vue+科大讯飞语音听写功能(解决针对vue new Worker报错问题)

Posted 2021-09-07 Saber Lily

1. 需求: 做语音听写然后转换成文字

2. 使用api: 科大讯飞语音听写

3. 在vue项目中配置文件步骤

4. 在配置完成后处理IatRecorder.js 里new Woeker()会报错的问题

现在开始: 

第一步：我们到官方先下载个js版的demo，拿到IatRecorder.js和transcode.worker.js文件复制到src下面

 transcode.worker.js：

// (function(){
  self.onmessage = function(e){
    transAudioData.transcode(e.data)
  }
  let transAudioData = {
    transcode(audioData) {
      let output = transAudioData.to16kHz(audioData)
      output = transAudioData.to16BitPCM(output)
      output = Array.from(new Uint8Array(output.buffer))
      self.postMessage(output)
      // return output
    },
    to16kHz(audioData) {
      var data = new Float32Array(audioData)
      var fitCount = Math.round(data.length * (16000 / 44100))
      var newData = new Float32Array(fitCount)
      var springFactor = (data.length - 1) / (fitCount - 1)
      newData[0] = data[0]
      for (let i = 1; i < fitCount - 1; i++) {
        var tmp = i * springFactor
        var before = Math.floor(tmp).toFixed()
        var after = Math.ceil(tmp).toFixed()
        var atPoint = tmp - before
        newData[i] = data[before] + (data[after] - data[before]) * atPoint
      }
      newData[fitCount - 1] = data[data.length - 1]
      return newData
    },
    to16BitPCM(input) {
      var dataLength = input.length * (16 / 8)
      var dataBuffer = new ArrayBuffer(dataLength)
      var dataView = new DataView(dataBuffer)
      var offset = 0
      for (var i = 0; i < input.length; i++, offset += 2) {
        var s = Math.max(-1, Math.min(1, input[i]))
        dataView.setInt16(offset, s < 0 ? s * 0x8000 : s * 0x7fff, true)
      }
      return dataView
    },
  }
// })()

IatRecorder.js：

这里需要你手动安装crypto-js,然后引入import CryptoJS from 'crypto-js'

npm install crypto-js --save-dev

const APPID = '你的APPID'
const API_SECRET = '你的API_SECRET '
const API_KEY = '你的API_KEY'
import CryptoJS from 'crypto-js'
import Worker from './transcode.worker.js'
const transWorker = new Worker()
console.log(transWorker)
var startTime = ""
var endTime = ""

function getWebSocketUrl(){
  return new Promise((resolve, reject) => {
 // 请求地址根据语种不同变化
 var url = 'wss://iat-api.xfyun.cn/v2/iat'
 var host = 'iat-api.xfyun.cn'
 var apiKey = API_KEY
 var apiSecret = API_SECRET
 var date = new Date().toGMTString()
 var algorithm = 'hmac-sha256'
 var headers = 'host date request-line'
 var signatureOrigin = `host: ${host}\\ndate: ${date}\\nGET /v2/iat HTTP/1.1`
 var signatureSha = CryptoJS.HmacSHA256(signatureOrigin, apiSecret)
 var signature = CryptoJS.enc.Base64.stringify(signatureSha)
 var authorizationOrigin = `api_key="${apiKey}", algorithm="${algorithm}", headers="${headers}", signature="${signature}"`
 var authorization = btoa(authorizationOrigin)
 url = `${url}?authorization=${authorization}&date=${date}&host=${host}`
 resolve(url)
 })
}
const IatRecorder = class  {
    constructor({ language, accent, appId } = {}) {
      let self = this
      this.status = 'null'
      this.language = language || 'zh_cn'
      this.accent = accent || 'mandarin'
      this.appId = appId || APPID
      // 记录音频数据
      this.audioData = []
      // 记录听写结果
      this.resultText = ''
      // wpgs下的听写结果需要中间状态辅助记录
      this.resultTextTemp = ''
      transWorker.onmessage = function (event) {
        // console.log("构造方法中",self.audioData)
        self.audioData.push(...event.data)
      }
    }

