nls.url}")
    private String url;

    @Override
    public void onApplicationEvent(ContextRefreshedEvent contextRefreshedEvent) {
   
        //启动客户端
        if (contextRefreshedEvent.getApplicationContext().getParent() == null) {
   
            log.info("客户端启动-------------------------->");
            synchronized (this) {
   
                ApplicationService.application = this;
                new NlsClientService(accessKeyId, accessKeySecret, url);
                log.info("阿里云 NlsClient 初始化完毕");
                AccessToken accessToken = NlsClientService.getAccessToken();
                redisTemplate.opsForValue().set("nlp:token", accessToken.getToken(), accessToken.getExpireTime(), TimeUnit.SECONDS);
            }
        }
    }
}

构建nls和获取token的具体实现

@Slf4j
public class NlsClientService {
   

    private static NlsClient client;

    private static AccessToken accessToken;

    public NlsClientService(String accessKeyId, String accessKeySecret, String url) {
   

        //创建NlsClient实例,应用全局创建一个即可,生命周期可和整个应用保持一致,默认服务地址为阿里云线上服务地址
        applyAccessToken(accessKeyId, accessKeySecret);
        if (url.isEmpty()) {
   
            client = new NlsClient(accessToken.getToken());
        } else {
   
            client = new NlsClient(url, accessToken.getToken());
        }
    }

    public static AccessToken getAccessToken() {
   
        return accessToken;
    }

    public static void applyAccessToken(String accessKeyId, String accessKeySecret) {
   
        accessToken = new AccessToken(accessKeyId, accessKeySecret);
        try {
   
            accessToken.apply();
            log.info("get token: " + accessToken.getToken() + ", expire time: " + accessToken.getExpireTime());
        } catch (IOException e) {
   
            log.error("https获取accessToken失败！" + e.getMessage());
        }
    }

    public static NlsClient getNlsClient() {
   
        return client;
    }
}

netty服务端

@Slf4j
@Configuration
public class NettyWebSocketServer {
   
    public static final int WEB_SOCKET_PORT = 9000;
    // 创建线程池执行器
    private EventLoopGroup bossGroup = new NioEventLoopGroup(1);
    private EventLoopGroup workerGroup = new NioEventLoopGroup(NettyRuntime.availableProcessors());

    /**
     * 启动 ws server
     *
     * @return
     * @throws InterruptedException
     */
    @PostConstruct
    public void start() throws InterruptedException {
   
        run();
    }

    /**
     * 销毁
     */
    @PreDestroy
    public void destroy() {
   
        Future<?> future = bossGroup.shutdownGracefully();
        Future<?> future1 = workerGroup.shutdownGracefully();
        future.syncUninterruptibly();
        future1.syncUninterruptibly();
        log.info("关闭 ws server 成功");
    }

    public void run() throws InterruptedException {
   
        // 服务器启动引导对象
        ServerBootstrap serverBootstrap = new ServerBootstrap();
        serverBootstrap.group(bossGroup, workerGroup)
                .channel(NioServerSocketChannel.class)
                .option(ChannelOption.SO_BACKLOG, 128)
                .option(ChannelOption.SO_KEEPALIVE, true)
                .handler(new LoggingHandler(LogLevel.INFO)) // 为 bossGroup 添加 日志处理器
                .childHandler(new ChannelInitializer<SocketChannel>() {
   
                    @Override
                    protected void initChannel(SocketChannel socketChannel) throws Exception {
   
