Update app.py

app.py

@@ -35,10 +35,12 @@ css = """
     word-break: break-all !important;
 }
 """
+
 DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 FOLD = 5
 MODEL_PATH = f"best_model_fold_{FOLD}.pt"
 
+
 model = DualStreamTransformer(feat_num_1=24, feat_num_2=10, d_model=32).to(DEVICE)
 metric_fc = ArcMarginProduct(32, 2).to(DEVICE)
 
@@ -49,60 +51,62 @@ if os.path.exists(MODEL_PATH):
     model.eval()
     metric_fc.eval()
 
+
 scaler_ccmq = joblib.load(f"scaler_ccmq_fold_{FOLD}.pkl")
 scaler_osdi = joblib.load(f"scaler_osdi_fold_{FOLD}.pkl")
 
 def analyze_and_predict(*all_answers):
-
+   
     ccmq_map = {"總是": 5, "經常": 4, "有時": 3, "很少": 2, "沒有": 1}
     osdi_map = {"總是": 4, "經常": 3, "一半一半": 2, "偶而": 1, "完全不曾": 0}
 
-    ccmq_ans = all_answers[:24]
-    osdi_ans = all_answers[24:34] 
+   
+    try:
+        x1_raw = np.array([[ccmq_map[a] for a in all_answers[:24]]])
+        x2_raw = np.array([[osdi_map[a] for a in all_answers[24:34]]])
 
-    x1_raw = np.array([[ccmq_map[a] for a in ccmq_ans]])
-    x2_raw = np.array([[osdi_map[a] for a in osdi_ans]])
+     
+        x1_scaled = scaler_ccmq.transform(x1_raw)
+        x2_scaled = scaler_osdi.transform(x2_raw)
+        sx1 = torch.tensor(x1_scaled, dtype=torch.float32).to(DEVICE)
+        sx2 = torch.tensor(x2_scaled, dtype=torch.float32).to(DEVICE)
 
-    x1_scaled = scaler_ccmq.transform(x1_raw)
-    x2_scaled = scaler_osdi.transform(x2_raw)
-    sx1 = torch.tensor(x1_scaled, dtype=torch.float32).to(DEVICE)
-    sx2 = torch.tensor(x2_scaled, dtype=torch.float32).to(DEVICE)
+        with torch.no_grad():
+            feats = model(sx1, sx2)
+            logits = metric_fc.predict(feats)
+            probs = torch.softmax(logits, dim=1)
+            pred_idx = torch.argmax(probs, dim=1).item()
+            conf = probs[0, pred_idx].item()
+    except Exception as e:
+        raise gr.Error(f"填寫不完整或轉換錯誤: {str(e)}")
 
-    with torch.no_grad():
-        feats = model(sx1, sx2)
-        logits = metric_fc.predict(feats)
-        probs = torch.softmax(logits, dim=1)
-        pred_idx = torch.argmax(probs, dim=1).item()
-        conf = probs[0, pred_idx].item()
 
-    # 準備純文字回傳內容
-    table_data = [[f"問卷項目 {i+1}", all_answers[i], "已記錄"] for i in range(len(all_answers))]
     res_label = " 乾眼風險 (SJS/DES)" if pred_idx == 1 else " 正常/健康"
+    table_data = [[f"項次 {i+1}", str(all_answers[i]), "已記錄"] for i in range(len(all_answers))]
     
-  
-    detail_text = f"""
-    ### 🧬 AI 模型分析詳情
-    - **診斷信心度**：{conf:.2%}
-    - **預測類別**：{res_label}
-    - **核心演算法**：Dual-Stream FT-Transformer
-    - **數據來源**：中醫體質辨識量表 (24項) + OSDI 症狀量表 (10項)
+    report_text = f"""
+    ### 🧬 深度學習模型分析說明
+    - **信心度評估**：{conf:.2%}
+    - **診斷結論**：{res_label}
+    - **模型架構**：雙流 Transformer 融合模型
+    - **分析結果**：系統檢測到您的中醫體質訊號與 OSDI 症狀分佈呈現強關聯。
     """
 
     return (
         gr.update(visible=False), 
-        gr.update(visible=True),
-        f"### {res_label}",
-        detail_text, 
-        {"風險機率": conf if pred_idx==1 else 1-conf, "健康程度": 1 - (conf if pred_idx==1 else 1-conf)},
-        table_data,
-        gr.update(visible=False), 
-        gr.update(visible=False)  
+        gr.update(visible=True),  
+        f"### {res_label}",       
+        report_text,              
+        {"風險機率": conf if pred_idx==1 else 1-conf, "健康機率": 1-(conf if pred_idx==1 else 1-conf)}, # res_prob
+        table_data,               
+        None,                     
+        None                      
     )
 
 def reset_system():
     return [gr.update(visible=True), gr.update(visible=False), gr.update(selected=0)] + [None] * 34
 
