基于python语音启动电脑应用程序

复制代码

osk模型进行输入语音转换

txt字典导航程序路径

pyttsx3引擎进行语音打印输出

关键词=程序路径

复制代码

import os
import json
import queue
import sounddevice as sd
from vosk import Model, KaldiRecognizer
import subprocess
import time
import pyttsx3
import threading

# 初始化 pyttsx3 引擎
engine = pyttsx3.init()
engine.setProperty('rate', 150)    # 设置语速
engine.setProperty('volume', 1.0)  # 设置音量

# 加载Vosk模型
model_path = r"D:\daku\yuyinshibie\vosk-model-small-cn-0.22"
if not os.path.exists(model_path):
    print(f"模型路径不存在: {model_path}")
    engine.say(f"模型路径不存在: {model_path}")
    engine.runAndWait()
    exit(1)

# 读取字典文件，格式为 "命令=程序路径"
def load_app_dict(file_path):
    app_dict = {}
    if not os.path.exists(file_path):
        print(f"字典文件不存在: {file_path}")
        engine.say(f"字典文件不存在: {file_path}")
        engine.runAndWait()
        return app_dict

    with open(file_path, 'r', encoding='utf-8') as file:
        for line in file:
            parts = line.strip().split('=')
            if len(parts) == 2:
                keys, value = parts
                # 处理可能存在的别名情况，例如 "微信,weixin"
                for key in keys.split('，'):
                    app_dict[key.strip()] = value.strip()
    return app_dict

# 启动应用程序
def launch_application(app_name, app_dict):
    if app_name in app_dict:
        app_path = app_dict[app_name]
        response = f"正在启动 {app_name}..."
        say(response)
        subprocess.Popen(app_path)
        time.sleep(2)  # 等待2秒再继续监听
    else:
        response = f"找不到与 '{app_name}' 对应的应用程序。"
        say(response)

# 定义一个函数用于语音输出，并在说的时候暂停监听
def say(text):
    global stream, callback_func
    if stream is not None:
        with stream_lock:
            stream.callback = None  # 移除回调函数以暂停监听
            stream.stop()           # 暂停音频流
    engine.say(text)
    engine.runAndWait()
    if stream is not None:
        with stream_lock:
            stream.start()          # 恢复音频流
            stream.callback = callback_func  # 重新设置回调函数

# 初始化模型和识别器
model = Model(model_path)
rec = KaldiRecognizer(model, 16000)

q = queue.Queue()

last_partial_result = ""
last_full_command = ""

stream_lock = threading.Lock()
stream = None
callback_func = None

def callback(indata, frames, time, status):
    if status:
        print(status, file=sys.stderr)
    q.put(bytes(indata))

# 主程序
if __name__ == "__main__":
    dict_file = r"D:\daku\yuyinshibie\zidian.txt"  # 字典文件路径
    app_dict = load_app_dict(dict_file)

    try:
        # 提前初始化音频流
        callback_func = callback
        stream = sd.RawInputStream(samplerate=16000, blocksize=8000, dtype='int16',
                                  channels=1, callback=callback)
        stream.start()

        say("请说：")
        while True:
            data = q.get()
            if rec.AcceptWaveform(data):
                result = json.loads(rec.Result())
                command = result['text'].strip()
                if command and command != last_full_command:
                    print(f"你说的是: {command}")
                    say(f"你说的是: {command}")
                    if "打开" in command:
                        app_to_open = command.replace("打开", "").strip()
                        launch_application(app_to_open, app_dict)
                    last_full_command = command
            elif rec.PartialResult():
                partial_result = json.loads(rec.PartialResult())['partial']
                if partial_result and "打开" in partial_result and partial_result != last_partial_result:
                    print(f"部分结果: {partial_result}")
                    say(f"部分结果: {partial_result}")
                    last_partial_result = partial_result
    except KeyboardInterrupt:
        say("\n退出程序。")
    finally:
        if stream is not None:
            stream.stop()
            stream.close()

关键词部分，为了识别准确以及出现谐音内容可以增添多个关键词使用，作为分割

字典路径如果出现中文字符有可能会报错！

代码意义不大，如果考虑深入：可以尝试增加快捷键，以及相关应用接口可以更好控制

上班族打开电脑i第一件事情是启动相关应用，同样可以尝试多应用编组启动