                        ChannelPipeline pipeline = socketChannel.pipeline();
                        //ws升级为wss
                        SslContext sslCtx = SslUtil.createSSLContext();
                        pipeline.addLast(sslCtx.newHandler(socketChannel.alloc()));
                        //10秒客户端没有向服务器发送心跳则关闭连接
                        pipeline.addLast(new IdleStateHandler(10, 10, 0));
                        // 因为使用http协议，所以需要使用http的编码器，解码器
                        pipeline.addLast(new HttpServerCodec());
                        // 以块方式写，添加 chunkedWriter 处理器
                        pipeline.addLast(new ChunkedWriteHandler());
                        /**
                         * 说明：
                         *  1. http数据在传输过程中是分段的，HttpObjectAggregator可以把多个段聚合起来；
                         *  2. 这就是为什么当浏览器发送大量数据时，就会发出多次 http请求的原因
                         */
                        pipeline.addLast(new HttpObjectAggregator(8192));
                        /**
                         * 说明：
                         *  1. 对于 WebSocket，它的数据是以帧frame 的形式传递的；
                         *  2. 可以看到 WebSocketFrame 下面有6个子类
                         *  3. 浏览器发送请求时： ws://localhost:7000/hello 表示请求的uri
                         *  4. WebSocketServerProtocolHandler 核心功能是把 http协议升级为 ws 协议，保持长连接；
                         *      是通过一个状态码 101 来切换的
                         */
                        pipeline.addLast(new WebSocketServerProtocolHandler("/websocket"));
                        // 自定义handler ，处理业务逻辑
                        pipeline.addLast(new NettyWebSocketServerHandler());
                    }
                });
        // 启动服务器，监听端口，阻塞直到启动成功
        serverBootstrap.bind(WEB_SOCKET_PORT).sync();
    }

}

语音识别handler

/**
 * 自定义handler
 */
@Slf4j
@Component
public class NettyWebSocketServerHandler extends SimpleChannelInboundHandler<BinaryWebSocketFrame> {
   

    private NlpService nlpService;

    // 当web客户端连接后，触发该方法
    @Override
    public void handlerAdded(ChannelHandlerContext ctx) throws Exception {
   
        this.nlpService = getService();
    }

    private NlpService getService() {
   
        return SpringUtil.getBean(NlpService.class);
    }

    /**
     * netty断联
     *
     * @param ctx
     * @throws Exception
     */
    @Override
    public void channelInactive(ChannelHandlerContext ctx) throws Exception {
   
        SpeechTranscriber transcriber = (SpeechTranscriber) transcriberMap.get(ctx.channel());
        if (Objects.nonNull(transcriber)) {
   
            //阿里云netty10s断开，连接状态变为STATE_CLOSED
            if (SpeechReqProtocol.State.STATE_CLOSED.toString() != transcriber.getState().toString()) {
   
                transcriber.stop();
            }
            transcriberMap.remove(ctx.channel());
            transcriber.close();
        }
        log.warn("{} 已经断开", ctx.channel());
    }


    /**
     * 处理异常
     *
     * @param ctx
     * @param cause
     * @throws Exception
     */
    @Override
    public void exceptionCaught(ChannelHandlerContext ctx, Throwable cause) throws Exception {
   
        log.warn("{} 已经异常断开 异常是{}", ctx.channel(), cause.getMessage());
        ctx.channel().close();
    }

    /**
     * 心跳检查
     *
     * @param ctx
     * @param evt
     * @throws Exception
     */
    @Override
    public void userEventTriggered(ChannelHandlerContext ctx, Object evt) throws Exception {
   
        if (evt instanceof IdleStateEvent) {
   
            IdleStateEvent idleStateEvent = (IdleStateEvent) evt;
            // 心跳检测超时事件
            if (idleStateEvent.state() == IdleState.READER_IDLE) {
   
                log.warn("{} 已经 10s 没有读到数据了，关闭连接", ctx.channel());
            } else if (idleStateEvent.state() == IdleState.WRITER_IDLE) {
   
                log.warn("{} 已经 10s 没有写出数据了，关闭连接", ctx.channel());
            }
            ctx.channel().close();
        }
    }

    @Override
    protected void channelRead0(ChannelHandlerContext ctx, BinaryWebSocketFrame msg) throws Exception {
   
        if (msg instanceof BinaryWebSocketFrame) {
   
            //读取音频二进制流
            ByteBuf byteBuf = msg.content();
            byte[] byteArray = new byte[byteBuf.readableBytes()];
            byteBuf.readBytes(byteArray);
            byteBuf.release();

            //调用阿里云语音翻译
            nlpService.speechTranslation(ctx.channel(), byteArray);
        }
    }
}

