From 8be35bfebfeb51aa0137ed4f69d8ffb761a89c61 Mon Sep 17 00:00:00 2001
From: tian <11429339@qq.com>
Date: Wed, 10 Jun 2026 10:44:08 +0800
Subject: [PATCH] =?UTF-8?q?chore:=20=E7=A7=BB=E9=99=A4=E5=9B=9E=E9=80=80?=
 =?UTF-8?q?=E9=80=BB=E8=BE=91=EF=BC=8C=E7=B2=BE=E7=AE=80=E9=A2=84=E6=B5=8B?=
 =?UTF-8?q?=E8=B7=AF=E7=94=B1;=20=E6=B8=85=E7=90=86=E6=97=A7=E6=9E=B6?=
 =?UTF-8?q?=E6=9E=84PyTorch=E6=A8=A1=E5=9E=8B?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 src/routes.py | 31 ++++++++-----------------------
 1 file changed, 8 insertions(+), 23 deletions(-)

diff --git a/src/routes.py b/src/routes.py
index daa4fe4..0d2308e 100644
--- a/src/routes.py
+++ b/src/routes.py
@@ -107,39 +107,24 @@ def predict():
         model_id = data.get('model_id')  # 可选：指定模型ID
         logger.info(f"Received prediction request for equipment type: {equipment_type}")
         
-        # 获取可用模型（默认R²最高的；可选指定model_id）
+        # 获取最优模型（默认R²最高的；可选指定model_id）
         with get_db_connection() as conn:
             cursor = conn.cursor()
             if model_id:
-                cursor.execute("""
-                    SELECT * FROM trained_models 
-                    WHERE id = ? AND is_active = TRUE
-                """, (model_id,))
-                models = cursor.fetchall()
+                cursor.execute("SELECT * FROM trained_models WHERE id = ? AND is_active = TRUE", (model_id,))
             else:
                 cursor.execute("""
                     SELECT * FROM trained_models 
-                    WHERE equipment_type = ? 
-                    AND model_type != 'pls'
-                    AND is_active = TRUE
-                    ORDER BY r2_score DESC
+                    WHERE equipment_type = ? AND model_type != 'pls' AND is_active = TRUE
+                    ORDER BY r2_score DESC LIMIT 1
                 """, (equipment_type,))
-                models = cursor.fetchall()
+            model = cursor.fetchone()
         
-        if not models:
-            return jsonify({'error': '未找到可用的模型'}), 404
+        if not model:
+            return jsonify({'error': '未找到可用的模型，请先训练'}), 404
         
-        # 逐个尝试加载，跳过架构不兼容的模型
         predictor = CostPredictor()
-        for model in models:
-            try:
-                prediction = predictor.predict(data, model)
-                break
-            except ValueError as e:
-                logger.warning(f"Model {model['model_name']} failed: {e}, trying next...")
-                continue
-        else:
-            return jsonify({'error': '所有激活模型加载失败，请重新训练'}), 500
+        prediction = predictor.predict(data, model)
         
         # 返回预测结果
         return jsonify({