xiaozhi-sdk 0.1.0__tar.gz

xiaozhi_sdk-0.1.0/PKG-INFO

@@ -0,0 +1,58 @@
+Metadata-Version: 2.4
+Name: xiaozhi-sdk
+Version: 0.1.0
+Summary: A short description of your package
+Home-page: https://github.com/dairoot/xiaozhi-sdk
+Author: dairoot
+Author-email: 623815825@qq.com
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+Requires-Dist: numpy
+Requires-Dist: requests>=2.32.1
+Requires-Dist: sounddevice>=0.4.2
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
+
+# 小智SDK (XiaoZhi SDK)
+
+[![Python Version](https://img.shields.io/badge/python-3.8+-blue.svg)](https://www.python.org/downloads/)
+[![License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
+[![PyPI](https://img.shields.io/badge/pypi-xiaozhi--sdk-blue.svg)](https://pypi.org/project/xiaozhi-sdk/)
+
+一个用于连接和控制小智智能设备的Python SDK，支持实时音频通信、MCP工具集成和设备管理功能。
+
+
+## 📦 安装
+
+```bash
+pip install xiaozhi-sdk
+```
+
+
+## 🚀 快速开始
+
+### 命令行使用
+
+最简单的使用方式是通过命令行连接设备：
+
+```bash
+# 查看帮助信息
+python -m xiaozhi_sdk -h
+
+# 连接设备（需要提供MAC地址）
+python -m xiaozhi_sdk 00:11:22:33:44:55
+```
+
+### 编程使用
+...
+

xiaozhi_sdk-0.1.0/README.md

@@ -0,0 +1,33 @@
+# 小智SDK (XiaoZhi SDK)
+
+[![Python Version](https://img.shields.io/badge/python-3.8+-blue.svg)](https://www.python.org/downloads/)
+[![License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
+[![PyPI](https://img.shields.io/badge/pypi-xiaozhi--sdk-blue.svg)](https://pypi.org/project/xiaozhi-sdk/)
+
+一个用于连接和控制小智智能设备的Python SDK，支持实时音频通信、MCP工具集成和设备管理功能。
+
+
+## 📦 安装
+
+```bash
+pip install xiaozhi-sdk
+```
+
+
+## 🚀 快速开始
+
+### 命令行使用
+
+最简单的使用方式是通过命令行连接设备：
+
+```bash
+# 查看帮助信息
+python -m xiaozhi_sdk -h
+
+# 连接设备（需要提供MAC地址）
+python -m xiaozhi_sdk 00:11:22:33:44:55
+```
+
+### 编程使用
+...
+

xiaozhi_sdk-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

xiaozhi_sdk-0.1.0/setup.py

@@ -0,0 +1,24 @@
+from setuptools import find_packages, setup
+
+setup(
+    name="xiaozhi-sdk",  # 包名
+    version="0.1.0",  # 版本号
+    packages=find_packages(),  # 自动发现包
+    install_requires=[  # 依赖
+        "numpy",
+        "requests>=2.32.1",
+        "sounddevice>=0.4.2",
+    ],
+    author="dairoot",
+    author_email="623815825@qq.com",  # 作者邮箱
+    description="A short description of your package",  # 简短描述
+    long_description=open('README.md').read(),  # 详细描述（通常从 README 读取）
+    long_description_content_type='text/markdown',  # README 文件格式
+    url="https://github.com/dairoot/xiaozhi-sdk",  # 项目主页
+    classifiers=[  # 分类元数据
+        "Programming Language :: Python :: 3",
+        "License :: OSI Approved :: MIT License",
+        "Operating System :: OS Independent",
+    ],
+    python_requires=">=3.8",  # 支持的 Python 版本
+)

