第十一篇【传奇开心果系列】Python的文本和语音相互转换库技术点案例示例：Microsoft Azure Cognitive Services开发聊天机器人经典案例_机器人语音处理services

print(“Say something…”)

Start speech recognition

result = speech_recognizer.recognize_once()

Get recognized text

if result.reason == speechsdk.ResultReason.RecognizedSpeech:
print(“Recognized: {}”.format(result.text))
elif result.reason == speechsdk.ResultReason.NoMatch:
print(“No speech could be recognized”)
elif result.reason == speechsdk.ResultReason.Canceled:
cancellation_details = result.cancellation_details
print(“Speech recognition canceled: {}”.format(cancellation_details.reason))

Set up the speech synthesizer

speech_synthesizer = speechsdk.SpeechSynthesizer(speech_config)

Synthesize text to speech

text_to_speak = “Hello! How can I help you today?”
result = speech_synthesizer.speak_text_async(text_to_speak).get()

if result.reason == speechsdk.ResultReason.SynthesizingAudioCompleted:
print(“Speech synthesized to audio.”)
elif result.reason == speechsdk.ResultReason.Canceled:
cancellation_details = result.cancellation_details
print(“Speech synthesis canceled: {}”.format(cancellation_details.reason))

这段代码演示了如何使用Azure Cognitive Services中的语音识别和语音合成功能。首先，我们设置了语音配置（speech config），然后创建了一个语音识别器（speech recognizer），并启动语音识别来识别用户说的话语。接着，根据识别结果，我们打印出识别的文本。


接下来，我们设置了语音合成器（speech synthesizer），并使用它将指定的文本转换为语音。在这个示例中，我们将文本“Hello! How can I help you today?”转换为语音。最后，我们打印出语音合成的结果。


通过结合语音识别和语音合成功能，你可以为聊天机器人添加语音交互的能力，使用户可以通过语音与机器人进行交流。这种功能可以提升用户体验，特别是在需要使用语音进行交互的场景下。


![在这里插入图片描述](https://img-blog.csdnimg.cn/a5a00bca6ff74f82a1455cf89491e801.jpg)当扩展示例代码以包括语音识别和语音合成功能时，你可以考虑以下示例代码来展示如何结合这两个功能，以支持语音交互的聊天机器人：

import azure.cognitiveservices.speech as speechsdk

Set up the speech config for speech recognition

speech_key = “YOUR_SPEECH_KEY”
service_region = “YOUR_SERVICE_REGION”
speech_config = speechsdk.SpeechConfig(subscription=speech_key, region=service_region)

Set up the speech recognizer

speech_recognizer = speechsdk.SpeechRecognizer(speech_config)

print(“Listening…”)

Start speech recognition

result = speech_recognizer.recognize_once()

Get recognized text

if result.reason == speechsdk.ResultReason.RecognizedSpeech:
user_input = result.text
print(“Recognized: {}”.format(user_input))
elif result.reason == speechsdk.ResultReason.NoMatch:
print(“No speech could be recognized”)
elif result.reason == speechsdk.ResultReason.Canceled:
cancellation_details = result.cancellation_details
print(“Speech recognition canceled: {}”.format(cancellation_details.reason))

Set up the speech synthesizer for text-to-speech

speech_synthesizer = speechsdk.SpeechSynthesizer(speech_config)

Define the response text

response_text = “You said: {}”.format(user_input)

Synthesize text to speech

result = speech_synthesizer.speak_text_async(response_text).get()

if result.reason == speechsdk.ResultReason.SynthesizingAudioCompleted:
print(“Speech synthesized to audio.”)
elif result.reason == speechsdk.ResultReason.Canceled:
cancellation_details = result.cancellation_details
print(“Speech synthesis canceled: {}”.format(cancellation_details.reason))

在这段代码中，首先设置了语音配置（speech config）用于语音识别和语音合成功能。然后创建了一个语音识别器（speech recognizer），并启动语音识别以识别用户说的话语。识别到的文本被存储在`user_input`变量中。


接着，设置了语音合成器（speech synthesizer），并定义了一个回复文本`response_text`，其中包含了用户输入的内容。使用语音合成器将回复文本转换为语音，并输出结果。


通过结合语音识别和语音合成功能，你可以构建一个支持语音交互的聊天机器人。用户可以通过语音输入与机器人交流，机器人可以识别用户的语音输入并以语音形式回复，从而实现更加自然和便捷的交互体验。


### 五、知识库示例代码和扩展


![在这里插入图片描述](https://img-blog.csdnimg.cn/741628294155451fba4314e9f0d9e599.jpg)以下是一个示例代码，演示如何使用Azure Cognitive Services中的知识库（QnA Maker）功能构建一个问答库，以便聊天机器人能够回答用户的常见问题：

from azure.cognitiveservices.knowledge.qnamaker import QnAMakerClient
from azure.cognitiveservices.knowledge.qnamaker.models import QueryDTO

Set up QnA Maker client

subscription_key = “YOUR_SUBSCRIPTION_KEY”
endpoint = “YOUR_QNA_MAKER_ENDPOINT”
kb_id = “YOUR_KNOWLEDGE_BASE_ID”
client = QnAMakerClient(endpoint, CognitiveServicesCredentials(subscription_key))

Define a function to query the QnA Maker knowledge base

def query_knowledge_base(question):
query = QueryDTO(question=question)
response = client.runtime.generate_answer(kb_id, query)
if response.answers:
return response.answers[0].answer
else:
return “Sorry, I don’t have an answer to that question.”

Example usage

question = “What is the capital of France?”
answer = query_knowledge_base(question)