Update src/streamlit_app.py

src/streamlit_app.py

@@ -1,129 +1,115 @@
 import streamlit as st
 import cv2
 import numpy as np
+from streamlit_webrtc import webrtc_streamer, VideoProcessorBase
 import mediapipe as mp
-from PIL import Image
+import av
+from collections import deque
 
-# Initialize MediaPipe
+# Инициализация MediaPipe Pose
 mp_pose = mp.solutions.pose
 mp_drawing = mp.solutions.drawing_utils
 
-# Configure page
-st.set_page_config(layout="wide")
-st.title("📷 Posture Analysis App")
-
-# Initialize session state
-if 'camera_on' not in st.session_state:
-    st.session_state.camera_on = False
-if 'posture_status' not in st.session_state:
-    st.session_state.posture_status = "Please enable camera"
-
-def analyze_posture(image):
-    """Analyze posture using MediaPipe"""
-    with mp_pose.Pose(
-        min_detection_confidence=0.5,
-        min_tracking_confidence=0.5,
-        model_complexity=1
-    ) as pose:
-        
-        image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
-        results = pose.process(image_rgb)
-        
-        annotated_image = image.copy()
+class PoseProcessor(VideoProcessorBase):
+    def __init__(self):
+        self.pose = mp_pose.Pose(min_detection_confidence=0.5,
+                                min_tracking_confidence=0.5,
+                                model_complexity=1)
+        self.status_deque = deque(maxlen=1)  # Для передачи текста в UI
+
+    def analyze_posture(self, landmarks, image_shape):
+        h, w, _ = image_shape
+        left_shoulder = landmarks.landmark[mp_pose.PoseLandmark.LEFT_SHOULDER]
+        right_shoulder = landmarks.landmark[mp_pose.PoseLandmark.RIGHT_SHOULDER]
+        left_hip = landmarks.landmark[mp_pose.PoseLandmark.LEFT_HIP]
+        right_hip = landmarks.landmark[mp_pose.PoseLandmark.RIGHT_HIP]
+        left_ear = landmarks.landmark[mp_pose.PoseLandmark.LEFT_EAR]
+        right_ear = landmarks.landmark[mp_pose.PoseLandmark.RIGHT_EAR]
+        nose = landmarks.landmark[mp_pose.PoseLandmark.NOSE]
+
+        sitting = left_hip.y < left_shoulder.y + 0.1 or right_hip.y < right_shoulder.y + 0.1
+
+        messages = []
+
+        head_forward = (left_ear.y > left_shoulder.y + 0.1 or right_ear.y > right_shoulder.y + 0.1) and \
+                       (nose.y > left_shoulder.y or nose.y > right_shoulder.y)
+        if head_forward:
+            messages.append("• Голова наклонена вперед (текстовая шея)")
+
+        shoulders_rounded = left_shoulder.x > left_hip.x + 0.05 or right_shoulder.x < right_hip.x - 0.05
+        if shoulders_rounded:
+            messages.append("• Плечи ссутулены (округлены вперед)")
+
+        shoulder_diff = abs(left_shoulder.y - right_shoulder.y)
+        hip_diff = abs(left_hip.y - right_hip.y)
+        if shoulder_diff > 0.05 or hip_diff > 0.05:
+            messages.append("• Наклон в сторону (несимметричная осанка)")
+
+        if sitting and (left_hip.y < left_shoulder.y + 0.15 or right_hip.y < right_shoulder.y + 0.15):
+            messages.append("• Таз наклонен вперед (сидя)")
+
+        if messages:
+            report = [
+                f"**{'Сидя' if sitting else 'Стоя'} - обнаружены проблемы:**",
+                *messages,
+                "\n**Рекомендации:**",
+                "• Держите голову прямо, уши должны быть над плечами",
+                "• Отведите плечи назад и вниз",
+                "• Держите спину прямой, избегайте наклонов в стороны",
+                "• При сидении опирайтесь на седалищные бугры"
+            ]
+        else:
+            report = [
+                f"**Отличная осанка ({'сидя' if sitting else 'стоя'})!**",
+                "Все ключевые точки находятся в правильном положении.",
+                "\n**Совет:**",
+                "• Продолжайте следить за осанкой в течение дня"
+            ]
+
+        return "\n\n".join(report)
+
+    def recv(self, frame: av.VideoFrame) -> av.VideoFrame:
+        img = frame.to_ndarray(format="bgr24")
+        img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+
+        results = self.pose.process(img_rgb)
         if results.pose_landmarks:
+            # Рисуем ключевые точки
             mp_drawing.draw_landmarks(
-                annotated_image, 
-                results.pose_landmarks, 
-                mp_pose.POSE_CONNECTIONS,
+                img, results.pose_landmarks, mp_pose.POSE_CONNECTIONS,
                 mp_drawing.DrawingSpec(color=(0, 255, 0), thickness=2, circle_radius=2),
                 mp_drawing.DrawingSpec(color=(0, 0, 255), thickness=2)
             )
-            posture_status = check_posture(results.pose_landmarks, image.shape)
+            status = self.analyze_posture(results.pose_landmarks, img.shape)
+            self.status_deque.append(status)
         else:
-            posture_status = "No pose detected"
-        
-        return annotated_image, posture_status
-
-def check_posture(landmarks, image_shape):
-    """Analyze body landmarks and generate posture report"""
-    h, w, _ = image_shape
-    
-    # Get key points
-    left_shoulder = landmarks.landmark[mp_pose.PoseLandmark.LEFT_SHOULDER]
-    right_shoulder = landmarks.landmark[mp_pose.PoseLandmark.RIGHT_SHOULDER]
-    left_hip = landmarks.landmark[mp_pose.PoseLandmark.LEFT_HIP]
-    right_hip = landmarks.landmark[mp_pose.PoseLandmark.RIGHT_HIP]
-    left_ear = landmarks.landmark[mp_pose.PoseLandmark.LEFT_EAR]
-    right_ear = landmarks.landmark[mp_pose.PoseLandmark.RIGHT_EAR]
-    nose = landmarks.landmark[mp_pose.PoseLandmark.NOSE]
-    
-    # Determine posture
-    sitting = left_hip.y < left_shoulder.y + 0.1 or right_hip.y < right_shoulder.y + 0.1
-    
-    messages = []
-    
-    # Forward head check
-    head_forward = (left_ear.y > left_shoulder.y + 0.1 or right_ear.y > right_shoulder.y + 0.1) and \
-                   (nose.y > left_shoulder.y or nose.y > right_shoulder.y)
-    if head_forward:
-        messages.append("• Forward head tilt detected")
-    
-    # Shoulders check
-    shoulders_rounded = left_shoulder.x > left_hip.x + 0.05 or right_shoulder.x < right_hip.x - 0.05
-    if shoulders_rounded:
-        messages.append("• Rounded shoulders detected")
-    
-    # Side tilt check
-    shoulder_diff = abs(left_shoulder.y - right_shoulder.y)
-    hip_diff = abs(left_hip.y - right_hip.y)
-    if shoulder_diff > 0.05 or hip_diff > 0.05:
-        messages.append("• Body leaning to one side")
-    
-    # Generate report
-    if messages:
-        report = [
-            f"**{'Sitting' if sitting else 'Standing'} posture issues:**",
-            *messages,
-            "\n**Recommendations:**",
-            "• Keep head straight",
-            "• Pull shoulders back",
-            "• Maintain straight back",
-            "• Sit on sitting bones"
-        ]
-    else:
-        report = [
-            f"**Excellent {'sitting' if sitting else 'standing'} posture!**",
-            "\n**Tips:**",
-            "• Continue monitoring your posture"
-        ]
-    
-    return "\n\n".join(report)
+            self.status_deque.append("Ключевые точки не обнаружены")
+
+        return av.VideoFrame.from_ndarray(img, format="bgr24")
 
