Python人工智能之路 - 第三篇 : PyAudio 实现录音 自动化交互实现问答

import pyaudio
import wave

CHUNK = 1024
FORMAT = pyaudio.paInt16
CHANNELS = 2
RATE = 16000
RECORD_SECONDS = 2
WAVE_OUTPUT_FILENAME = "Oldboy.wav"

p = pyaudio.PyAudio()

stream = p.open(format=FORMAT,
                channels=CHANNELS,
                rate=RATE,
                input=True,
                frames_per_buffer=CHUNK)

print("开始录音,请说话......")

frames = []

for i in range(0, int(RATE / CHUNK * RECORD_SECONDS)):
    data = stream.read(CHUNK)
    frames.append(data)

print("录音结束,请闭嘴!")

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()

# pyrec.py 文件内容
import pyaudio
import wave

CHUNK = 1024
FORMAT = pyaudio.paInt16
CHANNELS = 2
RATE = 16000
RECORD_SECONDS = 2

def rec(file_name):
    p = pyaudio.PyAudio()

    stream = p.open(format=FORMAT,
                    channels=CHANNELS,
                    rate=RATE,
                    input=True,
                    frames_per_buffer=CHUNK)

    print("开始录音,请说话......")

    frames = []

    for i in range(0, int(RATE / CHUNK * RECORD_SECONDS)):
        data = stream.read(CHUNK)
        frames.append(data)

    print("录音结束,请闭嘴!")

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

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

{‘err_msg‘: ‘speech quality error.‘, ‘err_no‘: 3301, ‘sn‘: ‘6397933501529645284‘} # 音质不清晰

# wav2pcm.py 文件内容
import os

def wav_to_pcm(wav_file):
    # 假设 wav_file = "音频文件.wav"
    # wav_file.split(".") 得到["音频文件","wav"] 拿出第一个结果"音频文件"  与 ".pcm" 拼接 等到结果 "音频文件.pcm"
    pcm_file = "%s.pcm" %(wav_file.split(".")[0])

    # 就是此前我们在cmd窗口中输入命令,这里面就是在让Python帮我们在cmd中执行命令
    os.system("ffmpeg -y  -i %s  -acodec pcm_s16le -f s16le -ac 1 -ar 16000 %s"%(wav_file,pcm_file))

    return pcm_file

{‘corpus_no‘: ‘6569869134617218414‘, ‘err_msg‘: ‘success.‘, ‘err_no‘: 0, ‘result‘: [‘老男孩教育‘], ‘sn‘: ‘8116162981529666859‘}

# playmp3.py 文件内容
import os

def play_mp3(file_name):
    os.system("ffplay  %s"%(file_name))

# baidu_ai.py 文件内容
from aip import AipSpeech

# 这里的三个参数,对应在百度语音创建的应用中的三个参数
APP_ID = "xxxxx"
API_KEY = "xxxxxxx"
SECRET_KEY = "xxxxxxxx"

client = AipSpeech(APP_ID, API_KEY, SECRET_KEY)


def audio_to_text(pcm_file):
    # 读取文件 , 终于得到了PCM文件
    with open(pcm_file, ‘rb‘) as fp:
        file_context = fp.read()

    # 识别本地文件
    res = client.asr(file_context, ‘pcm‘, 16000, {
        ‘dev_pid‘: 1536,
    })

    # 从字典里面获取"result"的value 列表中第1个元素,就是识别出来的字符串"老男孩教育"
    res_str = res.get("result")[0]

    return res_str


def text_to_audio(res_str):
    synth_file = "synth.mp3"
    synth_context = client.synthesis(res_str, "zh", 1, {
        "vol": 5,
        "spd": 4,
        "pit": 9,
        "per": 4
    })

    with open(synth_file, "wb") as f:
        f.write(synth_context)

    return synth_file

import pyrec  # 录音函数文件
import wav2pcm  # wav转换pcm 函数文件
import baidu_ai  # 语音合成函数,语音识别函数 文件
import playmp3  # 播放mp3 函数 文件

pyrec.rec("1.wav")  # 录音并生成wav文件,使用方式传入文件名

pcm_file = wav2pcm.wav_to_pcm("1.wav")  # 将wav文件 转换成pcm文件 返回 pcm的文件名

res_str = baidu_ai.audio_to_text(pcm_file) # 将转换后的pcm音频文件识别成 文字 res_str

synth_file = baidu_ai.text_to_audio(res_str) # 将res_str 字符串 合成语音 返回文件名 synth_file

playmp3.play_mp3(synth_file) # 播放 synth_file

# FAQ.py 文件内容
def faq(Q):
    if Q == "你叫什么名字": # 问题
        return "我的名字是老男孩教育" # 答案

　　return "我不知道你在说什么" #问题没有答案时返