xiaozhi_sdk-0.1.0/tests/test_pic.py

@@ -0,0 +1,46 @@
+import aiohttp
+import asyncio
+
+
+async def explain_image():
+    url = "http://api.xiaozhi.me/mcp/vision/explain"
+    question = "这个图片里有什么？"
+    image_path = "./file/leijun.jpg"
+
+    boundary = "----ESP32_CAMERA_BOUNDARY"
+    headers = {
+
+        "Content-Type": f"multipart/form-data; boundary={boundary}",
+        # 不设置Transfer-Encoding，aiohttp自动处理
+    }
+
+    with open(image_path, "rb") as f:
+        img_data = f.read()
+
+    # 手动构造multipart body
+    body = bytearray()
+
+    # question字段
+    body.extend(f"--{boundary}\r\n".encode())
+    body.extend(b'Content-Disposition: form-data; name="question"\r\n\r\n')
+    body.extend(question.encode("utf-8"))
+    body.extend(b"\r\n")
+
+    # 文件字段头
+    body.extend(f"--{boundary}\r\n".encode())
+    body.extend(b'Content-Disposition: form-data; name="file"; filename="camera.jpg"\r\n')
+    body.extend(b"Content-Type: image/jpeg\r\n\r\n")
+    body.extend(img_data)
+    body.extend(b"\r\n")
+
+    # multipart结束
+    body.extend(f"--{boundary}--\r\n".encode())
+
+    async with aiohttp.ClientSession() as session:
+        async with session.post(url, data=body, headers=headers) as resp:
+            print(f"Status: {resp.status}")
+            text = await resp.text()
+            print("Response:", text)
+
+if __name__ == "__main__":
+    asyncio.run(explain_image())

xiaozhi_sdk-0.1.0/tests/test_xiaozhi.py

@@ -0,0 +1,92 @@
+import asyncio
+import json
+import os
+import sys
+import time
+import wave
+
+import numpy as np
+import sounddevice as sd
+
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), ".."))
+
+from xiaozhi_sdk import XiaoZhiWebsocket
+from xiaozhi_sdk.utils import read_audio_file
+
+play_audio_over = asyncio.Event()
+
+
+async def assistant_audio_play(audio_queue):
+    # 创建一个持续播放的流
+    stream = sd.OutputStream(samplerate=16000, channels=1, dtype=np.int16)
+    stream.start()
+    last_time = None
+    while True:
+        if not audio_queue:
+            await asyncio.sleep(0.01)
+            if last_time and time.time() - last_time > 3:
+                break
+
+            continue
+
+        pcm_data = audio_queue.popleft()
+        stream.write(pcm_data)
+        last_time = time.time()
+
+    play_audio_over.set()
+    stream.stop()
+    stream.close()
+
+
+def mcp_tool_func():
+
+    def mcp_take_photo(data):
+        return open("./file/leijun.jpg", "rb")
+
+    def mcp_get_device_status(data):
+        return {
+                "audio_speaker": {"volume": 80},
+                "screen": {"brightness": 75, "theme": "light"},
+                "network": {"type": "wifi", "ssid": "wifi名称", "signal": "strong"},
+            }
+
+    def mcp_set_volume(data):
+        return {}
+
+    mcp_tool_func = {
+        "set_volume": mcp_set_volume,
+        "get_device_status": mcp_get_device_status,
+        "take_photo": mcp_take_photo,
+    }
+    return mcp_tool_func
+
+
+
+
+
+async def main():
+    async def message_handler_callback(message):
+        print("message received:", message)
+
+    MAC_ADDR = "fc:01:2c:c9:2b:31"
+    url = "ws://120.79.156.134:8380"
+    url = None
+    xiaozhi = XiaoZhiWebsocket(message_handler_callback, url=url)
+    await xiaozhi.set_mcp_tool_callback(mcp_tool_func())
+    await xiaozhi.init_connection(MAC_ADDR)
+    asyncio.create_task(assistant_audio_play(xiaozhi.audio_queue))
+    await asyncio.sleep(1)
+    
+    # 使用新的音频读取函数
+    for pcm in read_audio_file("./file/take_photo.wav"):
+    # for pcm in read_audio_file("./file/say_hello.wav"):
+        await xiaozhi.send_audio(pcm)
+
+    # 发送静音数据
+    await xiaozhi.send_silence_audio()
+    await asyncio.wait_for(play_audio_over.wait(), timeout=20.0)
+    await xiaozhi.close()
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