 def main():
-    col1, col2 = st.columns([2, 1])
-    
-    with col1:
-        st.header("Camera View")
-        
-        # Camera toggle
-        if st.button("Enable Camera" if not st.session_state.camera_on else "Disable Camera"):
-            st.session_state.camera_on = not st.session_state.camera_on
-            st.rerun()
-        
-        if st.session_state.camera_on:
-            # Placeholder for local camera implementation
-            demo_image = np.zeros((480, 640, 3), dtype=np.uint8)
-            st.image(demo_image, caption="Camera feed placeholder")
-            st.warning("For full camera functionality, please run locally")
-            
-            # Simulate analysis
-            if st.button("Analyze Current Frame"):
-                _, posture_status = analyze_posture(demo_image)
-                st.session_state.posture_status = posture_status
-    
-    with col2:
-        st.header("Analysis Results")
-        st.markdown(st.session_state.posture_status)
+    st.set_page_config(layout="wide")
+    st.title("📷 Анализатор осанки с веб-камеры (WebRTC)")
+    st.write("Приложение анализирует вашу осанку в реальном времени с помощью камеры браузера.")
+
+    # Запускаем WebRTC стример с нашим процессором
+    webrtc_ctx = webrtc_streamer(
+        key="pose-analyzer",
+        video_processor_factory=PoseProcessor,
+        media_stream_constraints={"video": True, "audio": False},
+        async_processing=True
+    )
+
+    # Выводим текст анализа из процесса
+    if webrtc_ctx.video_processor:
+        # Получаем последнюю доступную строку с анализом
+        if webrtc_ctx.video_processor.status_deque:
+            analysis_text = webrtc_ctx.video_processor.status_deque[-1]
+        else:
+            analysis_text = "Ожидание видео и анализа..."
+
+        st.markdown("### Анализ осанки:")
+        st.markdown(analysis_text)
 
 if __name__ == "__main__":
     main()