1# Copyright (c) 2024 Baidu, Inc. All Rights Reserved.
2#
3# Licensed under the Apache License, Version 2.0 (the "License");
4# you may not use this file except in compliance with the License.
5# You may obtain a copy of the License at
6#
7#     http://www.apache.org/licenses/LICENSE-2.0
8#
9# Unless required by applicable law or agreed to in writing, software
10# distributed under the License is distributed on an "AS IS" BASIS,
11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12# See the License for the specific language governing permissions and
13# limitations under the License.
14
15import os
16import time
17import wave
18import sys
19import pyaudio
20import webrtcvad
21import appbuilder
22import re
23
24# 请前往千帆AppBuilder官网创建密钥，流程详见：https://cloud.baidu.com/doc/AppBuilder/s/Olq6grrt6#1%E3%80%81%E5%88%9B%E5%BB%BA%E5%AF%86%E9%92%A5
25# 设置环境变量
26os.environ["APPBUILDER_TOKEN"] = (
27    "..."
28)
29# 已发布AppBuilder应用的ID
30app_id = "..."
31appbuilder.logger.setLoglevel("ERROR")
32
33CHUNK = 1024
34FORMAT = pyaudio.paInt16
35CHANNELS = 1 if sys.platform == "darwin" else 2
36RATE = 16000
37DURATION = 30  # ms
38CHUNK = RATE // 1000 * DURATION
39
40
41class Chatbot:
42    def __init__(self):
43        self.p = pyaudio.PyAudio()
44        self.tts = appbuilder.TTS()
45        self.asr = appbuilder.ASR()
46        self.agent = appbuilder.AppBuilderClient(app_id)
47        self.conversation_id = self.agent.create_conversation()
48
49    def run(self):
50        self.run_tts_and_play_audio(
51            "我是你的专属聊天机器人，如果你有什么问题，可以直接问我"
52        )
53        while True:
54            # Record
55            audio_path = "output.wav"
56            print("开始记录音频...")
57            if self.record_audio(audio_path) < 1000:
58                time.sleep(1)
59                continue
60            print("音频记录结束")
61
62            # ASR
63            print("开始执行ASR...")
64            query = self.run_asr(audio_path)
65            print("结束执行ASR")
66
67            # Agent
68            print("query: ", query)
69            if len(query) == 0:
70                continue
71            answer = self.run_agent(query)
72            results = re.findall(r"(https?://[^\s]+)", answer)
73            for result in results:
74                print("链接地址:", result)
75                answer = answer.replace(result, "")
76            print("answer:", answer)
77
78            # TTS
79            print("开始执行TTS并播报...")
80            self.run_tts_and_play_audio(answer)
81            print("结束TTS并播报结束")
82
83    def record_audio(self, path):
84        with wave.open(path, "wb") as wf:
85            wf.setnchannels(CHANNELS)
86            wf.setsampwidth(self.p.get_sample_size(FORMAT))
87            wf.setframerate(RATE)
88            stream = self.p.open(
89                format=FORMAT, channels=CHANNELS, rate=RATE, input=True
90            )
91            vad = webrtcvad.Vad(1)
92            not_speech_times = 0
93            speech_times = 0
94            total_times = 0
95            start_up_times = 33 * 5  # 初始时间设置为5秒
96            history_speech_times = 0
97            while True:
98                if history_speech_times > 33 * 10:
99                    break
100                data = stream.read(CHUNK, False)
101                if vad.is_speech(data, RATE):
102                    speech_times += 1
103                    wf.writeframes(data)
104                else:
105                    not_speech_times += 1
106                total_times += 1
107                if total_times >= start_up_times:
108                    history_speech_times += speech_times
109                    # 模拟滑窗重新开始计数
110                    if float(not_speech_times) / float(total_times) > 0.7:
111                        break
112                    not_speech_times = 0
113                    speech_times = 0
114                    total_times = 0
115                    start_up_times = start_up_times / 2
116                    if start_up_times < 33:
117                        start_up_times = 33
118            stream.close()
119            return history_speech_times * DURATION
120
121    def run_tts_and_play_audio(self, text: str):
122        # AppBuilder内置的TTS使用文档，用户可根据文档调整参数：https://github.com/baidubce/app-builder/tree/master/python/core/components/tts
123        msg = self.tts.run(
124            appbuilder.Message(content={"text": text}),
125            speed=5,
126            pitch=5,
127            volume=5,
128            person=0,
129            audio_type="pcm",
130            model="paddlespeech-tts",
131            stream=True,
132        )
133        stream = self.p.open(
134            format=self.p.get_format_from_width(2),
135            channels=1,
136            rate=24000,
137            output=True,
138            frames_per_buffer=2048,
139        )
140        for pcm in msg.content:
141            stream.write(pcm)
142        stream.stop_stream()
143        stream.close()
144
145    # AppBuilder内置的ASR使用文档，用户可根据文档调整参数：https://github.com/baidubce/app-builder/blob/master/python/core/components/asr/README.md
146    def run_asr(self, audio_path: str):
147        with open(audio_path, "rb") as f:
148            content_data = {"audio_format": "wav", "raw_audio": f.read(), "rate": 16000}
149            msg = appbuilder.Message(content_data)
150            out = self.asr.run(msg)
151            text = out.content["result"][0]
152            return text
153
154    def run_agent(self, query):
155        msg = self.agent.run(self.conversation_id, query, stream=True)
156        answer = ""
157        for content in msg.content:
158            answer += content.answer
159        return answer
160
161
162if __name__ == "__main__":
163    chatbot = Chatbot()
164    chatbot.run()