    // 修改录音听写状态
    setStatus(status) {
      this.onWillStatusChange && this.status !== status && this.onWillStatusChange(this.status, status)
      this.status = status
    }
    setResultText({ resultText, resultTextTemp } = {}) {
      this.onTextChange && this.onTextChange(resultTextTemp || resultText || '')
      resultText !== undefined && (this.resultText = resultText)
      resultTextTemp !== undefined && (this.resultTextTemp = resultTextTemp)
    }
    // 修改听写参数
    setParams({ language, accent } = {}) {
      language && (this.language = language)
      accent && (this.accent = accent)
    }
    // 连接websocket
    connectWebSocket() {
      return getWebSocketUrl().then(url => {
        let iatWS
        if ('WebSocket' in window) {
          iatWS = new WebSocket(url)
        } else if ('MozWebSocket' in window) {
          iatWS = new MozWebSocket(url)
        } else {
          alert('浏览器不支持WebSocket')
          return
        }
        this.webSocket = iatWS
        this.setStatus('init')
        iatWS.onopen = e => {
          this.setStatus('ing')
          // 重新开始录音
          setTimeout(() => {
            this.webSocketSend()
          }, 500)
        }
        iatWS.onmessage = e => {
          this.result(e.data)
        }
        iatWS.onerror = e => {
          this.recorderStop()
        }
        iatWS.onclose = e => {
          endTime = Date.parse(new Date())
          console.log("持续时间",endTime-startTime)
          this.recorderStop()
        }
      })
    }
    // 初始化浏览器录音
    recorderInit() {
      navigator.getUserMedia =
        navigator.getUserMedia ||
        navigator.webkitGetUserMedia ||
        navigator.mozGetUserMedia ||
        navigator.msGetUserMedia

      // 创建音频环境
      try {
        this.audioContext = new (window.AudioContext || window.webkitAudioContext)()
        this.audioContext.resume()
        if (!this.audioContext) {
          alert('浏览器不支持webAudioApi相关接口')
          return
        }
      } catch (e) {
        if (!this.audioContext) {
          alert('浏览器不支持webAudioApi相关接口')
          return
        }
      }

      // 获取浏览器录音权限
      if (navigator.mediaDevices && navigator.mediaDevices.getUserMedia) {
        navigator.mediaDevices
          .getUserMedia({
            audio: true,
            video: false,
          })
          .then(stream => {
            getMediaSuccess(stream)
          })
          .catch(e => {
            getMediaFail(e)
          })
      } else if (navigator.getUserMedia) {
        navigator.getUserMedia(
          {
            audio: true,
            video: false,
          },
          stream => {
            getMediaSuccess(stream)
          },
          function(e) {
            getMediaFail(e)
          }
        )
      } else {
        if (navigator.userAgent.toLowerCase().match(/chrome/) && location.origin.indexOf('https://') < 0) {
          alert('chrome下获取浏览器录音功能，因为安全性问题，需要在localhost或127.0.0.1或https下才能获取权限')
        } else {
          alert('无法获取浏览器录音功能，请升级浏览器或使用chrome')
        }
        this.audioContext && this.audioContext.close()
        return
      }
      // 获取浏览器录音权限成功的回调
      let getMediaSuccess = stream => {
        // 创建一个用于通过javascript直接处理音频
        this.scriptProcessor = this.audioContext.createScriptProcessor(0, 1, 1)
        this.scriptProcessor.onaudioprocess = e => {
          // 去处理音频数据
           if (this.status === 'ing') {
             transWorker.postMessage(e.inputBuffer.getChannelData(0))
            //  this.audioData.push(e.inputBuffer.getChannelData(0))
           }
        }
        // 创建一个新的MediaStreamAudiosourceNode 对象，使来自MediaStream的音频可以被播放和操作
        this.mediaSource = this.audioContext.createMediaStreamSource(stream)
        // 连接
        this.mediaSource.connect(this.scriptProcessor)
        this.scriptProcessor.connect(this.audioContext.destination)
        this.connectWebSocket()
      }

