前言

如題：做一個錄音文字識別功能，知識點有三個，分別是微信小程序的錄音功能、錄音文件直傳阿里云OSS、使用阿里云的錄音文件識別接口返回識別后的文字

一、微信小程序錄音

官方文檔:微信小程序全局唯一的錄音管理器 RecorderManager

wxml：

<view style="padding-top:40%"></view>
<button bindtap="start">開始錄音</button>
<!-- <button bindtap="pause">暫停錄音</button> -->
<!-- <button bindtap="resume">繼續(xù)錄音</button> -->
<view style="padding-top:10px"></view>
<button bindtap="stop">停止錄音并識別</button>
<!-- <button bindtap="play">播放錄音[本地]</button> -->
<!-- <button bindtap="play">播放錄音[線上]</button> -->
<view style="text-align:center;padding-top:20px">{{text}}</view>

js：需要注意的是，點擊開始錄音時要判斷當(dāng)前是否獲取到了錄音權(quán)限，如果沒有錄音權(quán)限進行提示，引導(dǎo)用戶重新授權(quán)

const recorderManager = wx.getRecorderManager()
const innerAudioContext = wx.createInnerAudioContext()

Page({
  data: {
    text: '語音識別后文字展示'
  },
  onLoad: function () {
  },
  //開始錄音的時候
  start: function () {
    var that = this
    const options = {
      duration: 100000, //指定錄音的時長，單位 ms
      sampleRate: 16000, //采樣率
      numberOfChannels: 1, //錄音通道數(shù)
      encodeBitRate: 96000, //編碼碼率
      format: 'mp3', //音頻格式，有效值 aac/mp3
      frameSize: 50, //指定幀大小，單位 KB
    }
    //開始錄音
    wx.authorize({
      scope: 'scope.record',
      success() {
        console.log("錄音授權(quán)成功");
        //第一次成功授權(quán)后 狀態(tài)切換為2
        that.setData({
          status: 2,
        })
        recorderManager.start(options);
        recorderManager.onStart(() => {
          //可以彈出模態(tài)框等友好提示
          console.log('錄音開始')
        });
        //錯誤回調(diào)
        recorderManager.onError((res) => {
          //進行錯誤提示
          console.log(res);
        })
      },
      fail() {
        console.log("第一次錄音授權(quán)失敗");
        wx.showModal({
          title: '提示',
          content: '您未授權(quán)錄音，功能將無法使用',
          showCancel: true,
          confirmText: "授權(quán)",
          confirmColor: "#52a2d8",
          success: function (res) {
            if (res.confirm) {
              //確認則打開設(shè)置頁面（重點）
              wx.openSetting({
                success: (res) => {
                  console.log(res.authSetting);
                  if (!res.authSetting['scope.record']) {
                    //未設(shè)置錄音授權(quán)
                    console.log("未設(shè)置錄音授權(quán)");
                    wx.showModal({
                      title: '提示',
                      content: '您未授權(quán)錄音，功能將無法使用',
                      showCancel: false
                    })
                  } else {
                    //第二次才成功授權(quán)
                    console.log("設(shè)置錄音授權(quán)成功");
                    that.setData({
                      status: 2
                    })
                  }
                },
                fail: function () {
                  console.log("授權(quán)設(shè)置錄音失敗");
                }
              })
            } else if (res.cancel) {
              console.log("cancel");
            }
          },
          fail: function () {
            console.log("openfail");
          }
        })
      }
    })
  },
  //暫停錄音-需求不需要暫停
  pause: function () {
    recorderManager.pause();
    recorderManager.onPause((res) => {
      console.log('暫停錄音')
    })
  },
  //繼續(xù)錄音-不需要暫停也就不需要繼續(xù)
  resume: function () {
    recorderManager.resume();
    recorderManager.onStart(() => {
      console.log('重新開始錄音')
    });
    //錯誤回調(diào)
    recorderManager.onError((res) => {
      console.log(res);
    })
  },
  //停止錄音
  stop: function () {
    recorderManager.stop();
    recorderManager.onStop((res) => {
      console.log('停止錄音', res.tempFilePath)
      this.getPolicy(res.tempFilePath)
    })
  },
  //播放聲音
  play: function () {
    innerAudioContext.autoplay = true
    innerAudioContext.src = this.tempFilePath,
      innerAudioContext.onPlay(() => {
        console.log('開始播放')
      })
    innerAudioContext.onError((res) => {
      console.log(res.errMsg)
      console.log(res.errCode)
    })
  }
})

二、微信小程序錄音直傳阿里云OSS

官方文檔：微信小程序直傳實踐

如文檔所示，首先進行配置Bucket跨域訪問，這個不做解釋，其次在微信小程序中配置域名白名單也就是uploadFile和downloadFile合法域名為對應(yīng)Bucket的外網(wǎng)訪問域名

Javascript直傳OSS需要提前獲取policy，服務(wù)端如下:

  public Map<String, String> ossPolicy() {
    DateFormat dateFormat = new SimpleDateFormat("yyyy-MM-dd");
    Date date = new Date();
    String path = "data/" + dateFormat.format(date) + "/";
    try {
      long expireTime = 3000;
      long expireEndTime = System.currentTimeMillis() + expireTime * 1000;
      Date expiration = new Date(expireEndTime);
      PolicyConditions policyConds = new PolicyConditions();
      policyConds.addConditionItem(PolicyConditions.COND_CONTENT_LENGTH_RANGE, 0, 1048576000);
      policyConds.addConditionItem(MatchMode.StartWith, PolicyConditions.COND_KEY, path);

