语音助手遇背景噪音时类型异常致崩溃的问题求助

当只有背景噪音无有效语音时，listener.recognize_google(voice, show_all=True)返回的actual_result不再是包含alternative键的字典，而是列表（或空结构）。后续代码直接以字典方式访问actual_result["alternative"]，触发TypeError。

修改take_command函数，增加对actual_result类型和有效性的校验，过滤无意义输入后重新监听，直到获取有效语音命令：

import PIL
import speech_recognition as sr
import pyttsx3
import random
import time

listener = sr.Recognizer()
engine = pyttsx3.init()


def talk(text):
    try:
        print("Assistant - " + text.capitalize())
        engine.say(text)
        engine.runAndWait()
    except ValueError:
        pass


def take_command():
    try:
        with sr.Microphone() as source:
            # 校准环境噪音，减少误识别
            listener.adjust_for_ambient_noise(source, duration=0.5)
            voice = listener.listen(source)
            actual_result = listener.recognize_google(voice, show_all=True)
            
            # 校验结果有效性：必须是字典且包含alternative键
            if not isinstance(actual_result, dict) or "alternative" not in actual_result:
                return take_command()
            
            alternatives = actual_result["alternative"]
            # 确保alternatives是有效列表
            if not isinstance(alternatives, list) or len(alternatives) == 0:
                return take_command()
            
            # 选择最优识别结果
            if any("confidence" in alt for alt in alternatives):
                best_hypothesis = max(alternatives, key=lambda alt: alt.get("confidence", 0))
            else:
                best_hypothesis = alternatives[0]
            
            # 确保存在有效转录文本
            if "transcript" not in best_hypothesis:
                return take_command()
            
            command = best_hypothesis["transcript"].lower().replace('alexa', '')
            # 过滤空命令
            if not command.strip():
                return take_command()
            
            print("\nUser - " + command.capitalize())
            return command

    # 捕获语音识别相关异常：无语音、API请求失败等
    except (sr.UnknownValueError, sr.RequestError, TypeError, ValueError):
        return take_command()

• 增加listener.adjust_for_ambient_noise：提前校准环境噪音，降低误识别概率
• 多层有效性校验：依次检查结果类型、alternative字段、列表有效性、转录文本存在性
• 过滤空命令：排除纯空格等无意义输入
• 扩展异常捕获：覆盖语音识别过程中常见的错误类型，确保程序稳定重试

内容的提问来源于stack exchange，提问作者Nir Malka