调试

tools/generate_content_user_feature.py

@@ -21,6 +21,7 @@ import time
 import pymysql
 from sklearn.preprocessing import LabelEncoder
 from elasticsearch import Elasticsearch
+from sklearn.preprocessing import OneHotEncoder
 
 
 class Generate_Feature_Info(object):
@@ -71,6 +72,16 @@ class Generate_Feature_Info(object):
             logging.error("catch exception,err_msg:%s" % traceback.format_exc())
             return None
 
+    @classmethod
+    def get_tagv3_label_onehot(cls,label_encoder):
+        try:
+            onehot_encoder = OneHotEncoder().fit(label_encoder)
+            return onehot_encoder
+        except:
+            logging.error("catch exception,err_msg:%s" % traceback.format_exc())
+            return None
+
+
     @classmethod
     def generate_content_feature_to_redis(cls,label_encoder,content_type="diary"):
         try:
@@ -207,9 +218,10 @@ if __name__ == "__main__":
     tagv3_name_set = Generate_Feature_Info.get_tagv3_word_list()
 
     label_encoder = Generate_Feature_Info.get_tagv3_label_encode(tagv3_name_list=list(tagv3_name_set))
+    onehot_encoder = Generate_Feature_Info.get_tagv3_label_onehot(label_encoder=label_encoder)
 
-    device_id="868771031984211"
-    Generate_Feature_Info.generate_user_feature_to_redis(device_id,label_encoder)
+    # device_id="868771031984211"
+    # Generate_Feature_Info.generate_user_feature_to_redis(device_id,label_encoder)
     # Generate_Feature_Info.generate_content_feature_to_redis(label_encoder)
     #
     #
@@ -238,6 +250,9 @@ if __name__ == "__main__":
     #     LinUCB.update_linucb_info(user_features=user_feature_list,reward=1,content_id=diary_id,
     #                               redis_name_linucb_matrix=linucb_matrix_redis_name,redis_cli=redis_client)
     # diary_fd.close()
-    # test_val_list = ["切开双眼皮"]
-    #
-    # label_results_list = label_encoder.transform(test_val_list)
+    test_val_list = ["切开双眼皮"]
+
+    label_results_list = label_encoder.transform(test_val_list)
+    label_one_hot_results_list = onehot_encoder.transform(label_results_list)
+
+    print(label_one_hot_results_list)
\ No newline at end of file