xiaozhi-sdk 0.1.1__tar.gz → 0.2.0__tar.gz

{xiaozhi_sdk-0.1.1/xiaozhi_sdk.egg-info → xiaozhi_sdk-0.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: xiaozhi-sdk
-Version: 0.1.1
+Version: 0.2.0
 Summary: 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。
 Author-email: dairoot <623815825@qq.com>
 License-Expression: MIT
@@ -22,6 +22,7 @@ Requires-Dist: click
 Requires-Dist: colorlog
 Requires-Dist: soundfile>=0.13.1
 Requires-Dist: pydub>=0.25.1
+Requires-Dist: pillow>=11.3.0
 Dynamic: license-file
 
 # 小智SDK (XiaoZhi SDK)

{xiaozhi_sdk-0.1.1 → xiaozhi_sdk-0.2.0}/pyproject.toml

@@ -23,6 +23,7 @@ dependencies = [
     "colorlog",
     "soundfile>=0.13.1",
     "pydub>=0.25.1",
+    "pillow>=11.3.0",
 ]
 classifiers = [
     "Programming Language :: Python :: 3",

xiaozhi_sdk-0.2.0/tests/test_wake_word.py

@@ -0,0 +1,31 @@
+import asyncio
+import os
+import sys
+
+import pytest
+
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), ".."))
+
+from xiaozhi_sdk import XiaoZhiWebsocket
+
+
+MAC_ADDR = "00:22:44:66:88:00"
+ota_url = None
+URL = None
+
+
+@pytest.mark.asyncio
+async def test_main():
+    is_end = asyncio.Event()
+    async def message_handler_callback(message):
+        if message.get("state") == "stop":
+            is_end.set()
+        print("message received:", message)
+
+    xiaozhi = XiaoZhiWebsocket(message_handler_callback, url=URL, ota_url=ota_url)
+    await xiaozhi.init_connection(MAC_ADDR)
+
+    await xiaozhi.send_wake_word("你是")
+
+    await asyncio.wait_for(is_end.wait(), timeout=20.0)
+    await xiaozhi.close()

{xiaozhi_sdk-0.1.1 → xiaozhi_sdk-0.2.0}/tests/test_xiaozhi.py

@@ -54,11 +54,13 @@ def mcp_tool_func():
     def mcp_set_volume(data) -> tuple[dict, bool]:
         return {}, False
 
-    return {
-        "set_volume": mcp_set_volume,
-        "get_device_status": mcp_get_device_status,
-        "take_photo": mcp_take_photo,
-    }
+    from xiaozhi_sdk.utils.mcp_tool import take_photo, get_device_status, set_volume
+
+    take_photo["tool_func"] = mcp_take_photo
+    get_device_status["tool_func"] = mcp_get_device_status
+    set_volume["tool_func"] = mcp_set_volume
+
+    return [take_photo, get_device_status, set_volume]
 
 
 async def message_handler_callback(message):
@@ -79,7 +81,8 @@ URL = None
 @pytest.mark.asyncio
 async def test_main():
     xiaozhi = XiaoZhiWebsocket(message_handler_callback, url=URL, ota_url=ota_url)
-    await xiaozhi.set_mcp_tool_callback(mcp_tool_func())
+
+    await xiaozhi.set_mcp_tool(mcp_tool_func())
     await xiaozhi.init_connection(MAC_ADDR)
 
     # # say hellow

{xiaozhi_sdk-0.1.1 → xiaozhi_sdk-0.2.0}/xiaozhi_sdk/__init__.py

@@ -1,3 +1,3 @@
-__version__ = "0.1.1"
+__version__ = "0.2.0"
 
 from xiaozhi_sdk.core import XiaoZhiWebsocket  # noqa

{xiaozhi_sdk-0.1.1 → xiaozhi_sdk-0.2.0}/xiaozhi_sdk/cli.py

@@ -1,4 +1,5 @@
 import asyncio
+import io
 import logging
 import time
 from collections import deque
@@ -8,6 +9,7 @@ import click
 import colorlog
 import numpy as np
 import sounddevice as sd
+from PIL import ImageGrab
 
 from xiaozhi_sdk import XiaoZhiWebsocket
 from xiaozhi_sdk.config import INPUT_SERVER_AUDIO_SAMPLE_RATE
@@ -22,19 +24,23 @@ logging.addLevelName(INFO1, "INFO1")
 logging.addLevelName(INFO2, "INFO2")
 logging.addLevelName(INFO3, "INFO3")
 
