接口默认返回100条数据

bin/test.py

@@ -46,7 +46,7 @@ def load_test_data():
 
 def evaluation(result_detail):
 
-    top_n_list = [1, 3, 5, 10]
+    top_n_list = [1, 3, 5, 10, 20, 50, 100]
     metrics = {
         'all_test_cnt': 0,
         'all_recall_cnt': 0,
@@ -69,7 +69,7 @@ def evaluation(result_detail):
             for top_n, counselor in enumerate(rd['recall_counselors']):
                 if counselor['counselor'] == rd['supplier_id']:
                     from_info = counselor['from'].split(' ')
-                    if from_info[0] == 'top_50':
+                    if from_info[0] == 'top_100':
                         metrics['default_recall_cnt'] += 1
                     else:
                         from_id = from_info[1]
@@ -201,7 +201,7 @@ if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     
     parser.add_argument('--mode', default=0, type=int, help='模式：0-推荐的咨询师列表去重（默认，与实际线上一样）；1-推荐的咨询师列表没有去重')
-    parser.add_argument('--k', default=10, type=int, help='召回相似用户的数量')
+    parser.add_argument('--k', default=20, type=int, help='召回相似用户的数量')
     parser.add_argument('--top_n', default=5, type=int, help='每个相似用户召回的咨询师数量')
     parser.add_argument('--max_test', default=0, type=int, help='最多测试数据量')
     

src/core/order_data_manager.py

@@ -63,7 +63,7 @@ class OrderDataManager():
         """
         构建索引
             用户-咨询师 索引
-            top50 咨询师列表 用于冷启动
+            top100 咨询师列表 用于冷启动
         """
         self.logger.info('')
         self.logger.info('开始构建 用户-咨询师 索引')
@@ -100,14 +100,14 @@ class OrderDataManager():
         
         # 订单最多的咨询师
         supplier_cnter = Counter(df['supplier_id'])
-        top50_supplier = []
-        for key, _ in supplier_cnter.most_common(50):
-            top50_supplier.append(str(key))
-
-        self.logger.info('top50 订单量咨询师统计完成')
-        with open(os.path.join(self.local_file_dir, 'top50_supplier.txt'), 'w', encoding='utf-8') as f:
-            f.write('\n'.join(top50_supplier))
-        self.logger.info('top50 订单量咨询师列表已保存')
+        top100_supplier = []
+        for key, _ in supplier_cnter.most_common(100):
+            top100_supplier.append(str(key))
+
+        self.logger.info('top100 订单量咨询师统计完成')
+        with open(os.path.join(self.local_file_dir, 'top100_supplier.txt'), 'w', encoding='utf-8') as f:
+            f.write('\n'.join(top100_supplier))
+        self.logger.info('top100 订单量咨询师列表已保存')
 
 
 if __name__ == '__main__':

src/core/recommender.py

@@ -24,7 +24,7 @@ class Recommender():
 
 class UserCFRecommender(Recommender):
 
-    def __init__(self, top_n=5, k=10, is_use_db=True) -> None:
+    def __init__(self, top_n=5, k=20, is_use_db=True) -> None:
         super().__init__()
         # 召回 top_n 个相似用户
         self.top_n = top_n
@@ -57,7 +57,7 @@ class UserCFRecommender(Recommender):
         with open(os.path.join(self.local_file_dir, 'user_doctor_index.json'), encoding='utf-8') as f:
             order_user_counselor_index = json.load(f)
 
-        with open(os.path.join(self.local_file_dir, 'top50_supplier.txt'), 'r', encoding='utf-8') as f:
+        with open(os.path.join(self.local_file_dir, 'top100_supplier.txt'), 'r', encoding='utf-8') as f:
             default_counselor = [line.strip() for line in f]
 
         self.order_user_embedding = order_user_embedding
@@ -66,7 +66,7 @@ class UserCFRecommender(Recommender):
         self.default_counselor = [{
             'counselor': str(user),
             'score': 1 - 0.01 * index,
-            'from': 'top_50',
+            'from': 'top_100',
         } for index, user in enumerate(default_counselor)]
 
         self.index = faiss.IndexFlatL2(len(self.order_user_embedding[0]))
@@ -91,7 +91,7 @@ class UserCFRecommender(Recommender):
         return self.manager.profile_to_embedding(user_profile)
 
 
-    def _recommend_top(self, size=50):
+    def _recommend_top(self, size=100):
         return self.default_counselor[:size]