smartpi 0.1.41__py3-none-any.whl → 0.1.42__py3-none-any.whl

smartpi/__init__.py

@@ -4,5 +4,5 @@ from .base_driver import P1, P2, P3, P4, P5, P6, M1, M2, M3, M4, M5, M6
 __all__ = ["base_driver","gui","ultrasonic","touch_sensor","temperature","humidity","light_sensor","color_sensor","motor","servo","led","flash",
            "P1", "P2", "P3", "P4", "P5", "P6", "M1", "M2", "M3", "M4", "M5", "M6"]
 
-__version__ = "0.1.41"
+__version__ = "0.1.42"
 

smartpi/ai_asr.py

@@ -798,7 +798,7 @@ def init(appid: str, secret_id: str, secret_key: str, voice_format: int = 8, sam
          hotword_id: str = "", result_type: int = 0, slice_size: int = 960, 
          vad_silence_timeout: int = 10000, vad_pause_timeout: int = 3000, 
          vad_mini_volume: int = 1000, vad_mini_length: int = 500, debug: bool = False,
-         vad_type: str = "energy", vad_threshold: float = 0.7, 
+         vad_threshold: float = 0.7, 
          sentence_callback=None, engine_model_type: str = "16k_zh", 
          interruption_callback=None) -> bool:
     """初始化语音识别
@@ -817,7 +817,6 @@ def init(appid: str, secret_id: str, secret_key: str, voice_format: int = 8, sam
         vad_mini_volume (int): 最小音量阈值，默认1000
         vad_mini_length (int): 最小语音长度(毫秒)，默认500
         debug (bool): 是否开启调试模式，默认False
-        vad_type (str): VAD类型，默认"energy"（目前只支持energy silero不稳定慎用！）
         vad_threshold (float): VAD阈值，默认0.7
         sentence_callback (callable): 句子识别完成回调函数
         engine_model_type (str): 引擎模型类型，默认"16k_zh"
@@ -839,28 +838,30 @@ def init(appid: str, secret_id: str, secret_key: str, voice_format: int = 8, sam
             except ImportError:
                 print("警告: ai_vad模块未找到")
                 ai_vad = None
-        
+    
         if ai_vad is not None:
             try:
                 vad = ai_vad.VADProviderBase.create_vad_instance(
-                    vad_type, threshold=vad_threshold
+                    vad_mini_volume, vad_mini_length,
+                    vad_pause_timeout, vad_silence_timeout,
+                    vad_threshold
                 )
-                print(f"VAD实例创建成功，类型: {vad_type}，阈值: {vad_threshold}")
+                print("VAD初始化成功，使用EnergyVADProvider")
             except Exception as e:
                 print(f"VAD实例创建失败: {e}")
                 vad = None
-        
-        _ai_asr_instance = TencentSpeechRecognizer(
-            appid=appid,
-            secret_id=secret_id,
-            secret_key=secret_key,
-            engine_model_type=engine_model_type,
-            sentence_callback=sentence_callback,
-            vad=vad,
-            interruption_callback=interruption_callback
-        )
-        print("语音识别初始化成功")
-        return True
+            
+            _ai_asr_instance = TencentSpeechRecognizer(
+                appid=appid,
+                secret_id=secret_id,
+                secret_key=secret_key,
+                engine_model_type=engine_model_type,
+                sentence_callback=sentence_callback,
+                vad=vad,
+                interruption_callback=interruption_callback
+            )
+            print("语音识别初始化成功")
+            return True
     except Exception as e:
         print(f"语音识别初始化失败: {e}")
         import traceback

smartpi/ai_vad.py

@@ -3,143 +3,27 @@ import sys
 import os
 import time
 import numpy as np
-import torch
-
-# 设置环境变量以优化性能
-os.environ['OMP_NUM_THREADS'] = '1'
-os.environ['KMP_AFFINITY'] = 'disabled'
 