+
 # 为logger添加自定义方法
 def info1(self, message, *args, **kwargs):
     if self.isEnabledFor(INFO1):
         self._log(INFO1, message, args, **kwargs)
 
+
 def info2(self, message, *args, **kwargs):
     if self.isEnabledFor(INFO2):
         self._log(INFO2, message, args, **kwargs)
 
+
 def info3(self, message, *args, **kwargs):
     if self.isEnabledFor(INFO3):
         self._log(INFO3, message, args, **kwargs)
 
+
 # 将自定义方法添加到Logger类
 logging.Logger.info1 = info1
 logging.Logger.info2 = info2
@@ -69,20 +75,47 @@ is_playing_audio = False
 is_end = False
 
 
+def get_image_byte(data):
+    logger.info("请截图需要识别的内容:")
+    time.sleep(1)
+    for _ in range(20):
+        im = ImageGrab.grabclipboard()
+        if not im:
+            time.sleep(0.3)
+            continue
+        if im.mode == "RGBA":
+            im = im.convert("RGB")
+
+        byte_io = io.BytesIO()
+        im.save(byte_io, format="JPEG", quality=30)
+        # im.save("./test.jpg", format='JPEG', quality=30)
+
+        img_bytes = byte_io.getvalue()
+        logger.info("截图成功")
+
+        # if platform.system() == "Darwin":
+        #     subprocess.run("pbcopy", input=b"")
+
+        return img_bytes, False
+
+    logger.error("截图失败, 请在10秒内完成截图")
+    return "截图失败", True
+
+
 async def handle_message(message):
     """处理接收到的消息"""
     global is_end
-    if message["type"] == "stt": # 人类语音
+    if message["type"] == "stt":  # 人类语音
         logger.info1("human: %s", message["text"])
-    elif message["type"] == "tts" and message["state"] == "sentence_start": # AI语音
+    elif message["type"] == "tts" and message["state"] == "sentence_start":  # AI语音
         logger.info2("AI: %s", message["text"])
     elif message["type"] == "tts" and message["state"] == "stop":
         pass
         # logger.info2("播放结束")
         # logger.info("聆听中...")
-    elif message["type"] == "llm": # 表情
+    elif message["type"] == "llm":  # 表情
         logger.info3("emotion: %s", message["text"])
-    else: # 其他消息
+    else:  # 其他消息
         pass
         # logger.info("other: %s", message)
 
@@ -124,17 +157,23 @@ class XiaoZhiClient:
         self,
         url: Optional[str] = None,
         ota_url: Optional[str] = None,
+        wake_word: str = "",
     ):
         self.xiaozhi: Optional[XiaoZhiWebsocket] = None
         self.url = url
         self.ota_url = ota_url
         self.mac_address = ""
+        self.wake_word = wake_word
 
     async def start(self, mac_address: str, serial_number: str, license_key: str, enable_audio):
         """启动客户端连接"""
         self.mac_address = mac_address
-        self.xiaozhi = XiaoZhiWebsocket(handle_message, url=self.url, ota_url=self.ota_url, send_wake=True)
+        self.xiaozhi = XiaoZhiWebsocket(handle_message, url=self.url, ota_url=self.ota_url, wake_word=self.wake_word)
+        from xiaozhi_sdk.utils.mcp_tool import take_photo
+
+        take_photo["tool_func"] = get_image_byte
 
+        await self.xiaozhi.set_mcp_tool([take_photo])
         await self.xiaozhi.init_connection(
             self.mac_address, aec=False, serial_number=serial_number, license_key=license_key
         )
@@ -162,10 +201,12 @@ class XiaoZhiClient:
                 await self.xiaozhi.send_audio(pcm_data)
 
