Create streamlit_app.py

streamlit_app.py

@@ -0,0 +1,44 @@
+import streamlit as st
+import cv2, numpy as np, base64, io, os
+import librosa, joblib
+from deepface import DeepFace
+
+# 1) 加载所有模型
+@st.cache_resource
+def load_models():
+    DeepFace.analyze(img_path=np.zeros((224,224,3),dtype=np.uint8),
+                     actions=['emotion'], enforce_detection=False)
+    voice_clf = joblib.load("voice_model.joblib")
+    return voice_clf
+
+voice_clf = load_models()
+
+st.title("📱 即時多模態情緒分析")
+
+# 2) 即时人脸
+st.header("🖼 實時人臉情緒")
+img_data = st.camera_input("對準鏡頭")
+if img_data is not None:
+    arr = np.frombuffer(img_data.read(), np.uint8)
+    img = cv2.imdecode(arr, cv2.IMREAD_COLOR)
+    res = DeepFace.analyze(img, actions=["emotion"], enforce_detection=False)
+    emo = (res[0] if isinstance(res,list) else res).get("dominant_emotion","unknown")
+    st.write("情緒：", emo)
+
+# 3) 語音上傳
+st.header("🎤 上傳語音情緒")
+audio = st.file_uploader("請上傳 WAV 音檔", type=["wav"])
+if audio is not None:
+    with open("tmp.wav","wb") as f: f.write(audio.getbuffer())
+    y, sr = librosa.load("tmp.wav", sr=None)
+    mf = np.mean(librosa.feature.mfcc(y=y, sr=sr, n_mfcc=13).T,axis=0)
+    emo = voice_clf.predict([mf])[0]
+    st.write("情緒：", emo)
+
+# 4) 文字輸入
+st.header("📝 輸入文字情緒")
+txt = st.text_input("打些文字…")
+if txt:
+    # copy 你的 analyze_text_fn
+    emo = analyze_text_fn(txt)
+    st.write("情緒：", emo)