xiaozhi_sdk-0.1.0/xiaozhi_sdk/__init__.py

@@ -0,0 +1,155 @@
+import asyncio
+import json
+import os
+import uuid
+import wave
+import websockets
+from collections import deque
+from typing import Dict
+
+from xiaozhi_sdk.config import INPUT_SERVER_AUDIO_SAMPLE_RATE, WSS_URL
+from xiaozhi_sdk.iot import OtaDevice
+from xiaozhi_sdk.mcp import McpTool
+from xiaozhi_sdk.opus import AudioOpus
+from xiaozhi_sdk.utils import read_audio_file, get_wav_info
+
+
+class XiaoZhiWebsocket(McpTool):
+
+    def __init__(self, message_handler_callback=None, url=None, audio_sample_rate=16000, audio_channels=1):
+        super().__init__()
+        self.url = url or WSS_URL
+        self.audio_sample_rate = audio_sample_rate
+        self.audio_channels = audio_channels
+        self.audio_opus = AudioOpus(audio_sample_rate, audio_channels)
+        self.client_id = str(uuid.uuid4())
+        self.mac_addr = None
+        self.message_handler_callback = message_handler_callback
+
+        self.hello_received = asyncio.Event()
+        self.session_id = ""
+        self.audio_queue = deque()
+        self.websocket = None
+        self.message_handler_task = None
+
+    async def send_hello(self, aec: bool):
+        hello_message = {
+            "type": "hello",
+            "version": 1,
+            "features": {"aec": aec, "mcp": True},
+            "transport": "websocket",
+            "audio_params": {
+                "format": "opus",
+                "sample_rate": INPUT_SERVER_AUDIO_SAMPLE_RATE,
+                "channels": 1,
+                "frame_duration": 60,
+            },
+        }
+        await self.websocket.send(json.dumps(hello_message))
+        await asyncio.wait_for(self.hello_received.wait(), timeout=10.0)
+
+    async def start_listen(self):
+        listen_message = {
+            "session_id": self.session_id,
+            "type": "listen",
+            "state": "start",
+            "mode": "realtime"
+        }
+        await self.websocket.send(json.dumps(listen_message))
+
+    async def set_mcp_tool_callback(self, tool_func: Dict[str, callable]):
+        self.tool_func = tool_func
+
+    async def activate_iot_device(self):
+        ota = OtaDevice(self.mac_addr, self.client_id)
+
+        data = await ota.activate_device()
+        if data.get("activation"):
+            await self.send_demo_get_code_audio()
+            challenge = data["activation"]["challenge"]
+            await asyncio.sleep(3)
+            for _ in range(10):
+                if await ota.check_activate(challenge):
+                    break
+                await asyncio.sleep(3)
+
+    async def init_connection(self, mac_addr: str, aec: bool = False):
+        self.mac_addr = mac_addr
+        headers = {
+            "Authorization": "Bearer test-token",
+            "Protocol-Version": "1",
+            "Device-Id": mac_addr,
+            "Client-Id": self.client_id,
+        }
+
+        self.websocket = await websockets.connect(uri=self.url, additional_headers=headers)
+        self.message_handler_task = asyncio.create_task(self.message_handler())
+        await self.send_hello(aec)
+        await self.start_listen()
+        asyncio.create_task(self.activate_iot_device())
+
+    async def send_demo_get_code_audio(self):
+        current_dir = os.path.dirname(os.path.abspath(__file__))
+        wav_path = os.path.join(current_dir, "../file/greet.wav")
+        framerate, nchannels = get_wav_info(wav_path)
+        audio_opus = AudioOpus(framerate, nchannels)
+
+        for pcm_data in read_audio_file(wav_path):
+            opus_data = await audio_opus.pcm_to_opus(pcm_data)
+            await self.websocket.send(opus_data)
+
+    async def send_silence_audio(self, duration_seconds: float = 1.2):
+        # 发送 静音数据
+        frames_count = int(duration_seconds * 1000 / 60)
+        pcm_frame = b"\x00\x00" * int(INPUT_SERVER_AUDIO_SAMPLE_RATE / 1000 * 60)
+
+        for _ in range(frames_count):
+            await self.send_audio(pcm_frame)
+
+    async def send_audio(self, pcm: bytes):
+        if not self.websocket:
+            return
+
+        state = self.websocket.state
+        if state == websockets.protocol.State.OPEN:
+            opus_data = await self.audio_opus.pcm_to_opus(pcm)
+            await self.websocket.send(opus_data)
+        elif state in [websockets.protocol.State.CLOSED, websockets.protocol.State.CLOSING]:
+            if self.message_handler_callback:
+                await self.message_handler_callback({"type": "websocket", "state": "close", "source": "sdk.send_audio"})
+            await asyncio.sleep(0.5)
+        else:
+            await asyncio.sleep(0.1)
+
+    async def message_handler(self):
+        try:
+            async for message in self.websocket:
+                if isinstance(message, bytes):
+                    pcm_array = await self.audio_opus.opus_to_pcm(message)
+                    self.audio_queue.extend(pcm_array)
+                else:
+                    data = json.loads(message)
+                    message_type = data["type"]
+
+                    if message_type == "hello":
+                        self.hello_received.set()
+                        self.session_id = data["session_id"]
+                    elif message_type == "mcp":
+                        await self.mcp(data)
+                    elif self.message_handler_callback:
+                        await self.message_handler_callback(data)
+        except websockets.ConnectionClosed:
+            if self.message_handler_callback:
+                await self.message_handler_callback(
+                    {"type": "websocket", "state": "close", "source": "sdk.message_handler"})
+
+    async def close(self):
+        if self.message_handler_task and not self.message_handler_task.done():
+            self.message_handler_task.cancel()
+            try:
+                await self.message_handler_task
+            except asyncio.CancelledError:
+                pass
+
+        if self.websocket:
+            await self.websocket.close()

