Merge branch 'feature/virtual_vote_num' into 'dev'

Feature/virtual vote num

See merge request !217

.gitignore

@@ -58,6 +58,7 @@ docs/_build/
 # config
 fabfile.py
 settings.online.py
+/gaia/settings.py
 settings_local.py
 media/
 log/
@@ -69,3 +70,16 @@ Vagrantfile
 
 *.DS_Store
 dump.rdb
+# .gitignore for yangchuncheng
+api/management/commands/ycc*
+settings_override*
+.script/
+.tmp.sql
+.env
+*.pem
+/gaia/hospital_list_settings.py
+coverage_html/
+gaia/rpcd.json
+*.swp
+
+dbmw_deploy/config.dir/

injection/data_sync/tasks.py

@@ -21,7 +21,6 @@ def write_to_es(es_type, pk_list, use_batch_query_set=False):
         type_info = type_info_map[es_type]
 
         logging.info("duan add,es_type:%s" % str(es_type))
-        logging.info("get es_type:%s"%es_type)
         type_info.insert_table_by_pk_list(
             sub_index_name=es_type,
             pk_list=pk_list,

libs/es.py

@@ -111,10 +111,6 @@ class ESPerform(object):
                 return False
 
             mapping_dict = cls.__load_mapping(sub_index_name)
-            logging.info("get write_alias_name:%s"%write_alias_name)
-            logging.info("get mapping_dict:%s"%mapping_dict)
-            logging.info("get mapping_type:%s"%mapping_type)
-
             es_cli.indices.put_mapping(index=write_alias_name,body=mapping_dict,doc_type=mapping_type)
 
             return True
@@ -158,14 +154,26 @@ class ESPerform(object):
                     cls.put_index_mapping(es_cli,sub_index_name)
 
             bulk_actions = []
-            for data in data_list:
-                bulk_actions.append({
-                    '_op_type': 'index',
-                    '_index': official_index_name,
-                    '_type': doc_type,
-                    '_id': data['id'],
-                    '_source': data,
-                })
+
+            if sub_index_name=="topic":
+                for data in data_list:
+                    bulk_actions.append({
+                        '_op_type': 'index',
+                        '_index': official_index_name,
+                        '_type': doc_type,
+                        '_id': data['id'],
+                        '_source': data,
+                        'routing': data["content_level"]
+                    })
+            else:
+                for data in data_list:
+                    bulk_actions.append({
+                        '_op_type': 'index',
+                        '_index': official_index_name,
+                        '_type': doc_type,
+                        '_id': data['id'],
+                        '_source': data,
+                    })
             elasticsearch.helpers.bulk(es_cli,bulk_actions)
 
             return True
@@ -226,17 +234,22 @@ class ESPerform(object):
             for item in es_nodes_info_list:
                 try:
                     item_list = item.split(" ")
-                    if len(item_list)>4:
+                    if len(item_list)==11:
+                        cpu_load = item_list[4]
+                    elif len(item_list)==10:
                         cpu_load = item_list[3]
-                        if int(cpu_load) > 60:
-                            high_num += 1
-                        es_nodes_list.append(int(cpu_load))
+                    else:
+                        continue
+                    int_cpu_load = int(cpu_load)
+                    if int_cpu_load > 60:
+                        high_num += 1
+                    es_nodes_list.append(int_cpu_load)
                 except:
                     logging.error("catch exception,item:%s,err_msg:%s" % (str(item),traceback.format_exc()))
                     return True
 
             if high_num > 3:
-                logging.info("check es_nodes_load high,cpu load:%s" % str(es_nodes_info_list))
+                logging.info("check es_nodes_load high,cpu load:%s,ori_cpu_info:%s" % (str(es_nodes_list), str(es_nodes_info_list)))
                 return True
             else:
                 return False

linucb/views/collect_data.py

@@ -9,6 +9,9 @@ import json
 from trans2es.models.tag import TopicTag
 import traceback
 from django.conf import settings
+from libs.es import ESPerform
+from search.utils.common import *
+
 
 class KafkaManager(object):
     consumser_obj = None
@@ -28,6 +31,8 @@ class CollectData(object):
     def __init__(self):
         self.linucb_matrix_redis_prefix = "physical:linucb:device_id:"
         self.linucb_recommend_redis_prefix = "physical:linucb:tag_recommend:device_id:"
+        self.linucb_recommend_topic_id_prefix = "physical:linucb:topic_recommend:device_id:"
+        self.tag_topic_id_redis_prefix = "physical:tag_id:topic_id_list:"
         # 默认
         self.user_feature = [0,1]
 