 
-async def run_client(mac_address: str, url: str, ota_url: str, serial_number: str, license_key: str, enable_audio: bool):
+async def run_client(
+    mac_address: str, url: str, ota_url: str, serial_number: str, license_key: str, enable_audio: bool, wake_word: str
+):
     """运行客户端的异步函数"""
     logger.debug("Recording... Press Ctrl+C to stop.")
-    client = XiaoZhiClient(url, ota_url)
+    client = XiaoZhiClient(url, ota_url, wake_word)
     await client.start(mac_address, serial_number, license_key, enable_audio)
 
     with sd.InputStream(callback=client.audio_callback, channels=1, samplerate=16000, blocksize=960):
@@ -179,9 +220,12 @@ async def run_client(mac_address: str, url: str, ota_url: str, serial_number: st
 @click.option("--serial_number", default="", help="设备的序列号")
 @click.option("--license_key", default="", help="设备的授权密钥")
 @click.option("--enable_audio", default=True, help="是否开启音频播放")
-def main(mac_address: str, url: str, ota_url: str, serial_number: str, license_key: str, enable_audio: bool):
+@click.option("--wake_word", default="", help="唤醒词")
+def main(
+    mac_address: str, url: str, ota_url: str, serial_number: str, license_key: str, enable_audio: bool, wake_word: str
+):
     """小智SDK客户端
 
     MAC_ADDRESS: 设备的MAC地址 (格式: XX:XX:XX:XX:XX:XX)
     """
-    asyncio.run(run_client(mac_address, url, ota_url, serial_number, license_key, enable_audio))
+    asyncio.run(run_client(mac_address, url, ota_url, serial_number, license_key, enable_audio, wake_word))

{xiaozhi_sdk-0.1.1 → xiaozhi_sdk-0.2.0}/xiaozhi_sdk/core.py

@@ -13,7 +13,6 @@ from xiaozhi_sdk.config import INPUT_SERVER_AUDIO_SAMPLE_RATE
 from xiaozhi_sdk.iot import OtaDevice
 from xiaozhi_sdk.mcp import McpTool
 from xiaozhi_sdk.utils import get_wav_info, read_audio_file, setup_opus
-from xiaozhi_sdk.utils.mcp_tool import async_mcp_play_music, async_search_custom_music
 
 setup_opus()
 from xiaozhi_sdk.opus import AudioOpus
@@ -30,14 +29,14 @@ class XiaoZhiWebsocket(McpTool):
         ota_url: Optional[str] = None,
         audio_sample_rate: int = 16000,
         audio_channels: int = 1,
-        send_wake: bool = False,
+        wake_word: str = "",
     ):
         super().__init__()
         self.url = url
         self.ota_url = ota_url
-        self.send_wake = send_wake
         self.audio_channels = audio_channels
         self.audio_opus = AudioOpus(audio_sample_rate, audio_channels)
+        self.wake_word = wake_word
 
         # 客户端标识
         self.client_id = str(uuid.uuid4())
@@ -62,10 +61,9 @@ class XiaoZhiWebsocket(McpTool):
         self.ota: Optional[OtaDevice] = None
         self.iot_task: Optional[asyncio.Task] = None
         self.wait_device_activated: bool = False
-        self.tool_func = {
-            "async_play_custom_music": async_mcp_play_music,
-            "async_search_custom_music": async_search_custom_music,
-        }
+
+        # mcp工具
+        self.mcp_tool_dict = {}
 
     async def _send_hello(self, aec: bool) -> None:
         """发送hello消息"""
@@ -122,7 +120,7 @@ class XiaoZhiWebsocket(McpTool):
             await self.websocket.send(opus_data)
         await self.send_silence_audio()
 
-    async def send_wake_word(self, wake_word: str = "你好，小智") -> None:
+    async def send_wake_word(self, wake_word: str) -> None:
         """发送唤醒词"""
         await self.websocket.send(
             json.dumps({"session_id": self.session_id, "type": "listen", "state": "detect", "text": wake_word})
@@ -175,9 +173,10 @@ class XiaoZhiWebsocket(McpTool):
                 )
                 logger.debug("[websocket] close")
 