xiaozhi_sdk-0.1.0/xiaozhi_sdk/__main__.py

@@ -0,0 +1,90 @@
+import argparse
+import asyncio
+import re
+import time
+from collections import deque
+
+import numpy as np
+import sounddevice as sd
+
+from xiaozhi_sdk import XiaoZhiWebsocket
+
+input_audio = deque()
+
+is_play_audio = False
+
+
+async def message_handler_callback(message):
+    print("message received:", message)
+
+
+async def assistant_audio_play(audio_queue):
+    global is_play_audio
+    # 创建一个持续播放的流
+    stream = sd.OutputStream(samplerate=16000, channels=1, dtype=np.int16)
+    stream.start()
+    last_time = None
+
+    while True:
+
+        if not audio_queue:
+            await asyncio.sleep(0.01)
+            if last_time and time.time() - last_time > 1:
+                is_play_audio = False
+            continue
+
+        is_play_audio = True
+        pcm_data = audio_queue.popleft()
+        stream.write(pcm_data)
+        last_time = time.time()
+
+
+class Client:
+    def __init__(self, mac_address):
+        self.mac_address = mac_address
+        self.xiaozhi = None
+        pass
+
+    async def start(self):
+        self.xiaozhi = XiaoZhiWebsocket(message_handler_callback)
+        await self.xiaozhi.init_connection(self.mac_address, aec=False)
+        asyncio.create_task(assistant_audio_play(self.xiaozhi.audio_queue))
+
+    def callback_func(self, indata, frames, time, status):
+        pcm = (indata.flatten() * 32767).astype(np.int16).tobytes()
+        input_audio.append(pcm)
+
+    async def process_audio(self):
+        while True:
+            if not input_audio:
+                await asyncio.sleep(0.02)
+                continue
+            pcm = input_audio.popleft()
+            if not is_play_audio:
+                await self.xiaozhi.send_audio(pcm)
+
+
+def mac_address(string):
+    """验证是否为有效的MAC地址"""
+    if re.fullmatch(r'([0-9A-Fa-f]{2}:){5}[0-9A-Fa-f]{2}', string):
+        return string
+    else:
+        raise argparse.ArgumentTypeError(f"无效的MAC地址格式: '{string}'")
+
+
+async def main():
+    parser = argparse.ArgumentParser(description="这是一个小智SDK。")
+    parser.add_argument('device', type=mac_address, help='你的小智设备的MAC地址 (格式: XX:XX:XX:XX:XX:XX)')
+
+    args = parser.parse_args()
+    client = Client(args.device)
+    await client.start()
+    await asyncio.sleep(2)
+
+    with sd.InputStream(callback=client.callback_func, channels=1, samplerate=16000, blocksize=960):
+        print("Recording... Press Ctrl+C to stop.")
+        await client.process_audio()  # 持续处理音频
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