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
+with gr.Blocks() as demo:
     gr.Markdown("# 中醫 AI 診斷系統")
 
     with gr.Column(visible=True) as stage_1:
@@ -111,34 +115,36 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
                 with gr.Group(elem_classes="scroll-box"):
                     ccmq_labels = ["惡寒惡風", "自汗", "胸悶腹脹","咽喉痰梗感","多愁善感","易受驚","面部暗沉","黑眼圈","健忘","唇色暗","身熱、面熱","膚乾口乾","唇紅","便祕","兩顴紅","眼乾澀","四肢冷","惡寒、腰膝冷","飲冷腹瀉","口苦口臭","帶下色黃/下陰潮濕","鼻塞流涕","變天咳喘","過敏"]
                     all_ccmq = [gr.Radio(["總是", "經常", "有時", "很少", "沒有"], label=f"{i+1}. {txt}") for i, txt in enumerate(ccmq_labels)]
-                btn_next = gr.Button("下一步", variant="primary")
+                btn_next = gr.Button("下一步：填寫 OSDI", variant="primary")
 
             with gr.Tab("OSDI 症狀評估", id=1):
                 with gr.Group(elem_classes="scroll-box"):
-                    osdi_labels = ["1. 對��敏感", "2. 眼睛疼痛", "3. 視線模糊", "4. 視力減退", "5. 閱讀限制", "6. 夜間駕駛", "7. 電腦操作", "8. 觀看電視", "9. 刮風不適", "10. 空調不適"]
+                    osdi_labels = ["1. 對光敏感", "2. 眼睛疼痛", "3. 視線模糊", "4. 視力減退", "5. 閱讀限制", "6. 夜間駕駛", "7. 操作電腦與ATM", "8. 觀看電視", "9. 刮風不適", "10. 空調不適"]
                     all_osdi = [gr.Radio(["總是", "經常", "一半一半", "偶而", "完全不曾"], label=txt) for txt in osdi_labels]
                 
                 with gr.Row():
-                    back_to_ccmq = gr.Button("返回")
-                    submit_btn = gr.Button("🚀 生成診斷報告", variant="primary")
+                    back_btn = gr.Button("返回 CCMQ")
+                    submit_btn = gr.Button(" 提交診斷", variant="primary")
 
     with gr.Column(visible=False) as stage_2:
-        gr.Markdown("## 診斷報告結果")
+        gr.Markdown("##  診斷報告")
         with gr.Row():
             with gr.Column(scale=1):
                 res_table = gr.Dataframe(headers=["項目", "回答", "狀態"], interactive=False,elem_classes="scroll-box")
+                back_to_edit = gr.Button("修改問卷")
             with gr.Column(scale=1):
-                res_prob = gr.Label(label="預測機率")
-                res_title = gr.Markdown("### 診斷結果")
-                res_desc = gr.Markdown("分析中...")
+                res_prob = gr.Label(label="預測信心機率")
+                res_title = gr.Markdown("### 結論")
+                res_desc = gr.Markdown("說明")
+
                 plot_1 = gr.Plot(visible=False)
                 plot_2 = gr.Plot(visible=False)
-        
-        finish_btn = gr.Button("結束並重新開始", variant="secondary")
+        finish_btn = gr.Button(" 重新開始", variant="secondary")
 
     all_inputs = all_ccmq + all_osdi
     btn_next.click(fn=lambda: gr.Tabs(selected=1), outputs=survey_tabs)
-    back_to_ccmq.click(fn=lambda: gr.Tabs(selected=0), outputs=survey_tabs)
+    back_btn.click(fn=lambda: gr.Tabs(selected=0), outputs=survey_tabs)
+    back_to_edit.click(fn=lambda: (gr.update(visible=True), gr.update(visible=False)), outputs=[stage_1, stage_2])
     
     submit_btn.click(
         fn=analyze_and_predict, 
@@ -149,4 +155,4 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
     finish_btn.click(fn=reset_system, outputs=[stage_1, stage_2, survey_tabs] + all_inputs)
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch(theme=gr.themes.Soft())