 class VADProviderBase:
     """VAD provider base class"""
     @staticmethod
     def create_vad_instance(
-        class_name: str = "silero", 
-        model_dir=None, 
-        threshold=0.5, 
-        min_silence_duration_ms=800,
-        energy_threshold=0.02
+        vad_mini_volume=1000, 
+        vad_mini_length=500,
+        vad_pause_timeout=3000, 
+        vad_silence_timeout=10000,
+        threshold=0.02
     ):
-        """Factory method to create VAD instance"""
-        if class_name == "silero":
-            try:
-                return SileroVADProvider(
-                    model_dir=model_dir, 
-                    threshold=threshold, 
-                    min_silence_duration_ms=min_silence_duration_ms
-                )
-            except Exception as e:
-                print(f"Failed to create SileroVAD instance: {e}")
-                print("Falling back to EnergyVADProvider...")
-                return EnergyVADProvider(
-                    threshold=energy_threshold, 
-                    min_silence_duration_ms=min_silence_duration_ms
-                )
-        elif class_name == "energy":
-            return EnergyVADProvider(
-                threshold=energy_threshold, 
-                min_silence_duration_ms=min_silence_duration_ms
-            )
-        raise ValueError(f"Unsupported VAD type: {class_name}")
-
-class SileroVADProvider(VADProviderBase):
-    """Silero VAD provider"""
-    def __init__(self, model_dir=None, threshold=0.5, min_silence_duration_ms=800):
-        # 获取模型目录，默认为当前文件所在目录下的models/snakers4_silero-vad
-        if not model_dir:
-            model_dir = os.path.join(os.path.dirname(os.path.abspath(__file__)), "models/snakers4_silero-vad")
-        
-        print(f"Initializing SileroVAD: model_dir={model_dir}, threshold={threshold}, min_silence_duration_ms={min_silence_duration_ms}")
-        
-        # 加载模型
-        try:
-            # 尝试检查torchaudio是否安装
-            import importlib.util
-            torchaudio_available = importlib.util.find_spec('torchaudio') is not None
-            
-            if torchaudio_available:
-                # 如果有torchaudio，使用默认的PyTorch模型
-                self.model, self.utils = torch.hub.load(
-                    repo_or_dir=model_dir,
-                    source="local",
-                    model="silero_vad",
-                    force_reload=False,
-                )
-            else:
-                # 如果没有torchaudio，尝试使用ONNX版本
-                print("torchaudio not found, trying ONNX version...")
-                self.model, self.utils = torch.hub.load(
-                    repo_or_dir=model_dir,
-                    source="local",
-                    model="silero_vad",
-                    onnx=True,
-                    force_reload=False,
-                )
-        except Exception as e:
-            print(f"Failed to load Silero model (both PyTorch and ONNX versions): {e}")
-            # 降级到能量检测
-            raise Exception("Model loading failed, will use energy detection")
-        
-        # 初始化VAD配置，使用默认值作为备选
-        self.vad_threshold = float(threshold) if threshold else 0.5
-        self.silence_threshold_ms = int(min_silence_duration_ms) if min_silence_duration_ms else 800
-        print("SileroVAD初始化成功")
-    
-    def is_vad(self, conn, audio_data):
-        """检测音频数据中的语音活动"""
-        try:
-            # 确保conn对象有必要的属性
-            if not hasattr(conn, 'client_audio_buffer'):
-                conn.client_audio_buffer = []
-            if not hasattr(conn, 'client_have_voice'):
-                conn.client_have_voice = False
-            if not hasattr(conn, 'client_have_voice_last_time'):
-                conn.client_have_voice_last_time = 0
-            if not hasattr(conn, 'client_voice_stop'):
-                conn.client_voice_stop = False
-            
-            # 将新数据加入缓冲区
-            conn.client_audio_buffer.extend(audio_data)
-            
-            # 处理缓冲区中的完整帧（每次处理512采样点，每个采样点2字节）
-            client_have_voice = False
-            while len(conn.client_audio_buffer) >= 512 * 2:
-                # 提取前512个采样点（1024字节）
-                chunk = bytes(conn.client_audio_buffer[:512 * 2])
-                conn.client_audio_buffer = conn.client_audio_buffer[512 * 2:]
-                
-                # 转换为模型需要的格式
-                audio_int16 = np.frombuffer(chunk, dtype=np.int16)
-                audio_float32 = audio_int16.astype(np.float32) / 32768.0
-                
-                # 检测语音活动
-                with torch.no_grad():
-                    speech_prob = self.model(torch.from_numpy(audio_float32), 16000).item()
-                
-                client_have_voice = speech_prob >= self.vad_threshold
-                
-                # 更新VAD状态
-                if client_have_voice:
-                    conn.client_have_voice = True
-                    conn.client_have_voice_last_time = time.time() * 1000
-                    conn.client_voice_stop = False
-                elif conn.client_have_voice:
-                    # 检查静默时间
-                    stop_duration = time.time() * 1000 - conn.client_have_voice_last_time
-                    if stop_duration >= self.silence_threshold_ms:
-                        conn.client_have_voice = False
-                        conn.client_voice_stop = True
-            
-            return client_have_voice
-        except Exception as e:
-            print(f"VAD detection error: {e}")
-            return False
-    
-    def reset_states(self):
-        """Reset VAD states"""
-        pass  # States are now stored in conn object, no need to reset here
+        """Factory method to create VAD instance - always returns EnergyVADProvider"""
+        # 忽略vad_mini_volume, vad_mini_length, vad_silence_timeout参数
+        # 将vad_pause_timeout作为min_silence_duration_ms传递
+        return EnergyVADProvider(
+            threshold=threshold, 
+            min_silence_duration_ms=vad_pause_timeout
+        )
 