      String postPolicy = ossClient.generatePostPolicy(expiration, policyConds);
      byte[] binaryData = postPolicy.getBytes("utf-8");
      String encodedPolicy = BinaryUtil.toBase64String(binaryData);
      String postSignature = ossClient.calculatePostSignature(postPolicy);

      Map<String, String> respMap = new LinkedHashMap<String, String>();
      respMap.put("accessid", ossClient.getCredentialsProvider().getCredentials().getAccessKeyId());
      respMap.put("policy", encodedPolicy);
      respMap.put("signature", postSignature);
      respMap.put("dir", path);
      respMap.put("host", configService.findConfigValueBySuffix(ConfigConstant.CONFIG_SUFFIX_RESOURCE_URL));
      respMap.put("expire", String.valueOf(expireEndTime / 1000));
      return respMap;
    } catch (UnsupportedEncodingException e) {
      e.printStackTrace();
    }
    return null;
  }

獲取到policy以及簽名等后，直傳錄音文件到阿里云OSS:

  getPolicy: function (path) {
    wx.request({
      url: 'https://xxx.com/oss/policy',
      method: 'GET',
      success: (res) => {
        console.log('獲取policy成功' + path)
        this.upload(res, path)
      }
    })
  },
  upload: function (res, path) {
    const host = res.data.datas.host;
    const signature = res.data.datas.signature;
    const ossAccessKeyId = res.data.datas.accessid;
    const policy = res.data.datas.policy;
    //這里的key實際是新起的文件名和路徑
    const key = res.data.datas.dir + Date.now() + '.mp3';
    wx.uploadFile({
      url: host,
      filePath: path,
      name: 'file', // 必須填file。
      formData: {
        key,
        policy,
        OSSAccessKeyId: ossAccessKeyId,
        signature,
      },
      success: (res) => {
        console.log(res);
        if (res.statusCode === 204) {
          console.log('上傳成功');
          this.toText(key)
        }
      },
      fail: err => {
        console.log(err);
      }
    });
  }

三、阿里云錄音文字識別極速版

進行錄音識別之前需要先創(chuàng)建智能語音交互服務(wù)對應(yīng)的項目，獲得AppKey

官方文檔:創(chuàng)建智能語音交互服務(wù)項目

官方文檔:錄音文件識別極速版

由于將錄音文件直傳OSS了，所以不需要將錄音文件上傳到語音識別里，所以采取第二種方式，使用音頻文件鏈接進行語音識別：文章來源地址http://www.zghlxwxcb.cn/news/detail-496624.html

  public ResultBody record2text(String recordUrl) throws Exception {
    Assert.isBlank(recordUrl, JobErrorInfoEnum.PARAM_NOT_NULL);

    String prefix = configService.findConfigValueBySuffix(ConfigConstant.CONFIG_SUFFIX_RESOURCE_URL);

    AccessToken accessToken = new AccessToken(id, secret);
    accessToken.apply();
    String token = accessToken.getToken();

    String fileName = prefix + recordUrl;
    String format = "mp3";
    int sampleRate = 16000;

    String allText = process(fileName, appKey, token, format, sampleRate);
    Assert.isBlank(allText, JobErrorInfoEnum.RECORD_TRANS_ERROR);

    return ResultBody.success(allText);
  }

  public String process(String fileName, String appKey, String token, String format, int sampleRate) {
    /**
     * 設(shè)置HTTPS REST POST請求
     * 1.使用http協(xié)議
     * 2.語音識別服務(wù)域名：nls-gateway.cn-shanghai.aliyuncs.com
     * 3.語音識別接口請求路徑：/stream/v1/FlashRecognizer
     * 4.設(shè)置必須請求參數(shù)：appkey、token、format、sample_rate
     */
    String request = url;
    request = request + "?appkey=" + appKey;
    request = request + "&token=" + token;
    request = request + "&format=" + format;
    request = request + "&sample_rate=" + sampleRate;

    /**
     * 設(shè)置HTTPS頭部字段
     * 發(fā)送HTTPS POST請求，返回服務(wù)端的響應(yīng)。
     *
     * 1.Content-Type：application/octet-stream
     */
    HashMap<String, String> headers = new HashMap<String, String>();
    String response;
    if (new File(fileName).isFile()) {
      headers.put("Content-Type", "application/octet-stream");
      response = HttpUtil.sendPostFile(request, headers, fileName);
    } else {
      headers.put("Content-Type", "application/text");
      response = HttpUtil.sendPostLink(request, headers, fileName);
    }

    String allText = "";
    if (response != null) {
      Map map = JsonUtils.json2Map(response);
      if ((int) map.get("status") == 20000000) {
        Map result = (Map) map.get("flash_result");
        List<Map> sentences = (List<Map>) result.get("sentences");
        if (sentences != null && sentences.size() > 0) {
          for (Map sentence : sentences) {
            allText += sentence.get("text");
          }
        }
      }
    }
    return allText;
  }

到了這里，關(guān)于微信小程序錄音直傳阿里云OSS并語音識別的文章就介紹完了。如果您還想了解更多內(nèi)容，請在右上角搜索TOY模板網(wǎng)以前的文章或繼續(xù)瀏覽下面的相關(guān)文章，希望大家以后多多支持TOY模板網(wǎng)！