Update app.py

app.py

@@ -1,40 +1,89 @@
 # app.py
-import os
-# —— 把 DeepFace 缓存目录指向可写的 /tmp/.deepface
-os.environ["DEEPFACE_HOME"] = "/tmp/.deepface"
-os.makedirs(os.environ["DEEPFACE_HOME"], exist_ok=True)
-
 import gradio as gr
-import cv2
 import numpy as np
+import joblib, io
+import librosa
 from deepface import DeepFace
 
-def face_emotion(frame: np.ndarray) -> str:
+# —— 1. 预加载模型 —— 
+# DeepFace 热身 + 载入语音模型
+audio_model = joblib.load("src/voice_model.joblib")
+# 你也可以包一层 try/except
+
+def analyze_face(frame: np.ndarray):
     """
-    接收 gr.Camera 给出的 RGB ndarray，
-    转成 BGR 后交给 DeepFace 分析情绪。
+    输入：摄像头采到的 RGB numpy 数组
+    输出：DeepFace 分析出的 dominant_emotion（字符串）
     """
-    # RGB → BGR
-    bgr = cv2.cvtColor(frame, cv2.COLOR_RGB2BGR)
+    # DeepFace.analyze 接受 RGB np.array
     res = DeepFace.analyze(
-        bgr,
+        img_path=frame,
         actions=['emotion'],
         enforce_detection=False
     )
-    # DeepFace 支持 list 或 dict 返回
+    # 兼容 dict / list 返回
     if isinstance(res, list):
         emo = res[0].get('dominant_emotion', 'unknown')
     else:
         emo = res.get('dominant_emotion', 'unknown')
     return emo
 
-# —— Gradio 前端 —— 
-with gr.Blocks() as demo:
-    gr.Markdown("## 📱 多模態即時情緒分析（示範：即時人臉情緒）")
-    camera = gr.Camera(label="請對準鏡頭", type="numpy")
-    output = gr.Textbox(label="偵測到的情緒")
-    # 实时流：fps 可以调低一点，减轻服务器压力
-    camera.stream(face_emotion, camera, output, fps=5)
+def analyze_audio(wav_file):
+    """
+    输入：上传的 wav（二进制）
+    输出：语音情绪分类标签
+    """
+    data = wav_file.read()
+    y, sr = librosa.load(io.BytesIO(data), sr=None)
+    mfccs = librosa.feature.mfcc(y=y, sr=sr, n_mfcc=13)
+    mf = np.mean(mfccs.T, axis=0)
+    return audio_model.predict([mf])[0]
+
+def analyze_text(txt):
+    """
+    简单的基于关键词的中文情绪分析
+    """
+    if any(w in text for w in ["開心","快樂","愉快","喜悅","歡喜","興奮","歡","高興"]): return "happy"
+    if any(w in text for w in ["生氣","憤怒","不爽","發火","火大","氣憤"]): return "angry"
+    if any(w in text for w in ["傷心","難過","哭","難受","心酸","憂","悲","哀","痛苦","慘","愁"]): return "sad"
+    if any(w in text for w in ["驚訝","意外","嚇","驚詫","詫異","訝異","好奇"]):       return "surprise"
+    if any(w in text for w in ["怕","恐懼","緊張","懼","膽怯","畏"]):       return "fear"
+    return "neutral"
+
+with gr.Blocks(title="多模態即時情緒分析") as demo:
+    gr.Markdown("## 📱 多模態即時情緒分析")
+
+    tabs = gr.Tabs()
+    with tabs:
+        with gr.TabItem("🔴 Face（即時）"):
+            gr.Markdown("⚠️ Spaces 無法直接打開攝像頭，請本機 `python app.py` 測試；手機／電腦瀏覽器可用以下方式：")
+            camera = gr.Image(
+                source="webcam",
+                type="numpy",
+                label="請對準鏡頭"
+            )
+            face_out = gr.Textbox(label="偵測結果")
+            camera.change(fn=analyze_face, inputs=camera, outputs=face_out)
 
-if __name__ == "__main__":
-    demo.launch()
+        with gr.TabItem("🎤 上傳語音"):
+            wav = gr.File(
+                label="請選擇 .wav 音檔",
+                file_types=["wav"]
+            )
+            audio_out = gr.Textbox(label="偵測結果")
+            wav.upload(fn=analyze_audio, inputs=wav, outputs=audio_out)
+
+        with gr.TabItem("⌨️ 文本輸入"):
+            txt = gr.Textbox(
+                label="在此輸入文字",
+                placeholder="輸入想要分析的句子…"
+            )
+            text_btn = gr.Button("開始分析")
+            text_out = gr.Textbox(label="偵測結果")
+            text_btn.click(fn=analyze_text, inputs=txt, outputs=text_out)
+
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=7860,
+        share=False
+    )