xiaozhi_sdk-0.1.0/xiaozhi_sdk/config.py

@@ -0,0 +1,5 @@
+INPUT_SERVER_AUDIO_SAMPLE_RATE = 16000
+
+WSS_URL = "wss://api.tenclass.net/xiaozhi/v1/"
+OTA_URL = "https://api.tenclass.net/xiaozhi/ota/"
+VL_URL = "http://api.xiaozhi.me/mcp/vision/explain"

xiaozhi_sdk-0.1.0/xiaozhi_sdk/data.py

@@ -0,0 +1,58 @@
+mcp_initialize_payload = {
+    "jsonrpc": "2.0",
+    "id": 1,
+    "result": {
+        "protocolVersion": "2024-11-05",
+        "capabilities": {"tools": {}},
+        "serverInfo": {"name": "", "version": "0.0.1"},
+    },
+}
+
+mcp_tool_conf = {
+    "get_device_status": {
+        "description": "Provides the real-time information of the device, including the current status of the audio speaker, screen, battery, network, etc.\nUse this tool for: \n1. Answering questions about current condition (e.g. what is the current volume of the audio speaker?)\n2. As the first step to control the device (e.g. turn up / down the volume of the audio speaker, etc.)",
+        "inputSchema": {"type": "object", "properties": {}},
+    },
+    "set_volume": {
+        "description": "Set the volume of the audio speaker. If the current volume is unknown, you must call `self.get_device_status` tool first and then call this tool.",
+        "inputSchema": {
+            "type": "object",
+            "properties": {"volume": {"type": "integer", "minimum": 0, "maximum": 100}},
+            "required": ["volume"],
+        },
+    },
+    "set_brightness": {
+        "description": "Set the brightness of the screen.",
+        "inputSchema": {
+            "type": "object",
+            "properties": {"brightness": {"type": "integer", "minimum": 0, "maximum": 100}},
+            "required": ["brightness"],
+        },
+    },
+    "set_theme": {
+        "description": "Set the theme of the screen. The theme can be `light` or `dark`.",
+        "inputSchema": {"type": "object", "properties": {"theme": {"type": "string"}}, "required": ["theme"]},
+    },
+    "take_photo": {
+        "description": "Take a photo and explain it. Use this tool after the user asks you to see something.\nArgs:\n  `question`: The question that you want to ask about the photo.\nReturn:\n  A JSON object that provides the photo information.",
+        "inputSchema": {
+            "type": "object",
+            "properties": {"question": {"type": "string"}},
+            "required": ["question"],
+        },
+    },
+    "open_tab": {
+        "description": "Open a web page in the browser. 小智后台：https://xiaozhi.me",
+        "inputSchema": {
+            "type": "object",
+            "properties": {"url": {"type": "string"}},
+            "required": ["url"],
+        },
+    },
+}
+
+mcp_tools_payload = {
+    "jsonrpc": "2.0",
+    "id": 2,
+    "result": {"tools": []},
+}