@@ -44,24 +49,88 @@ class CollectData(object):
             logging.error("catch exception,err_msg:%s" % traceback.format_exc())
             return dict()
 
+    def get_tag_topic_list(self,tag_id):
+        try:
+            q = {
+                "query":{
+                    "bool":{
+                        "must":[
+                            {"term":{"is_online": True}},
+                            {"term":{"is_deleted": False}},
+                            {"term":{"tag_list":tag_id}}
+                        ]
+                    }
+                },
+                "_source":{
+                    "include":["id"]
+                },
+                "sort":[
+                    {"create_time_val":{"order":"desc"}},
+                    {"language_type":{"order":"asc"}},
+                ]
+            }
+            result_dict = ESPerform.get_search_results(ESPerform.get_cli(), sub_index_name="topic-high-star", query_body=q,
+                                                       offset=0, size=5000)
+
+            topic_id_list = [item["_source"]["id"] for item in result_dict["hits"]]
+            return topic_id_list
+        except:
+            logging.error("catch exception,err_msg:%s" % traceback.format_exc())
+            return list()
+
     def update_recommend_tag_list(self, device_id,user_feature=None):
         try:
+            recommend_tag_set = set()
             recommend_tag_list = list()
+            recommend_tag_dict = dict()
             redis_linucb_tag_data_dict = self._get_user_linucb_info(device_id)
             if len(redis_linucb_tag_data_dict) == 0:
                 recommend_tag_list = LinUCB.get_default_tag_list()
                 LinUCB.init_device_id_linucb_info(redis_client, self.linucb_matrix_redis_prefix,device_id,recommend_tag_list)
             else:
                 user_feature = user_feature if user_feature else self.user_feature
-                recommend_tag_list = LinUCB.linucb_recommend_tag(device_id,redis_linucb_tag_data_dict,user_feature,list(redis_linucb_tag_data_dict.keys()))
+                (recommend_tag_dict,recommend_tag_set) = LinUCB.linucb_recommend_tag(device_id,redis_linucb_tag_data_dict,user_feature,list(redis_linucb_tag_data_dict.keys()))
 
-            logging.info("duan add,device_id:%s,recommend_tag_list:%s" % (str(device_id), str(recommend_tag_list)))
-            if len(recommend_tag_list) > 0:
+            if len(recommend_tag_dict) > 0:
+                recommend_tag_list = list(recommend_tag_set)
                 tag_recommend_redis_key = self.linucb_recommend_redis_prefix + str(device_id)
                 redis_client.set(tag_recommend_redis_key, json.dumps(recommend_tag_list))
                 # Todo:设置过期时间，调研set是否支持
                 redis_client.expire(tag_recommend_redis_key, 7*24*60*60)
 