      let getMediaFail = (e) => {
        this.audioContext && this.audioContext.close()
        this.audioContext = undefined
        // 关闭websocket
        if (this.webSocket && this.webSocket.readyState === 1) {
          this.webSocket.close()
        }
      }
    }
    recorderStart() {
      if (!this.audioContext) {
        this.recorderInit()
      } else {
        this.audioContext.resume()
        this.connectWebSocket()
      }
    }
    // 暂停录音
    recorderStop() {
      // safari下suspend后再次resume录音内容将是空白，设置safari下不做suspend
      if (!(/Safari/.test(navigator.userAgent) && !/Chrome/.test(navigator.userAgen))){
        this.audioContext && this.audioContext.suspend()
      }
      this.setStatus('end')
    }
    // 处理音频数据
    // transAudioData(audioData) {
    //   audioData = transAudioData.transaction(audioData)
    //   this.audioData.push(...audioData)
    // }
    // 对处理后的音频数据进行base64编码，
    toBase64(buffer) {
      var binary = ''
      var bytes = new Uint8Array(buffer)
      var len = bytes.byteLength
      for (var i = 0; i < len; i++) {
        binary += String.fromCharCode(bytes[i])
      }
      return window.btoa(binary)
    }
    // 向webSocket发送数据
    webSocketSend() {
      if (this.webSocket.readyState !== 1) {
        return
      }
      let audioData = this.audioData.splice(0, 1280)
      var params = {
        common: {
          app_id: this.appId,
        },
        business: {
          language: this.language, //小语种可在控制台--语音听写（流式）--方言/语种处添加试用
          domain: 'iat',
          accent: this.accent, //中文方言可在控制台--语音听写（流式）--方言/语种处添加试用
        },
        data: {
          status: 0,
          format: 'audio/L16;rate=16000',
          encoding: 'raw',
          audio: this.toBase64(audioData),
        },
      }
      console.log("参数language：",this.language)
      console.log("参数accent：",this.accent)
      this.webSocket.send(JSON.stringify(params))
      startTime = Date.parse(new Date())
      this.handlerInterval = setInterval(() => {
        // websocket未连接
        if (this.webSocket.readyState !== 1) {
          console.log("websocket未连接")
          this.audioData = []
          clearInterval(this.handlerInterval)
          return
        }
        if (this.audioData.length === 0) {
          console.log("自动关闭",this.status)
          if (this.status === 'end') {
            this.webSocket.send(
              JSON.stringify({
                data: {
                  status: 2,
                  format: 'audio/L16;rate=16000',
                  encoding: 'raw',
                  audio: '',
                },
              })
            )
            this.audioData = []
            clearInterval(this.handlerInterval)
          }
          return false
        }
        audioData = this.audioData.splice(0, 1280)
        // 中间帧
        this.webSocket.send(
          JSON.stringify({
            data: {
              status: 1,
              format: 'audio/L16;rate=16000',
              encoding: 'raw',
              audio: this.toBase64(audioData),
            },
          })
        )
      }, 40)
    }
    result(resultData) {
      // 识别结束
      let jsonData = JSON.parse(resultData)
      if (jsonData.data && jsonData.data.result) {
        let data = jsonData.data.result
        let str = ''
        let resultStr = ''
        let ws = data.ws
        for (let i = 0; i < ws.length; i++) {
          str = str + ws[i].cw[0].w
        }
        console.log("识别的结果为：",str)
        // 开启wpgs会有此字段(前提：在控制台开通动态修正功能)
        // 取值为 "apd"时表示该片结果是追加到前面的最终结果；取值为"rpl" 时表示替换前面的部分结果，替换范围为rg字段
        if (data.pgs) {
          if (data.pgs === 'apd') {
            // 将resultTextTemp同步给resultText
            this.setResultText({
              resultText: this.resultTextTemp,
            })
          }
          // 将结果存储在resultTextTemp中
          this.setResultText({
            resultTextTemp: this.resultText + str,
          })
        } else {
          this.setResultText({
            resultText: this.resultText + str,
          })
        }
      }
      if (jsonData.code === 0 && jsonData.data.status === 2) {
        this.webSocket.close()
      }
      if (jsonData.code !== 0) {
        this.webSocket.close()
        console.log(`${jsonData.code}:${jsonData.message}`)
      }
    }
    start() {
      this.recorderStart()
      this.setResultText({ resultText: '', resultTextTemp: '' })
    }
    stop() {
      this.recorderStop()
    }
  }