-    async def set_mcp_tool_callback(self, tool_func: Dict[str, Callable[..., Any]]) -> None:
-        """设置MCP工具回调函数"""
-        self.tool_func.update(tool_func)
+    async def set_mcp_tool(self, mcp_tool_list) -> None:
+        """设置MCP工具"""
+        for mcp_tool in mcp_tool_list:
+            self.mcp_tool_dict[mcp_tool["name"]] = mcp_tool
 
     async def connect_websocket(self, websocket_token):
         """连接websocket"""
@@ -229,8 +228,8 @@ class XiaoZhiWebsocket(McpTool):
             self.iot_task = asyncio.create_task(self._activate_iot_device(license_key, ota_info))
             logger.debug("[IOT] 设备未激活")
 
-        if self.send_wake:
-            await self.send_wake_word()
+        if self.wake_word:
+            await self.send_wake_word(self.wake_word)
 
     async def send_audio(self, pcm: bytes) -> None:
         """发送音频数据"""

{xiaozhi_sdk-0.1.1 → xiaozhi_sdk-0.2.0}/xiaozhi_sdk/mcp.py

@@ -1,15 +1,33 @@
 import asyncio
+import copy
 import json
 import logging
+import time
+from typing import Any, Dict
 
 import numpy as np
 import requests
 
-from xiaozhi_sdk.utils.mcp_data import mcp_initialize_payload, mcp_tool_conf, mcp_tools_payload
 from xiaozhi_sdk.utils.mcp_tool import _get_random_music_info
 
 logger = logging.getLogger("xiaozhi_sdk")
 
+mcp_initialize_payload: Dict[str, Any] = {
+    "jsonrpc": "2.0",
+    "id": 1,
+    "result": {
+        "protocolVersion": "2024-11-05",
+        "capabilities": {"tools": {}},
+        "serverInfo": {"name": "", "version": "0.0.1"},
+    },
+}
+
+mcp_tools_payload: Dict[str, Any] = {
+    "jsonrpc": "2.0",
+    "id": 2,
+    "result": {"tools": []},
+}
+
 
 class McpTool(object):
 
@@ -18,8 +36,9 @@ class McpTool(object):
         self.explain_url = ""
         self.explain_token = ""
         self.websocket = None
-        self.tool_func = {}
+        self.mcp_tool_dict = {}
         self.is_playing = False
+        self.message_handler_callback = None
 
     def get_mcp_json(self, payload: dict):
         return json.dumps({"session_id": self.session_id, "type": "mcp", "payload": payload})
@@ -40,13 +59,16 @@ class McpTool(object):
         headers = {"Authorization": f"Bearer {self.explain_token}"}
         files = {"file": ("camera.jpg", img_byte, "image/jpeg")}
         payload = {"question": question}
+        init_time = time.time()
         try:
-            response = requests.post(self.explain_url, files=files, data=payload, headers=headers, timeout=5)
+            response = requests.post(self.explain_url, files=files, data=payload, headers=headers, timeout=8)
             res_json = response.json()
-        except Exception:
+        except Exception as e:
+            logger.error("[MCP] 图片解析 error: %s", e)
             return "网络异常", True
         if res_json.get("error"):
             return res_json, True
+        logger.debug("[MCP] 图片解析耗时：%s", time.time() - init_time)
         return res_json, False
 
     async def play_custom_music(self, tool_func, arguments):
@@ -60,11 +82,10 @@ class McpTool(object):
 
     async def mcp_tool_call(self, mcp_json: dict):
         tool_name = mcp_json["params"]["name"]
-        tool_func = self.tool_func[tool_name]
+        mcp_tool = self.mcp_tool_dict[tool_name]
         arguments = mcp_json["params"]["arguments"]
         try:
-            if tool_name == "async_play_custom_music":
-
+            if tool_name == "play_custom_music":
                 # v1 返回 url
                 music_info = await _get_random_music_info(arguments["id_list"])
                 if not music_info.get("url"):
@@ -83,14 +104,18 @@ class McpTool(object):
                 # v2 音频放到输出
                 # asyncio.create_task(self.play_custom_music(tool_func, arguments))
 