+
+                redis_key = "physical:home_recommend" + ":device_id:" + device_id + ":query_type:" + str(TopicPageType.HOME_RECOMMEND)
+
+                have_read_topic_id_list = list()
+                redis_field_list = [b'have_read_topic_list']
+                redis_field_val_list = redis_client.hmget(redis_key, redis_field_list)
+                if redis_field_val_list[0]:
+                    have_read_topic_id_list = list(json.loads(redis_field_val_list[0]))
+
+
+                recommend_topic_id_list = list()
+                for index in range(0,1000):
+                    for tag_id in recommend_tag_list[0:5]:
+                        redis_tag_id_key = self.tag_topic_id_redis_prefix + str(tag_id)
+                        redis_tag_id_data = redis_client.get(redis_tag_id_key)
+                        tag_topic_id_list = json.loads(redis_tag_id_data) if redis_tag_id_data else []
+                        if not redis_tag_id_data:
+                            tag_topic_id_list = self.get_tag_topic_list(tag_id)
+                            redis_client.set(redis_tag_id_key,json.dumps(tag_topic_id_list))
+                            redis_client.expire(redis_tag_id_key,1*24*60*60)
+
+                        if len(tag_topic_id_list)>index:
+                            for topic_id in tag_topic_id_list[index:]:
+                                if topic_id not in have_read_topic_id_list and topic_id not in recommend_topic_id_list:
+                                    recommend_topic_id_list.append(topic_id)
+                                    break
+
+                topic_recommend_redis_key = self.linucb_recommend_topic_id_prefix + str(device_id)
+                redis_data_dict = {
+                    "data": json.dumps(recommend_topic_id_list),
+                    "cursor":0
+                }
+                redis_client.hmset(topic_recommend_redis_key,redis_data_dict)
             return True
         except:
             logging.error("catch exception,err_msg:%s" % traceback.format_exc())
@@ -98,9 +167,10 @@ class CollectData(object):
                                 logging.info("consume topic_id:%s,device_id:%s" % (str(topic_id), str(device_id)))
 
                                 tag_list = list()
-                                sql_query_results = TopicTag.objects.filter(is_online=True, topic_id=topic_id)
-                                for sql_item in sql_query_results:
-                                    tag_list.append(sql_item.tag_id)
+                                click_sql_query_results = TopicTag.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=topic_id).values_list("tag_id","is_online")
+                                for tag_id,is_online in click_sql_query_results:
+                                    if is_online:
+                                        tag_list.append(tag_id)
 
                                 is_click = 1
                                 is_vote = 0
@@ -130,17 +200,19 @@ class CollectData(object):
                                     exposure_topic_id = item["card_id"]
                                     logging.info(
                                         "consume exposure topic_id:%s,device_id:%s" % (str(exposure_topic_id), str(device_id)))
-                                    exposure_topic_id_list.append(exposure_topic_id)
+                                    if exposure_topic_id:
+                                        exposure_topic_id_list.append(exposure_topic_id)
 
                                 topic_tag_id_dict = dict()
                                 tag_list = list()
-                                sql_query_results = TopicTag.objects.filter(is_online=True, topic_id__in=exposure_topic_id_list)
-                                for sql_item in sql_query_results:
-                                    tag_list.append(sql_item.tag_id)
-
-                                    if sql_item.topic_id not in topic_tag_id_dict:
-                                        topic_tag_id_dict[sql_item.topic_id] = list()
-                                    topic_tag_id_dict[sql_item.topic_id].append(sql_item.tag_id)
+                                exposure_sql_query_results = TopicTag.objects.using(settings.SLAVE_DB_NAME).filter(topic_id__in=exposure_topic_id_list).values_list("topic_id","tag_id","is_online")
+                                for topic_id,tag_id,is_online in exposure_sql_query_results:
+                                    if is_online:
+                                        tag_list.append(tag_id)
+
+                                    if topic_id not in topic_tag_id_dict:
+                                        topic_tag_id_dict[topic_id] = list()
+                                    topic_tag_id_dict[topic_id].append(tag_id)
 
                                 is_click = 0
                                 is_vote = 0

linucb/views/linucb.py

@@ -9,12 +9,14 @@ import logging
 import traceback
 import json
 import pickle
+from django.conf import settings
+
 
 class LinUCB:
     d = 2
-    alpha = 0.25
-    r1 = 1
-    r0 = -0.5
+    alpha = 0.01
+    r1 = 10
+    r0 = -0.1
     default_tag_list = list()
 
     @classmethod
@@ -22,12 +24,9 @@ class LinUCB:
         try:
             if len(cls.default_tag_list) == 0:
 
-                query_item_results = Tag.objects.filter(is_online=True)
-
-                for item in query_item_results:
-                    cls.default_tag_list.append(item.id)
+                cls.default_tag_list = Tag.objects.using(settings.SLAVE_DB_NAME).filter(is_online=True,collection=1).values_list("id",flat=True)[0:100]
 
