最近有點頹，還是寫寫python緩解一下（不務正業x

idea

之前看到有拆筆記本螢幕做魔鏡的覺得特別神奇，當時記得是要樹莓派來著。現在自己搞搞發現百度的語音識別api還不錯，加上以前註冊的圖靈ai大概也能搞個語音聊天機器人，主要還是筆記本上終於帶了一個麥克風（哭泣

然後搞搞就出來了;-P雖然寫得略醜但是還是很有成就感的

Code

# -*- coding: utf-8 -*-
import wave, pyaudio
from aip import AipSpeech
import numpy as np
import requests, json

APP_ID = '不能告訴你'
 

API_KEY = '不能告訴你'
SECRET_KEY = '不能告訴你'

aipSpeech = AipSpeech(APP_ID, API_KEY, SECRET_KEY)

CHUNK = 1024
FORMAT = pyaudio.paInt16
RATE = 16000
CHANNELS = 1
RECORD_SECONDS = 2
WAVE_OUTPUT_FILENAME = "D:/audio.wav"

def record_wave():
    p = pyaudio.PyAudio()
    stream = p.open(format = FORMAT,
                    channels = CHANNELS,
                    rate = RATE,
                    input = True
 
,
                    frames_per_buffer=CHUNK)
    # print "* recording"
    frames = []
    for i in range(0, int(RATE / CHUNK * RECORD_SECONDS)):
        data = stream.read(CHUNK)
        frames.append(data)

    # print "* done recording"

    stream.stop_stream()
    stream.close()
    p.terminate()

    wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb'
 
)
    wf.setnchannels(CHANNELS)
    wf.setsampwidth(p.get_sample_size(FORMAT))
    wf.setframerate(RATE)
    wf.writeframes(b''.join(frames))
    wf.close()

def get_file_content(filePath):
    with open(filePath, 'rb') as fp:
        return fp.read()

def identify():
    res = aipSpeech.asr(get_file_content(WAVE_OUTPUT_FILENAME), 'wav', 16000, {'lan': 'zh',})
    # print res["err_msg"]
    position = "廣東省東莞市南城區"
    if res["err_msg"] == "success.":
        print(res["result"][0])
        cont = requests.get('http://www.tuling123.com/openapi/api?key=不能告訴你&info=%s&userid=111' % (res["result"][0], )).content
        m = json.loads(cont)
        print (m['text'])

def Monitor():
    p = pyaudio.PyAudio()
    stream = p.open(format=FORMAT,
                    channels=CHANNELS,
                    rate=RATE,
                    input=True,
                    frames_per_buffer=CHUNK)
    # print("開始緩存錄音")
    frames = []
    rec = []
    flag = False
    while (True):
        # print 'begin '
        data = stream.read(CHUNK)
        if flag == True:
            rec.append(data)
        frames.append(data)
        audio_data = np.fromstring(data, dtype=np.short)
        large_sample_count = np.sum( audio_data > 2000 )
        temp = np.max(audio_data)

        # print temp

        if temp > 2000:
            flag = True
            # print "檢測到訊號"
            # print '當前閾值：',temp

        if temp <= 2000:
            # record_wave()
            if flag == True:
                flag = False
                wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb')
                wf.setnchannels(CHANNELS)
                wf.setsampwidth(p.get_sample_size(FORMAT))
                wf.setframerate(RATE)
                wf.writeframes(b''.join(rec))
                wf.close()
                rec = []
                identify()

    stream.stop_stream()
    stream.close()
    p.terminate()

Monitor()

思想就是簡單粗暴地一小段一小段錄音，如果遇到某一段音量先變大後變小就擷取這一段作為一句話，扔到百度語音api識別，再把識別的結果扔到圖靈api裡面

受網速限制，在家裡測試的時候有點延時，連著說幾句會出現漏掉中間的情況。再加上圖靈機器人又有點傻，迴應是純文字而沒做語音導致體驗不太好
懶得改進了，玩玩還是夠用的