xiaozhi_sdk-0.1.0/xiaozhi_sdk/iot.py

@@ -0,0 +1,50 @@
+import aiohttp
+import json
+
+from xiaozhi_sdk.config import OTA_URL
+
+USER_AGENT = "XiaoXhi-SDK/1.0"
+
+
+class OtaDevice(object):
+
+    def __init__(self, mac_addr: str, client_id: str, serial_number: str = ""):
+        self.mac_addr = mac_addr
+        self.client_id = client_id
+        self.serial_number = serial_number
+
+    async def activate_device(self):
+        header = {
+            "user-agent": USER_AGENT,
+            "Device-Id": self.mac_addr,
+            "Client-Id": self.client_id,
+            "Content-Type": "application/json",
+            "serial-number": self.serial_number,
+        }
+        payload = {
+            "application": {"version": "1.0.0"},
+            "board": {
+                "type": "xiaozhi-sdk-box",
+                "name": "xiaozhi-sdk-main",
+            },
+        }
+        async with aiohttp.ClientSession() as session:
+            async with session.post(OTA_URL, headers=header, data=json.dumps(payload)) as response:
+                data = await response.json()
+                return data
+
+    async def check_activate(self, challenge: str):
+        url = OTA_URL + "/activate"
+        header = {
+            "user-agent": USER_AGENT,
+            "Device-Id": self.mac_addr,
+            "Client-Id": self.client_id,
+            "Content-Type": "application/json",
+        }
+        payload = {
+            "serial_number": self.serial_number,
+            "challenge": challenge,
+        }
+        async with aiohttp.ClientSession() as session:
+            async with session.post(url, headers=header, data=json.dumps(payload)) as response:
+                return response.status == 200

xiaozhi_sdk-0.1.0/xiaozhi_sdk/mcp.py

@@ -0,0 +1,75 @@
+import json
+
+import requests
+
+from xiaozhi_sdk.config import VL_URL
+from xiaozhi_sdk.data import mcp_initialize_payload, mcp_tools_payload, mcp_tool_conf
+
+
+class McpTool(object):
+
+    def __init__(self):
+        self.session_id = ""
+        self.vl_token = ""
+        self.websocket = None
+        self.tool_func = {}
+
+    def get_mcp_json(self, payload: dict):
+        return json.dumps({"session_id": self.session_id, "type": "mcp", "payload": payload})
+
+    def _build_response(self, request_id: str, content: str, is_error: bool = False):
+        return self.get_mcp_json({
+            "jsonrpc": "2.0",
+            "id": request_id,
+            "result": {
+                "content": [{"type": "text", "text": content}],
+                "isError": is_error,
+            },
+        })
+
+    async def analyze_image(self, img_byte: bytes, question: str = "这张图片里有什么？"):
+        headers = {"Authorization": f"Bearer {self.vl_token}"}
+        files = {"file": ("camera.jpg", img_byte, "image/jpeg")}
+        payload = {"question": question}
+
+        response = requests.post(VL_URL, files=files, data=payload, headers=headers)
+        return response.json()
+
+    async def mcp_tool_call(self, mcp_json: dict):
+        tool_name = mcp_json["params"]["name"]
+        tool_func = self.tool_func[tool_name]
+        
+        if tool_name == "take_photo":
+            res = await self.analyze_image(tool_func(None), mcp_json["params"]["arguments"]["question"])
+        else:
+            res = tool_func(mcp_json["params"]["arguments"])
+        
+        content = json.dumps(res, ensure_ascii=False)
+        return self._build_response(mcp_json["id"], content)
+
+    async def mcp(self, data: dict):
+        payload = data["payload"]
+        method = payload["method"]
+
+        if method == "initialize":
+            self.vl_token = payload["params"]["capabilities"]["vision"]["token"]
+            mcp_initialize_payload["id"] = payload["id"]
+            await self.websocket.send(self.get_mcp_json(mcp_initialize_payload))
+
+        elif method == "tools/list":
+            mcp_tools_payload["id"] = payload["id"]
+            for name, func in self.tool_func.items():
+                if func:
+                    mcp_tool_conf[name]["name"] = name
+                    mcp_tools_payload["result"]["tools"].append(mcp_tool_conf[name])
+
+            await self.websocket.send(self.get_mcp_json(mcp_tools_payload))
+
+        elif method == "tools/call":
+            print("tools/call", payload)
+            tool_name = payload["params"]["name"]
+            if not self.tool_func.get(tool_name):
+                raise Exception("Tool not found")
+
+            mcp_res = await self.mcp_tool_call(payload)
+            await self.websocket.send(mcp_res)