-            return cls.default_tag_list[:20]
+            return cls.default_tag_list
         except:
             logging.error("catch exception,err_msg:%s" % traceback.format_exc())
             return list()
@@ -72,6 +71,7 @@ class LinUCB:
 
 
             top_tag_set = set()
+            top_tag_dict = dict()
             np_score_list = list()
             np_score_dict = dict()
 
@@ -86,16 +86,20 @@ class LinUCB:
             sorted_np_score_list = sorted(np_score_list,reverse=True)
             for top_score in sorted_np_score_list:
                 for top_score_index in np_score_dict[top_score]:
-                    top_tag_set.add(str(tag_list[top_score_index], encoding="utf-8"))
+                    tag_id = str(tag_list[top_score_index], encoding="utf-8")
+                    top_tag_dict[tag_id] = top_score
+                    top_tag_set.add(tag_id)
+                    if len(top_tag_dict) >= 20:
+                        break
 
-                if len(top_tag_set) >= 10:
+                if len(top_tag_dict) >= 20:
                     break
 
             logging.info("duan add,device_id:%s,sorted_np_score_list:%s,np_score_dict:%s" % (str(device_id), str(sorted_np_score_list), str(np_score_dict)))
-            return list(top_tag_set)
+            return (top_tag_dict,top_tag_set)
         except:
             logging.error("catch exception,err_msg:%s" % traceback.format_exc())
-            return []
+            return ({},())
 
     @classmethod
     def init_device_id_linucb_info(cls, redis_cli,redis_prefix, device_id, tag_list):

search/views/group.py

@@ -13,7 +13,7 @@ from libs.es import ESPerform
 
 
 @bind("physical/search/query_pictorial")
-def query_group(query="", offset=0, size=10):
+def query_pictorial(query="", offset=0, size=10):
     """
     :remark:小组搜索排序策略，缺少排序策略
     :param query:
@@ -105,7 +105,7 @@ def pictorial_topic(topic_id=-1, offset=0, size=10):
             }
         }
         q["_source"] = {
-            "include": ["id", "pictorial_id", "tag_list"]
+            "includes": ["id", "pictorial_id", "tag_list"]
         }
         result_dict = ESPerform.get_search_results(es_cli_obj, "topic", q, offset, size)
         logging.info("get result_dict:%s" % result_dict)
@@ -142,7 +142,7 @@ def pictorial_topic(topic_id=-1, offset=0, size=10):
 
             }
             q["_source"] = {
-                "include": ["id", "update_time"]
+                "includes": ["id", "update_time"]
             }
             q["sort"] = {
                 'update_time': {
@@ -155,8 +155,6 @@ def pictorial_topic(topic_id=-1, offset=0, size=10):
                 for item in result_dict["hits"]:
                     pictorial_id = item["_source"]["id"]
                     pictorial_id_list.append(pictorial_id)
-        logging.info("get pictorial_id_list:%s" % pictorial_id_list)
-        logging.info("get topic_tag_list:%s" % topic_tag_list)
 
         if len(pictorial_id_list) < 10:
             num = 10 - len(pictorial_id_list)
@@ -177,7 +175,7 @@ def pictorial_topic(topic_id=-1, offset=0, size=10):
                 }
             }
             q["_source"] = {
-                "include": ["id", "tag_id"]}
+                "includes": ["id", "tag_id"]}
             q["sort"] = {
                 'update_time': {
                     'order': 'desc'
@@ -194,9 +192,6 @@ def pictorial_topic(topic_id=-1, offset=0, size=10):
                     if id not in pictorial_id_list:
                         pictorial_id_list.append(id)  #
 
-            logging.info("get result_dict tag:%s" % result_dict)
-            logging.info("get pictorial_id_list tag:%s" % pictorial_id_list)
-
             pictorial_list = pictorial_id_list if len(pictorial_id_list) < 10 else pictorial_id_list[:10]
 
             return {"pictorial_ids_list": pictorial_list}

search/views/topic.py

@@ -27,7 +27,7 @@ def get_discover_page_topic_ids(user_id, device_id, size, query_type=TopicPageTy
 