-            elif tool_name.startswith("async_"):
-                tool_res, is_error = await tool_func(arguments)
+            elif mcp_tool.get("is_async"):
+                tool_res, is_error = await mcp_tool["tool_func"](arguments)
             else:
-                tool_res, is_error = tool_func(arguments)
+                tool_res, is_error = mcp_tool["tool_func"](arguments)
         except Exception as e:
-            logger.error("[MCP] tool_func error: %s", e)
+            logger.error("[MCP] tool_name: %s, error: %s", tool_name, e)
             return self._build_response(mcp_json["id"], "工具调用失败", True)
 
+        if is_error:
+            logger.error("[MCP] tool_name: %s, error: %s", tool_name, tool_res)
+            return self._build_response(mcp_json["id"], "工具调用失败, {}".format(tool_res), True)
+
         if tool_name == "take_photo":
             tool_res, is_error = await self.analyze_image(tool_res, mcp_json["params"]["arguments"]["question"])
 
@@ -103,6 +128,7 @@ class McpTool(object):
 
         if method == "initialize":
             self.explain_url = payload["params"]["capabilities"]["vision"]["url"]
+            # self.explain_url = "http://82.157.143.133:8000/vision/explain"
             self.explain_token = payload["params"]["capabilities"]["vision"]["token"]
 
             mcp_initialize_payload["id"] = payload["id"]
@@ -117,19 +143,24 @@ class McpTool(object):
 
         elif method == "tools/list":
             mcp_tools_payload["id"] = payload["id"]
-            tool_list = []
-            for name, func in self.tool_func.items():
-                if func:
-                    tool_list.append(name)
-                    target_name = name.removeprefix("async_")
-                    mcp_tool_conf[target_name]["name"] = name
-                    mcp_tools_payload["result"]["tools"].append(mcp_tool_conf[target_name])
+            tool_name_list = []
+            mcp_tool_dict = copy.deepcopy(self.mcp_tool_dict)
+            for _, mcp_tool in mcp_tool_dict.items():
+                tool_name_list.append(mcp_tool["name"])
+                tool_func = mcp_tool.pop("tool_func", None)
+                if not tool_func:
+                    logger.error("[MCP] Tool %s has no tool_func", mcp_tool["name"])
+                    return
+                mcp_tool.pop("is_async", None)
+                mcp_tools_payload["result"]["tools"].append(mcp_tool)
+
             await self.websocket.send(self.get_mcp_json(mcp_tools_payload))
-            logger.debug("[MCP] 加载成功，当前可用工具列表为：%s", tool_list)
+            logger.debug("[MCP] 加载成功，当前可用工具列表为：%s", tool_name_list)
 
         elif method == "tools/call":
             tool_name = payload["params"]["name"]
-            if not self.tool_func.get(tool_name):
+
+            if not self.mcp_tool_dict.get(tool_name):
                 logger.warning("[MCP] Tool not found: %s", tool_name)
                 return
 