 class EnergyVADProvider(VADProviderBase):
-    """Audio energy-based VAD provider (fallback solution)"""
+    """Audio energy-based VAD provider"""
     def __init__(self, threshold=0.02, min_silence_duration_ms=300):
         print(f"Initializing EnergyVAD: threshold={threshold}, min_silence_duration_ms={min_silence_duration_ms}")
         # 初始化VAD配置，使用默认值作为备选

smartpi/onnx_pose_workflow.py

@@ -5,7 +5,7 @@ import onnx
 import json
 import cv2
 import time  # 用于实时耗时计算
-from lib.posenet_utils import get_posenet_output
+from posenet_utils import get_posenet_output
 
 
 class PoseWorkflow:

smartpi/rknn_pose_workflow.py

@@ -5,7 +5,7 @@ import json
 import os
 import time
 from rknnlite.api import RKNNLite  # RKNN核心库
-from lib.posenet_utils import get_posenet_output  # 姿态关键点提取逻辑（需支持返回3个值）
+from posenet_utils import get_posenet_output  # 姿态关键点提取逻辑（需支持返回3个值）
 
 
 class PoseWorkflow:

{smartpi-0.1.41.dist-info → smartpi-0.1.42.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: smartpi
-Version: 0.1.41
+Version: 0.1.42
 Summary: A library use for H2-RCU
 Author: ZMROBO
 Classifier: Programming Language :: Python :: 3