         recommend_topic_ids = TopicUtils.get_recommend_topic_ids(user_id=user_id, tag_id=0, offset=0, size=size,single_size=size,
                                                            query_type=query_type,
-                                                           filter_topic_id_list=have_read_topic_id_list)
+                                                           filter_topic_id_list=have_read_topic_id_list,index_type="topic-high-star")
 
 
         have_read_topic_id_list.extend(recommend_topic_ids)
@@ -46,45 +46,56 @@ def get_home_recommend_topic_ids(user_id, device_id, tag_id, offset, size, query
                                  query_type=TopicPageType.HOME_RECOMMEND):
     try:
         if query is None:
-            if user_id == -1:
-                redis_key = "physical:home_recommend" + ":user_id:" + str(
-                    user_id) + ":device_id:" + device_id + ":query_type:" + str(query_type)
-            else:
-                redis_key = "physical:home_recommend" + ":user_id:" + str(user_id) + ":query_type:" + str(query_type)
+            # redis_key = "physical:home_recommend" + ":user_id:" + str(
+            #     user_id) + ":device_id:" + device_id + ":query_type:" + str(query_type)
+            redis_key = "physical:home_recommend" + ":device_id:" + device_id + ":query_type:" + str(query_type)
         else:
-            if user_id == -1:
-                redis_key = "physical:home_query" + ":user_id:" + str(
-                    user_id) + ":device_id:" + device_id + ":query:" + str(query) + ":query_type:" + str(query_type)
-            else:
-                redis_key = "physical:home_query" + ":user_id:" + str(user_id) + ":query:" + str(
-                    query) + ":query_type:" + str(query_type)
+            # redis_key = "physical:home_query" + ":user_id:" + str(
+            #     user_id) + ":device_id:" + device_id + ":query:" + str(query) + ":query_type:" + str(query_type)
+            redis_key = "physical:home_query" + ":device_id:" + device_id + ":query:" + str(query) + ":query_type:" + str(query_type)
+
 
         redis_field_list = [b'have_read_topic_list']
         redis_field_val_list = redis_client.hmget(redis_key, redis_field_list)
 
-        tag_recommend_redis_key = "physical:linucb:tag_recommend:device_id:" + str(device_id)
+        topic_recommend_redis_key = "physical:linucb:topic_recommend:device_id:" + str(device_id)
+
 
-        recommend_tag_list = []
-        tag_recommend_val = redis_client.get(tag_recommend_redis_key)
-        if tag_recommend_val:
-            recommend_tag_list = json.loads(str(tag_recommend_val, encoding="utf-8"))
+        # recommend_tag_dict = dict()
+        # tag_recommend_val = redis_client.get(tag_recommend_redis_key)
+        # if tag_recommend_val:
+        #     recommend_tag_dict = json.loads(str(tag_recommend_val, encoding="utf-8"))
+
+        recommend_topic_list=list()
+        recommend_topic_dict = redis_client.hgetall(topic_recommend_redis_key)
+        if b"data" in recommend_topic_dict:
+            recommend_topic_id_list = json.loads(recommend_topic_dict[b"data"])
+            cursor = int(str(recommend_topic_dict[b"cursor"], encoding="utf-8"))
+            newcursor = cursor + 5
+            if len(recommend_topic_id_list) > newcursor:
+                recommend_topic_list = recommend_topic_id_list[cursor:newcursor]
+                redis_client.hset(topic_recommend_redis_key,"cursor",newcursor)
 
-        recommend_topic_ids = []
         have_read_topic_id_list = list()
 
-        if redis_field_val_list[0] and query is None:
-            have_read_topic_id_list = list(json.loads(redis_field_val_list[0]))
+        if redis_field_val_list[0]:
+            if query is None:
+                have_read_topic_id_list = list(json.loads(redis_field_val_list[0]))
+            else:
+                if offset>0:
+                    have_read_topic_id_list = list(json.loads(redis_field_val_list[0]))
 
         user_similar_score_redis_key = "physical:user_similar_score:user_id:" + str(user_id)
         redis_user_similar_score_redis_val = redis_client.get(user_similar_score_redis_key)
         user_similar_score_redis_list = json.loads(
             redis_user_similar_score_redis_val) if redis_user_similar_score_redis_val else []
 