xiaozhi_sdk-0.2.0/xiaozhi_sdk/utils/mcp_tool.py

@@ -0,0 +1,185 @@
+import io
+import random
+
+import aiohttp
+import numpy as np
+
+
+async def async_search_custom_music(data) -> tuple[dict, bool]:
+    search_url = f"https://music-api.gdstudio.xyz/api.php?types=search&name={data['music_name']}&count=100&pages=1"
+
+    # 为搜索请求设置 10 秒超时
+    timeout = aiohttp.ClientTimeout(total=10)
+    async with aiohttp.ClientSession(timeout=timeout) as session:
+        async with session.get(search_url) as response:
+            response_json = await response.json()
+
+    music_list = []
+    first_music_list = []
+    other_music_list1 = []
+    other_music_list2 = []
+    for line in response_json:
+        if data.get("author_name") and data["author_name"] in line["artist"][0]:
+            first_music_list.append(line)
+        elif data.get("author_name") and (data["author_name"] in line["artist"] or data["author_name"] in line["name"]):
+            other_music_list1.append(line)
+        else:
+            other_music_list2.append(line)
+
+    if len(first_music_list) <= 10:
+        music_list = first_music_list
+        random.shuffle(other_music_list2)
+        music_list = music_list + other_music_list1[: 20 - len(music_list)]
+        music_list = music_list + other_music_list2[: 20 - len(music_list)]
+
+    # print(data)
+    # print("找到音乐，数量：", len(first_music_list), len(music_list))
+
+    if not music_list:
+        return {}, False
+    return {"message": "已找到歌曲", "music_list": music_list}, False
+
+
+async def _get_random_music_info(id_list: list) -> dict:
+    timeout = aiohttp.ClientTimeout(total=10)
+    async with aiohttp.ClientSession(timeout=timeout) as session:
+        random.shuffle(id_list)
+
+        for music_id in id_list:
+            url = f"https://music-api.gdstudio.xyz/api.php?types=url&id={music_id}&br=128"
+            async with session.get(url) as response:
+                res_json = await response.json()
+            if res_json.get("url"):
+                break
+
+    return res_json
+
+
+async def async_mcp_play_music(data) -> tuple[list, bool]:
+    try:
+        from pydub import AudioSegment
+    except ImportError:
+        return [], True
+
+    id_list = data["id_list"]
+    res_json = await _get_random_music_info(id_list)
+
+    if not res_json:
+        return [], False
+
+    pcm_list = []
+    buffer = io.BytesIO()
+    # 为下载音乐文件设置 60 秒超时（音乐文件可能比较大）
+    download_timeout = aiohttp.ClientTimeout(total=60)
+    async with aiohttp.ClientSession(timeout=download_timeout) as session:
+        async with session.get(res_json["url"]) as resp:
+            async for chunk in resp.content.iter_chunked(1024):
+                buffer.write(chunk)
+
+    buffer.seek(0)
+    audio = AudioSegment.from_mp3(buffer)
+    audio = audio.set_frame_rate(16000).set_channels(1).set_sample_width(2)  # 2 bytes = 16 bits
+    pcm_data = audio.raw_data
+
+    chunk_size = 960 * 2
+    for i in range(0, len(pcm_data), chunk_size):
+        chunk = pcm_data[i : i + chunk_size]
+
+        if chunk:  # 确保不添加空块
+            chunk = np.frombuffer(chunk, dtype=np.int16)
+            pcm_list.extend(chunk)
+
+    return pcm_list, False
+
+
+search_custom_music = {
+    "name": "search_custom_music",
+    "description": "Search music and get music IDs. Use this tool when the user asks to search or play music. This tool returns a list of music with their IDs, which are required for playing music. Args:\n  `music_name`: The name of the music to search\n  `author_name`: The name of the music author (optional)",
+    "inputSchema": {
+        "type": "object",
+        "properties": {"music_name": {"type": "string"}, "author_name": {"type": "string"}},
+        "required": ["music_name"],
+    },
+    "tool_func": async_search_custom_music,
+    "is_async": True,
+}
+
+play_custom_music = {
+    "name": "play_custom_music",
+    "description": "Play music using music IDs. IMPORTANT: You must call `search_custom_music` first to get the music IDs before using this tool. Use this tool after getting music IDs from search results. Args:\n  `id_list`: The id list of the music to play (obtained from search_custom_music results). The list must contain more than 2 music IDs, and the system will randomly select one to play.\n  `music_name`: The name of the music (obtained from search_custom_music results)",
+    "inputSchema": {
+        "type": "object",
+        "properties": {
+            "music_name": {"type": "string"},
+            "id_list": {"type": "array", "items": {"type": "string"}, "minItems": 3},
+        },
+        "required": ["music_name", "id_list"],
+    },
+    "tool_func": async_mcp_play_music,
+    "is_async": True,
+}
+
+stop_music = {
+    "name": "stop_music",
+    "description": "Stop playing music.",
+    "inputSchema": {"type": "object", "properties": {}},
+    "tool_func": None,
+}
+
+get_device_status = {
+    "name": "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": {}},
+    "tool_func": None,
+}
+
+set_volume = {
+    "name": "set_volume",
+    "description": "Set the volume of the audio speaker. If the current volume is unknown, you must call `get_device_status` tool first and then call this tool.",
+    "inputSchema": {
+        "type": "object",
+        "properties": {"volume": {"type": "integer", "minimum": 0, "maximum": 100}},
+        "required": ["volume"],
+    },
+    "tool_func": None,
+}
+
+set_brightness = {
+    "name": "set_brightness",
+    "description": "Set the brightness of the screen.",
+    "inputSchema": {
+        "type": "object",
+        "properties": {"brightness": {"type": "integer", "minimum": 0, "maximum": 100}},
+        "required": ["brightness"],
+    },
+    "tool_func": None,
+}
+
+set_theme = {
+    "name": "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"]},
+    "tool_func": None,
+}
+
+take_photo = {
+    "name": "take_photo",
+    "description": "Use this tool when the user asks you to look at something, take a picture, or solve a problem based on what is captured.\nArgs:\n`question`: A clear question or task you want to ask about the captured photo (e.g., identify objects, read text, explain content, or solve a math/logic problem).\nReturn:\n  A JSON object that provides the photo information, including answers, explanations, or problem-solving results if applicable.",
+    "inputSchema": {
+        "type": "object",
+        "properties": {"question": {"type": "string"}},
+        "required": ["question"],
+    },
+    "tool_func": None,
+}
+
+open_tab = {
+    "name": "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"],
+    },
+    "tool_func": None,
+}

