调试

linucb/core/Linucb.py

@@ -7,7 +7,10 @@ from __future__ import print_function
 
 import numpy as np
 import redis
-# from libs.cache import redis_client
+import sys
+sys.path.append("/home/gmuser/gm_mab/")
+from libs.cache import redis_client
+
 import logging
 import traceback
 import json
@@ -17,9 +20,6 @@ import random
 import time
 
 
-redis_url = "redis://:ReDis!GmTx*0aN9@172.16.40.173:6379"
-redis_client = redis.StrictRedis.from_url(redis_url)
-
 class LinUCB:
     d = 6
     alpha = 0.01

tools/generate_content_user_feature.py

@@ -6,8 +6,10 @@ from __future__ import division
 from __future__ import print_function
 
 import numpy as np
-import redis
-# from libs.cache import redis_client
+import sys
+sys.path.append("/home/gmuser/gm_mab/")
+from libs.cache import redis_client
+
 import logging
 import traceback
 import json
@@ -17,7 +19,7 @@ import random
 import time
 import pymysql
 from sklearn.preprocessing import LabelEncoder
-
+from elasticsearch import Elasticsearch
 
 
 class Generate_Feature_Info(object):
@@ -27,8 +29,9 @@ class Generate_Feature_Info(object):
     password = "BJQaT9VzDcuPBqkd"
     database = "zhengxing"
 
-    # 新项目标签encode对象
-    tagv3_label_encoder = None
+    redis_name_content_linucb_feature_prefix = "strategy:linucb:feature:content_type:"
+    redis_name_user_linucb_feature_prefix = "strategy:linucb:feature:user"
+
 
     @classmethod
     def get_tagv3_word_list(cls):
@@ -67,8 +70,94 @@ class Generate_Feature_Info(object):
             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:
+            redis_name_content_linucb_feature = cls.redis_name_content_linucb_feature_prefix + content_type
+
+
+            es = Elasticsearch([
+                {
+                    'host': '172.16.31.17',
+                    'port': 9000,
+                }
+            ])
+            page = es.search(
+                index='gm-dbmw-diary-read',
+                doc_type='diary',
+                scroll='10m',
+                search_type='scan',
+                size=10,
+                body={
+                    "query": {
+                        "filtered": {
+                            "filter": {
+                                "bool": {
+                                    "must": [
+                                        {"term": {"is_online": True}},
+                                        {"terms": {"content_level": [5,6]}}
+                                    ]
+                                }
+                            }
+                        }
+                    },
+                    "_source": {"include": ["id", "tags_v3", "content_level"]}
+                }
+            )
+            sid = page['_scroll_id']
+            scroll_size = page['hits']['total']
+            while (scroll_size > 0):
+                try:
+                    page = es.scroll(scroll_id=sid, scroll='10m')
+                    sid = page['_scroll_id']
+                    scroll_size = len(page['hits']['hits'])
+
+                    for item in page['hits']['hits']:
+                        diary_id = item["_source"]["id"]
+                        tags_v3 = item["_source"]["tags_v3"]
+                        content_level = item["_source"]["content_level"] if item["_source"]["content_level"] else -1
 
+                        offi_tags_v3 = list()
+                        for tag_item in tags_v3:
+                            offi_tags_v3.append(tag_item.encode("utf-8"))
 
+                        tag_label = -1
+                        if len(offi_tags_v3)>0:
+                            tag_label = label_encoder.transform(test_val_list)[0]
+
+                        diary_feature_list = [tag_label,content_level]
+                        redis_client.hset(redis_name_content_linucb_feature,diary_id,json.dumps(diary_feature_list))
+
+                except:
+                    logging.error("catch exception,err_msg:%s" % traceback.format_exc())
+
+        except:
+            logging.error("catch exception,err_msg:%s" % traceback.format_exc())
+
+
+    @classmethod
+    def _get_user_portrait_tag3_redis_key(cls,device_id):
+        return "doris:user_portrait:tag3:device_id:" + str(device_id)
+
+    @classmethod
+    def generate_user_feature_to_redis(cls,device_id,label_encoder):
+        try:
+            redis_name_user_linucb_feature = cls.redis_name_user_linucb_feature_prefix
+
+            user_portrait_redis_name = cls._get_user_portrait_tag3_redis_key(device_id)
+            user_portrait_redis_data = redis_client.get(user_portrait_redis_name)
+            user_portrait_dict = json.loads(user_portrait_redis_data) if user_portrait_redis_data else dict()
+
+            user_tag_label = -1
+            if len(user_portrait_dict)>0:
+                sorted_user_portrait_list = sorted(user_portrait_dict["projects"].items(), key=lambda x: x[1], reverse=True)
+                user_tag_label = label_encoder.transform(sorted_user_portrait_list[0][0])[0]
+
+            user_feature_list = [user_tag_label]
+            redis_client.hset(redis_name_user_linucb_feature,device_id,json.dumps(user_feature_list))
+
+        except:
+            logging.error("catch exception,err_msg:%s" % traceback.format_exc())
 
 
 
@@ -79,6 +168,8 @@ if __name__ == "__main__":
 
     label_encoder = Generate_Feature_Info.get_tagv3_label_encode(tagv3_name_list=list(tagv3_name_set))
 
-    test_val = "切开双眼皮"
-
-    label_encoder.transform(test_val)
+    device_id="868771031984211"
+    Generate_Feature_Info.generate_user_feature_to_redis(device_id,label_encoder)
+    # test_val_list = ["切开双眼皮"]
+    #
+    # label_results_list = label_encoder.transform(test_val_list)