+        size = size-len(recommend_topic_list)
         topic_id_list = TopicUtils.get_recommend_topic_ids(user_id=user_id, tag_id=tag_id, offset=offset, size=size,
                                                            single_size=size,query=query, query_type=query_type,
                                                            filter_topic_id_list=have_read_topic_id_list,
-                                                           recommend_tag_list=recommend_tag_list,
-                                                           user_similar_score_list=user_similar_score_redis_list)
+                                                           recommend_tag_list=recommend_topic_list,
+                                                           user_similar_score_list=user_similar_score_redis_list,index_type="topic-high-star")
         have_read_group_id_set = set()
         have_read_user_id_set = set()
         unread_topic_id_dict = dict()
@@ -125,18 +136,19 @@ def get_home_recommend_topic_ids(user_id, device_id, tag_id, offset, size, query
         #         else:
         #             break
 
-        have_read_topic_id_list.extend(topic_id_list)
-        if len(have_read_topic_id_list) > 5000:
-            cut_len = len(have_read_topic_id_list)-5000
+        recommend_topic_list.extend(topic_id_list)
+        have_read_topic_id_list.extend(recommend_topic_list)
+        if len(have_read_topic_id_list) > 30000:
+            cut_len = len(have_read_topic_id_list)-30000
             have_read_topic_id_list = have_read_topic_id_list[cut_len:]
         redis_dict = {
             "have_read_topic_list": json.dumps(have_read_topic_id_list),
         }
         redis_client.hmset(redis_key, redis_dict)
         # 每个session key保存15分钟
-        redis_client.expire(redis_key, 60 * 60 * 24 * 3)
+        redis_client.expire(redis_key, 60 * 60 * 24 * 30)
 
-        return topic_id_list
+        return recommend_topic_list
     except:
         logging.error("catch exception,err_msg:%s" % traceback.format_exc())
         return []
@@ -170,11 +182,11 @@ def home_recommend(device_id="", user_id=-1, offset=0, size=10, query_type=Topic
         recommend_topic_ids = list()
 
         es_node_load_high_flag = False
-        try:
-            es_node_load_high_flag = ESPerform.if_es_node_load_high(ESPerform.get_cli())
-        except:
-            logging.error("catch exception,err_msg:%s" % traceback.format_exc())
-            es_node_load_high_flag = True
+        # try:
+        #     es_node_load_high_flag = ESPerform.if_es_node_load_high(ESPerform.get_cli())
+        # except:
+        #     logging.error("catch exception,err_msg:%s" % traceback.format_exc())
+        #     es_node_load_high_flag = True
 
         if es_node_load_high_flag:
             temp_downgrading_key = "physical:home_recommend:user_id:241407656:query_type:1"
@@ -273,7 +285,7 @@ def topic_detail_page_recommend(device_id="", user_id=-1, topic_id=-1, topic_pic
 
         result_list = TopicUtils.get_topic_detail_recommend_list(user_id, topic_id, topic_tag_list, topic_pictorial_id,
                                                                  topic_user_id, filter_topic_user_id,
-                                                                 have_read_topic_list, offset, size, es_cli_obj)
+                                                                 have_read_topic_list, offset, size, es_cli_obj,index_type="topic-high-star")
         recommend_topic_ids_list = list()
         if len(result_list) > 0:
             recommend_topic_ids_list = [item["_source"]["id"] for item in result_list]
@@ -347,7 +359,7 @@ def query_topic_by_user_similarity(topic_similarity_score_dict, offset=0, size=1
     try:
         must_topic_id_list = list(topic_similarity_score_dict.keys())
         topic_id_list = TopicUtils.get_recommend_topic_ids(tag_id=0, user_id=-1, offset=offset, size=size,single_size=size,
-                                                           must_topic_id_list=must_topic_id_list)
+                                                           must_topic_id_list=must_topic_id_list,index_type="topic-high-star")
 
         return {"recommend_topic_ids": topic_id_list}
     except:

trans2es/management/commands/trans2es_data2es_parallel.py

@@ -49,6 +49,7 @@ class Job(object):
     def __call__(self):
         type_info = get_type_info_map()[self._type_name]
         assert isinstance(type_info, TypeInfo)
+
         result = type_info.insert_table_chunk(
             sub_index_name=self._sub_index_name,
             table_chunk=self._chunk,