音频流格式，3200B的字节数组

byte[] byteArray = {
   -1, -2, -51, -4, 105, -7, 84, -9, -2, -11, -56, -12, -64, -13, 47, -13, -28, -13, 110, -12, 25, -11, 78, -10, -95, -9, -21, -7, -123, -4, 106, -2, -109, 0, 85, 3, -119, 5, -63, 6, -19, 7, -33, 9, -35, 10, 72, 11, -72, 11, -75, 11, 34, 11, 103, 10, 89, 9, 74, 7, -99, 5, 11, 4, -59, 1, -109, -1, 66, -2, -47, -4, 39, -5, -111, -6, -46, -7, 71, -7, -127, -7, 116, -7, -105, -7, 93, -6, 118, -5, -111, -4, 25, -2, -23, -1, -119, 1, 103, 3, -4, 4, 45, 6, 32, 7, -41, 7, 70, 8, -91, 8, 116, 8, -25, 7, -94, 7, -96, 6, 52, 5, -79, 3, -41, 1, -38, -1, 65, -2, 100, -4, 42, -6, 113, -8, 33, -9, -118, -11, 12, -12, 77, -13, 23, -13, 75, -13, -99, -13, 37, -12, 9, -11, 40, -10, -85, -9, 9, -7, 2, -5, 92, -3, 126, -1, -47, 1, -25, 3, -19, 5, 112, 7, -106, 8, -37, 9, -85, 10, 93, 11, -77, 11, 72, 11, 2, 11, 32, 10, -4, 8, -72, 7, -33, 5, 123, 4, -51, 2, -125, 0, 59, -2, 48, -4, -57, -6, -27, -8, 106, -9, -70, -10, 51, -10, -27, -11, -53, -11, -10, -11, -91, -10, -59, -9, -26, -8, -8, -7, -121, -5, -120, -3, 88, -1, -44, 0, -60, 2, -80, 4, 21, 6, 90, 7, 92, 8, 3, 9, -97, 9, 115, 9, -69, 8, 1, 8, 70, 7, 12, 6, 84, 4, -34, 2, 127, 1, -6, -1, -20, -3, 8, -4, -28, -6, -21, -7, 18, -7, 71, -8, 22, -8, 123, -8, -53, -8, 12, -7, -88, -7, 46, -5, -28, -4, 17, -2, -75, -1

http://www.kler.cn/news/108327.html

相关文章：

  • DevOps与CI/CD的最佳实践
  • python随手小练10（南农作业题）
  • 【赠书活动】从瀑布模式到水母模式：ChatGPT如何赋能软件研发全流程
  • 79 电话号码的字母组合
  • 【Linux】解决缓存锁问题：无法获得锁 /var/lib/dpkg/lock-frontend
  • Qt配置OpenCV教程，亲测已试过
  • 如何使用 nvm-windows 这个工具来管理你电脑上的Node.js版本
  • uniapp vue国际化 i18n
  • springcloud技术栈以及相关组件
  • 计算机中了locked勒索病毒怎么办，locked勒索病毒解密，数据恢复
  • SpringSecurity 认证实战
  • 拉扎维模拟CMOS集成电路设计西交张鸿老师课程P2~5视频学习记录
  • c++ 线程安全的string类
  • 雨云虚拟主机使用教程WordPress博客网站搭建教程
  • 客户中心模拟（Queue and A, ACM/ICPC World Finals 2000, UVa822）rust解法
  • Element 多个Form表单 同时验证
  • SpringCore完整学习教程7，入门级别
  • ubuntu 下载Python
  • WPS中图的自动编号及引用
  • 怎么从休学证明中取出休学原因（python自动化办公，涉及word和excel）
  • 精讲双向链表的销毁
  • 【算法|动态规划 | 01背包问题No.1】AcWing 426. 开心的金明
  • springboot 项目非docker 部署自动启动
  • 【教3妹学编程-java实战5】结构体字段赋值的几种方式
  • 推理还是背诵？通过反事实任务探索语言模型的能力和局限性
  • [双指针](一) Leetcode 283.移动零和1089.复写零
  • 2.MySQL的调控按钮——启动选项和系统变量
  • 什么是离岸金融 （OFFSHORE FINANCE）
  • 关于FTP的一些往事
  • Linux多线程服务端编程：使用muduo C++网络库 学习笔记 第四章 C++多线程系统编程精要