{xiaozhi_sdk-0.1.1 → xiaozhi_sdk-0.2.0/xiaozhi_sdk.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: xiaozhi-sdk
-Version: 0.1.1
+Version: 0.2.0
 Summary: 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。
 Author-email: dairoot <623815825@qq.com>
 License-Expression: MIT
@@ -22,6 +22,7 @@ Requires-Dist: click
 Requires-Dist: colorlog
 Requires-Dist: soundfile>=0.13.1
 Requires-Dist: pydub>=0.25.1
+Requires-Dist: pillow>=11.3.0
 Dynamic: license-file
 
 # 小智SDK (XiaoZhi SDK)

{xiaozhi_sdk-0.1.1 → xiaozhi_sdk-0.2.0}/xiaozhi_sdk.egg-info/SOURCES.txt

@@ -14,6 +14,7 @@ file/opus/macos-x64-libopus.dylib
 file/opus/windows-opus.dll
 tests/test_iot.py
 tests/test_pic.py
+tests/test_wake_word.py
 tests/test_xiaozhi.py
 xiaozhi_sdk/__init__.py
 xiaozhi_sdk/__main__.py
@@ -39,5 +40,4 @@ xiaozhi_sdk/../file/opus/macos-arm64-libopus.dylib
 xiaozhi_sdk/../file/opus/macos-x64-libopus.dylib
 xiaozhi_sdk/../file/opus/windows-opus.dll
 xiaozhi_sdk/utils/__init__.py
-xiaozhi_sdk/utils/mcp_data.py
 xiaozhi_sdk/utils/mcp_tool.py

{xiaozhi_sdk-0.1.1 → xiaozhi_sdk-0.2.0}/xiaozhi_sdk.egg-info/requires.txt

@@ -10,3 +10,4 @@ click
 colorlog
 soundfile>=0.13.1
 pydub>=0.25.1
+pillow>=11.3.0

xiaozhi_sdk-0.1.1/xiaozhi_sdk/utils/mcp_data.py

@@ -1,79 +0,0 @@
-from typing import Any, Dict
-
-mcp_initialize_payload: Dict[str, Any] = {
-    "jsonrpc": "2.0",
-    "id": 1,
-    "result": {
-        "protocolVersion": "2024-11-05",
-        "capabilities": {"tools": {}},
-        "serverInfo": {"name": "", "version": "0.0.1"},
-    },
-}
-
-mcp_tool_conf: Dict[str, Dict[str, Any]] = {
-    "search_custom_music": {
-        "description": "Search music and get music IDs. Use this tool when the user asks to search or play music. This tool returns a list of music with their IDs, which are required for playing music. Args:\n  `music_name`: The name of the music to search\n  `author_name`: The name of the music author (optional)",
-        "inputSchema": {
-            "type": "object",
-            "properties": {"music_name": {"type": "string"}, "author_name": {"type": "string"}},
-            "required": ["music_name"],
-        },
-    },
-    "play_custom_music": {
-        "description": "Play music using music IDs. IMPORTANT: You must call `search_custom_music` first to get the music IDs before using this tool. Use this tool after getting music IDs from search results. Args:\n  `id_list`: The id list of the music to play (obtained from search_custom_music results). The list must contain more than 2 music IDs, and the system will randomly select one to play.\n  `music_name`: The name of the music (obtained from search_custom_music results)",
-        "inputSchema": {
-            "type": "object",
-            "properties": {
-                "music_name": {"type": "string"},
-                "id_list": {"type": "array", "items": {"type": "string"}, "minItems": 3},
-            },
-            "required": ["music_name", "id_list"],
-        },
-    },
-    "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 `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: Dict[str, Any] = {
-    "jsonrpc": "2.0",
-    "id": 2,
-    "result": {"tools": []},
-}