  export default IatRecorder

第二步:  在你使用的页面配置： 

<template>
  <div class="conter">
      <button @click="translationStart">开始</button>
      <button @click="translationEnd">停止</button>
  </div>
</template>
<script>
import IatRecorder from '@/assets/js/IatRecorder';
const iatRecorder = new IatRecorder('en_us','mandarin','9abbbfb0')//小语种-中文方言-appId
export default {
  data() {
    return {
    };
  },
  methods: {
    translationStart() {
      iatRecorder.start();
    },
    translationEnd() {
      iatRecorder.onTextChange = (text) => {
        let inputText = text;
        this.searchData = inputText.substring(0, inputText.length - 1); //文字处理，因为不知道为什么识别输出的后面都带‘。’，这个方法是去除字符串最后一位
        console.log(this.searchData);
      };
      iatRecorder.stop();
    },
    }
};
</script>

 在你引用IatRecorder .js的时候问题来了

IatRecorder .js文件里面const transWorker = new Worker() 就会报错，原因是vue里面不能直接使用原生的new Worker,也会跟webpack版本有关系.

众所周知，JavaScript是单线程的，一些复杂比较耗时的操作，会阻塞页面的渲染交互，引起页面卡顿，影响用户体验。web worker是html5的新特性之一，主要就是用来解决此类问题，为页面额外开启一个线程，用来处理一些比较耗时操作，不影响主线程的进行。

在实际vue项目的开发使用过程中，还是遇到不少坑

其不同模块间的通信主要通过postMessage进行消息推送，通过onmessage进行消息接收，所以vue项目不能直接使用，得配置，而科大讯飞语音官网上给的不是vue项目示例，也没有特别说明，所以我也踩了一个坑.

第三步：解决new Worker()报错问题

首先两个js文件都不用改动

然后安装(这里安装最新版后运行会报错，建议使用2.0.0版本，而不是最新版)

//最新版本
npm install worker-loader -D
//指定版本
npm install worker-loader@2.0.0 -D

vue.config.js要添加以下配置：

configureWebpack: config => {
    config.module.rules.push({
      test: /\\.worker.js$/,
      use: {
        loader: 'worker-loader',
        options: { inline: true, name: 'workerName.[hash].js' }
      }
    })
  },

在你运行时候,会发现控制台会报错，“window is undefined”，这个是因为worker线程中不存在window对象，因此不能直接使用，要用this代替，要在vue.config.js中添加以下配置

chainWebpack: config => {
    config.output.globalObject('this')
 }

打包的时候报错就加上:

parallel: false

合成一起就是

module.exports = {
  configureWebpack: config => {
    config.module.rules.push({
      test: /\\.worker.js$/,
      use: {
        loader: 'worker-loader',
        options: { inline: true, name: 'workerName.[hash].js' }
      }
    })
  },
  parallel: false,
  chainWebpack: config => {
    config.output.globalObject('this')
  }
}

配置完后你会发现不会报错了，然后就可以正常运行了!

到此结束，希望猿友们一键三连~~~