#32 语音识别

example/lib/wxbot.js

@@ -2,6 +2,7 @@
 const Wechat = require('../../index')
 const debug = require('debug')('wxbot')
 const fs = require('fs')
+const ffmpeg = require('ffmpeg')
 
 class WxBot extends Wechat {
 
@@ -12,7 +13,9 @@ class WxBot extends Wechat {
 
     this.replyUsers = new Set()
     this.on('text-message', msg => this._botReply(msg))
-
+    this.on('voice-message', msg => this._botReply(msg))
+    this.on('image-message', msg => this._botReply(msg))
+
     this.superviseUsers = new Set()
     this.openTimes = 0
     this.on('init-message', () => this._botSupervise())
@@ -84,12 +87,70 @@ class WxBot extends Wechat {
       return '现在思路很乱，最好联系下我哥 T_T...'
     })
   }
+
+  _getWavBuf(mp3Buf) {
+    return new Promise((resolve, reject) => {
+      fs.writeFile('1.mp3', mp3Buf, (err,file) => {
+        if(err) {
+          reject(err)
+        }
+        let process = new ffmpeg('1.mp3')
+        process.then( (video) => {
+          video.save('1.wav', () => {
+            fs.readFile('1.wav', (err, buf) => {
+              if(err) {
+                reject(err)
+              }
+              fs.unlink('1.wav')
+              fs.unlink('1.mp3')
+              resolve(buf)
+            })
+          })
+        }).catch(err => {
+          reject(err)
+        })
+      })
+    })
+  }
+
+  _speechRecognize(buf) {
+    let params = {
+      'cuid': 'wechat4u',
+      'lan': 'zh',
+      'token': '24.9a233a67c12f73276a026357b6079dc0.2592000.1462584892.282335-6332804'
+    }
+    return this.request({
+      method: 'POST',
+      url: 'http://vop.baidu.com/server_api',
+      headers: {
+        'Content-Type': 'audio/wav; rate=8000',
+      },
+      params: params,
+      data: buf
+    }).then(res => {
+      const data = res.data
+      debug('语音识别结果：', data.result[0])
+      return data.result[0]
+    })
+  }
 
   _botReply(msg) {
     if (this.replyUsers.has(msg['FromUserName'])) {
-      this._tuning(msg['Content']).then((reply) => {
-        this.sendMsg(reply, msg['FromUserName'])
-        debug(reply)
+      new Promise((resolve, reject) => {
+        if(msg['Content'].type === 'audio/mp3') {
+          this._getWavBuf(msg['Content'].data)
+            .then(buf => this._speechRecognize(buf))
+            .then(content => this._tuning(content))
+            .then(replyContent => resolve(replyContent))
+        } else if(msg['Content'].type === 'image/jpeg') {
+          resolve('好美的照片啊！')
+        } else {
+          this._tuning(msg['Content'])
+              .then(replyContent => resolve(replyContent))
+        }
+      }).then(replyContent => {
+        this.sendMsg(replyContent, msg['FromUserName'])
+        debug('回复消息：', replyContent)
       })
     }
   }

package.json

@@ -42,6 +42,7 @@
     "vue-strap": "^1.0.7",
     "vue-style-loader": "^1.0.0",
     "webpack": "^1.12.2",
+    "ffmpeg": "0.0.4",
     "cross-env": "^1.0.7"
   },
   "repository": {
@@ -68,13 +69,19 @@
     "rules": {
       "camelcase": 2,
       "curly": 2,
-      "brace-style": [2, "1tbs"],
-      "quotes": [2, "single"],
+      "brace-style": [
+        2,
+        "1tbs"
+      ],
+      "quotes": [
+        2,
+        "single"
+      ],
       "space-infix-ops": 2
     },
     "env": {
       "browser": true,
       "node": true
     }
   }
-}
+}