xiaozhi_sdk-0.1.1/xiaozhi_sdk/utils/mcp_tool.py

@@ -1,92 +0,0 @@
-import io
-import random
-
-import aiohttp
-import numpy as np
-
-
-async def async_search_custom_music(data) -> tuple[dict, bool]:
-    search_url = f"https://music-api.gdstudio.xyz/api.php?types=search&name={data['music_name']}&count=100&pages=1"
-
-    # 为搜索请求设置 10 秒超时
-    timeout = aiohttp.ClientTimeout(total=10)
-    async with aiohttp.ClientSession(timeout=timeout) as session:
-        async with session.get(search_url) as response:
-            response_json = await response.json()
-
-    music_list = []
-    first_music_list = []
-    other_music_list1 = []
-    other_music_list2 = []
-    for line in response_json:
-        if data.get("author_name") and data["author_name"] in line["artist"][0]:
-            first_music_list.append(line)
-        elif data.get("author_name") and (data["author_name"] in line["artist"] or data["author_name"] in line["name"]):
-            other_music_list1.append(line)
-        else:
-            other_music_list2.append(line)
-
-    if len(first_music_list) <= 10:
-        music_list = first_music_list
-        random.shuffle(other_music_list2)
-        music_list = music_list + other_music_list1[: 20 - len(music_list)]
-        music_list = music_list + other_music_list2[: 20 - len(music_list)]
-
-    # print(data)
-    # print("找到音乐，数量：", len(first_music_list), len(music_list))
-
-    if not music_list:
-        return {}, False
-    return {"message": "已找到歌曲", "music_list": music_list}, False
-
-
-async def _get_random_music_info(id_list: list) -> dict:
-    timeout = aiohttp.ClientTimeout(total=10)
-    async with aiohttp.ClientSession(timeout=timeout) as session:
-        random.shuffle(id_list)
-
-        for music_id in id_list:
-            url = f"https://music-api.gdstudio.xyz/api.php?types=url&id={music_id}&br=128"
-            async with session.get(url) as response:
-                res_json = await response.json()
-            if res_json.get("url"):
-                break
-
-    return res_json
-
-
-async def async_mcp_play_music(data) -> tuple[list, bool]:
-    try:
-        from pydub import AudioSegment
-    except ImportError:
-        return [], True
-
-    id_list = data["id_list"]
-    res_json = await _get_random_music_info(id_list)
-
-    if not res_json:
-        return [], False
-
-    pcm_list = []
-    buffer = io.BytesIO()
-    # 为下载音乐文件设置 60 秒超时（音乐文件可能比较大）
-    download_timeout = aiohttp.ClientTimeout(total=60)
-    async with aiohttp.ClientSession(timeout=download_timeout) as session:
-        async with session.get(res_json["url"]) as resp:
-            async for chunk in resp.content.iter_chunked(1024):
-                buffer.write(chunk)
-
-    buffer.seek(0)
-    audio = AudioSegment.from_mp3(buffer)
-    audio = audio.set_frame_rate(16000).set_channels(1).set_sample_width(2)  # 2 bytes = 16 bits
-    pcm_data = audio.raw_data
-
-    chunk_size = 960 * 2
-    for i in range(0, len(pcm_data), chunk_size):
-        chunk = pcm_data[i : i + chunk_size]
-
-        if chunk:  # 确保不添加空块
-            chunk = np.frombuffer(chunk, dtype=np.int16)
-            pcm_list.extend(chunk)
-
-    return pcm_list, False