xiaozhi_sdk-0.1.0/xiaozhi_sdk/opus.py

@@ -0,0 +1,59 @@
+import os
+from xiaozhi_sdk import INPUT_SERVER_AUDIO_SAMPLE_RATE
+
+# 设置 opus 库路径
+os.environ["DYLD_LIBRARY_PATH"] = "/opt/homebrew/lib:" + os.environ.get("DYLD_LIBRARY_PATH", "")
+os.environ["LIBRARY_PATH"] = "/opt/homebrew/lib:" + os.environ.get("LIBRARY_PATH", "")
+import av
+import numpy as np
+import opuslib
+
+
+class AudioOpus:
+
+    def __init__(self, sample_rate, channels):
+        self.sample_rate = sample_rate
+        self.channels = channels
+
+        # 创建 Opus 编码器
+        self.opus_encoder = opuslib.Encoder(
+            fs=sample_rate, channels=channels, application=opuslib.APPLICATION_VOIP  # 采样率  # 单声道  # 语音应用
+        )
+
+        # 创建 Opus 解码器
+        self.opus_decoder = opuslib.Decoder(
+            fs=INPUT_SERVER_AUDIO_SAMPLE_RATE,  # 采样率
+            channels=1,  # 单声道
+        )
+
+        self.resampler = av.AudioResampler(format="s16", layout="mono", rate=sample_rate)
+
+    async def pcm_to_opus(self, pcm):
+        pcm_array = np.frombuffer(pcm, dtype=np.int16)
+        pcm_bytes = pcm_array.tobytes()
+        return self.opus_encoder.encode(pcm_bytes, 960)
+
+    async def change_sample_rate(self, pcm_array):
+        if self.sample_rate == INPUT_SERVER_AUDIO_SAMPLE_RATE:
+            return pcm_array.reshape(1, 960)
+
+        c = int(self.sample_rate / INPUT_SERVER_AUDIO_SAMPLE_RATE)
+        frame = av.AudioFrame.from_ndarray(np.array(pcm_array).reshape(1, -1), format="s16", layout="mono")
+        frame.sample_rate = INPUT_SERVER_AUDIO_SAMPLE_RATE  # Assuming input is 16kHz
+        resampled_frames = self.resampler.resample(frame)
+        samples = resampled_frames[0].to_ndarray().flatten()
+        new_frame = av.AudioFrame.from_ndarray(
+            samples.reshape(1, -1),
+            format="s16",
+            layout="mono",
+        )
+        new_frame.sample_rate = self.sample_rate
+        new_samples = new_frame.to_ndarray().flatten()
+        arr_padded = np.pad(new_samples, (0, 960 * c - new_samples.shape[0]), mode="constant", constant_values=0)
+        return arr_padded.reshape(c, 960)
+
+    async def opus_to_pcm(self, opus):
+        pcm_data = self.opus_decoder.decode(opus, 960)
+        pcm_array = np.frombuffer(pcm_data, dtype=np.int16)
+        samples = await self.change_sample_rate(pcm_array)
+        return samples