{smartpi-0.1.41.dist-info → smartpi-0.1.42.dist-info}/RECORD

@@ -1,9 +1,9 @@
-smartpi/__init__.py,sha256=ioz08SpgjM2Df14M57OTfaAqxYbI-A6h2BeKEQ--8GI,356
+smartpi/__init__.py,sha256=7AdxZqsEjhv8Ht_bLXOoF3T3jGZ0TEnG-SVYdjs4NR0,356
 smartpi/_gui.py,sha256=ij-6HZAEIwdy_hvU7f0NkyQjx_-eephijlKbGUhf8Uo,2177
-smartpi/ai_asr.py,sha256=eoF5z4oRJWpSk13tqO2M2N7ZBCkNWHPhQDt4-uUo_ko,40079
+smartpi/ai_asr.py,sha256=wxh_1Klh8vJottAt19jq3qpOmM_Cw4DHQEPKpsHGhmE,40063
 smartpi/ai_llm.py,sha256=-khBK2PMwbYUDvbaCTRCktE4dFloqfai4mHI-V8GEXM,36751
 smartpi/ai_tts.py,sha256=OjAJs3XOykiufXuEDiaD0coKsWjknPdhwRpRzGLDZKU,36061
-smartpi/ai_vad.py,sha256=RJBieDmH5A0FITfULhpcn6Vpy_NhVXK5s416D-Lm2nU,8983
+smartpi/ai_vad.py,sha256=9J1xxuKF8GvojkL_JuXv7_xE8bQAo5D49Xasi8piKK8,3630
 smartpi/base_driver.py,sha256=nGkGp7cAgeNo7BijHMJKHfdVZKIKuZNTx9GqLFPP5TI,31486
 smartpi/camera.py,sha256=AVpZsMpW-24CP3BOfarjmRawMJdTOZY7Crq7FeLOqb4,3292
 smartpi/color_sensor.py,sha256=ckIXD81YnqPo6nENAnipNp3gY12FJ235QKj0e8Cul9E,521
@@ -20,13 +20,13 @@ smartpi/motor.py,sha256=DvFzREEzyRafGmSCagU6ASeoE8fnAsKYI4oYMrkXXJc,5351
 smartpi/move.py,sha256=s1ZnkFtp6SCBnxhBxp6qQjGbifdsY5hjVCwlarTsZto,6688
 smartpi/onnx_hand_workflow.py,sha256=ZCoaWC6GygZSrhM6jhsuB6qmQ6GiAFFrso6rKAGmue8,8157
 smartpi/onnx_image_workflow.py,sha256=-saM_NxR6yDz06xlWZOvHf6cq3zmtOCFhCyZTGqvuOk,6188
-smartpi/onnx_pose_workflow.py,sha256=7hoZ31XfZRAbgmdQbgfK-xePniMa5mDEggV12F-Uq5c,20970
+smartpi/onnx_pose_workflow.py,sha256=w0zGlXXBgIcxwjLF7eYHPEePqRw3KXvKQUqzVd_UrzQ,20966
 smartpi/onnx_text_workflow.py,sha256=6l9MTT2T1-rNye3_dSHLI2U749Z94aoRdkSe6CNXfHw,7191
 smartpi/onnx_voice_workflow.py,sha256=jkMFzy3RUnLo8LZAuCUfsS3YCJWSZzZuiE4RFoQ2HZw,17440
 smartpi/posenet_utils.py,sha256=o3scK41Eqvftav4y4vp6_6HinQWNCLeLpArXAzqQ-7s,8983
 smartpi/rknn_hand_workflow.py,sha256=wsVN_PYP9M-1AFaN4yqrGksUBoamYfujW0nQq4nv3kU,10160
 smartpi/rknn_image_workflow.py,sha256=4lTtcdmQ9KN5WiEnHayvqAd-dA0tiap5YXIqAMn5SoI,18444
-smartpi/rknn_pose_workflow.py,sha256=LA6tXOI81R1IQhQvgBWLGV_I8Qa-ROUgXqj3kTEMfmc,27840
+smartpi/rknn_pose_workflow.py,sha256=T09F9vcQ6-QG3PlFRpCI4AuTPzEl_s0C38wfMD21Ujo,27836
 smartpi/rknn_text_workflow.py,sha256=KNBSetj3tmlLxdZOm0yzbiDnjH8S5191fuxh5Mi-uCY,9632
 smartpi/rknn_voice_workflow.py,sha256=T8iRQWPtJYXqoHIZH2FiT1WLxwN3HQg4D-mg-5KvYdA,16326
 smartpi/servo.py,sha256=0p09Jk-IVk5nLXz2AqFvytiYSSe4sMxdy1FaNMQijoY,5770
@@ -70,7 +70,7 @@ smartpi/text_gte_model/config/tokenizer_config.json,sha256=w5RiDifbeIYy6vyGX5v94
 smartpi/text_gte_model/config/vocab.txt,sha256=oi9hP3uz_8h8XoHNh6rgLnVdJbIEm75zKoSKM8HzsC8,84758
 smartpi/text_gte_model/gte/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 smartpi/text_gte_model/gte/gte_model.onnx,sha256=XXYg6TUhzOx1SqAhp6ePDU0QgeK6DQEqHATMuQQJCNE,30468366
-smartpi-0.1.41.dist-info/METADATA,sha256=446DzYjpDU_MxEdwmQyiyx7pf6K-OAQIOW3sek2gnZU,614
-smartpi-0.1.41.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
-smartpi-0.1.41.dist-info/top_level.txt,sha256=PoLhUCmWAiQUg5UeN2fS-Y1iQyBbF2rdUlizXtpHGRQ,8
-smartpi-0.1.41.dist-info/RECORD,,
+smartpi-0.1.42.dist-info/METADATA,sha256=FxIXEB2WMTsKc8Fs8ukVfdnD3--vqy7fOEXFT5AQLq8,614
+smartpi-0.1.42.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
+smartpi-0.1.42.dist-info/top_level.txt,sha256=PoLhUCmWAiQUg5UeN2fS-Y1iQyBbF2rdUlizXtpHGRQ,8
+smartpi-0.1.42.dist-info/RECORD,,