xiaozhi_sdk-0.1.0/xiaozhi_sdk/utils.py

@@ -0,0 +1,23 @@
+import wave
+
+
+def get_wav_info(file_path):
+    with wave.open(file_path, "rb") as wav_file:
+        return wav_file.getframerate(), wav_file.getnchannels()
+
+def read_audio_file(file_path):
+    """
+    读取音频文件并通过yield返回PCM流
+    
+    Args:
+        file_path (str): 音频文件路径
+        
+    Yields:
+        bytes: PCM音频数据块
+    """
+    with wave.open(file_path, "rb") as wav_file:
+        while True:
+            pcm = wav_file.readframes(960)  # 每次读取960帧（60ms的音频数据）
+            if not pcm:
+                break
+            yield pcm

xiaozhi_sdk-0.1.0/xiaozhi_sdk.egg-info/PKG-INFO

@@ -0,0 +1,58 @@
+Metadata-Version: 2.4
+Name: xiaozhi-sdk
+Version: 0.1.0
+Summary: A short description of your package
+Home-page: https://github.com/dairoot/xiaozhi-sdk
+Author: dairoot
+Author-email: 623815825@qq.com
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+Requires-Dist: numpy
+Requires-Dist: requests>=2.32.1
+Requires-Dist: sounddevice>=0.4.2
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
+
+# 小智SDK (XiaoZhi SDK)
+
+[![Python Version](https://img.shields.io/badge/python-3.8+-blue.svg)](https://www.python.org/downloads/)
+[![License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
+[![PyPI](https://img.shields.io/badge/pypi-xiaozhi--sdk-blue.svg)](https://pypi.org/project/xiaozhi-sdk/)
+
+一个用于连接和控制小智智能设备的Python SDK，支持实时音频通信、MCP工具集成和设备管理功能。
+
+
+## 📦 安装
+
+```bash
+pip install xiaozhi-sdk
+```
+
+
+## 🚀 快速开始
+
+### 命令行使用
+
+最简单的使用方式是通过命令行连接设备：
+
+```bash
+# 查看帮助信息
+python -m xiaozhi_sdk -h
+
+# 连接设备（需要提供MAC地址）
+python -m xiaozhi_sdk 00:11:22:33:44:55
+```
+
+### 编程使用
+...
+

xiaozhi_sdk-0.1.0/xiaozhi_sdk.egg-info/SOURCES.txt

@@ -0,0 +1,17 @@
+README.md
+setup.py
+tests/test_pic.py
+tests/test_xiaozhi.py
+xiaozhi_sdk/__init__.py
+xiaozhi_sdk/__main__.py
+xiaozhi_sdk/config.py
+xiaozhi_sdk/data.py
+xiaozhi_sdk/iot.py
+xiaozhi_sdk/mcp.py
+xiaozhi_sdk/opus.py
+xiaozhi_sdk/utils.py
+xiaozhi_sdk.egg-info/PKG-INFO
+xiaozhi_sdk.egg-info/SOURCES.txt
+xiaozhi_sdk.egg-info/dependency_links.txt
+xiaozhi_sdk.egg-info/requires.txt
+xiaozhi_sdk.egg-info/top_level.txt

xiaozhi_sdk-0.1.0/xiaozhi_sdk.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

xiaozhi_sdk-0.1.0/xiaozhi_sdk.egg-info/requires.txt

@@ -0,0 +1,3 @@
+numpy
+requests>=2.32.1
+sounddevice>=0.4.2

xiaozhi_sdk-0.1.0/xiaozhi_sdk.egg-info/top_level.txt

@@ -0,0 +1 @@
+xiaozhi_sdk