trans2es/mapping/topic-high-star.json

+{
+  "dynamic":"strict",
+  "properties": {
+    "id":{"type":"long"},
+    "is_online":{"type":"boolean"},//上线
+    "is_deleted":{"type":"boolean"},
+    "vote_num":{"type":"long"},
+    "reply_num":{"type":"long"},
+    "name":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
+    "description":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
+    "content":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
+    "content_level":{"type":"text"},
+    "user_id":{"type":"long"},
+    "user_nick_name":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},//帖子用户名
+    "group_id":{"type":"long"}, //所在组ID
+    "tag_list":{"type":"long"},//标签属性
+    "edit_tag_list":{"type":"long"},//编辑标签
+    "tag_name_list":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
+    "share_num":{"type":"long"},
+    "pick_id_list":{"type":"long"},
+    "offline_score":{"type":"double"},//离线算分
+    "manual_score":{"type":"double"},//人工赋分
+    "has_image":{"type":"boolean"},//是否有图
+    "has_video":{"type":"boolean"},//是否是视频
+    "create_time":{"type":"date", "format":"date_time_no_millis"},
+    "update_time":{"type":"date", "format":"date_time_no_millis"},
+    "create_time_val":{"type":"long"},
+    "update_time_val":{"type":"long"},
+    "language_type":{"type":"long"},
+
+    "is_shadow": {"type": "boolean"},
+    "is_recommend": {"type": "boolean"},
+    "is_complaint": {"type": "boolean"},  // 是否被举报
+
+    "virtual_content_level":{"type": "text"},
+    "like_num_crawl": {"type": "long"}, // 爬取点赞数
+    "comment_num_crawl": {"type": "long"}, // 爬取评论数
+
+    "is_crawl": {"type": "boolean"},
+    "platform": {"type": "long"},
+    "platform_id": {"type": "long"},
+
+    "drop_score":{"type": "double"},  // 人工降分
+    "sort_score":{"type": "double"},  // 排序分
+
+    "pictorial_id":{"type": "long"}, //所在组ID
+    "pictorial_name":{  // 所在组名称
+        "type": "text",
+        "analyzer": "gm_default_index",
+        "search_analyzer": "gm_default_index"
+    }
+
+  }
+}

trans2es/mapping/topic-star-routing.json

+{
+  "dynamic":"strict",
+  "properties": {
+    "id":{"type":"long"},
+    "is_online":{"type":"boolean"},//上线
+    "is_deleted":{"type":"boolean"},
+    "vote_num":{"type":"long"},
+    "reply_num":{"type":"long"},
+    "name":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
+    "description":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
+    "content":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
+    "content_level":{"type":"text"},
+    "user_id":{"type":"long"},
+    "user_nick_name":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},//帖子用户名
+    "group_id":{"type":"long"}, //所在组ID
+    "tag_list":{"type":"long"},//标签属性
+    "edit_tag_list":{"type":"long"},//编辑标签
+    "tag_name_list":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
+    "share_num":{"type":"long"},
+    "pick_id_list":{"type":"long"},
+    "offline_score":{"type":"double"},//离线算分
+    "manual_score":{"type":"double"},//人工赋分
+    "has_image":{"type":"boolean"},//是否有图
+    "has_video":{"type":"boolean"},//是否是视频
+    "create_time":{"type":"date", "format":"date_time_no_millis"},
+    "update_time":{"type":"date", "format":"date_time_no_millis"},
+    "create_time_val":{"type":"long"},
+    "update_time_val":{"type":"long"},
+    "language_type":{"type":"long"},
+
+    "is_shadow": {"type": "boolean"},
+    "is_recommend": {"type": "boolean"},
+    "is_complaint": {"type": "boolean"},  // 是否被举报
+
+    "virtual_content_level":{"type": "text"},
+    "like_num_crawl": {"type": "long"}, // 爬取点赞数
+    "comment_num_crawl": {"type": "long"}, // 爬取评论数
+
+    "is_crawl": {"type": "boolean"},
+    "platform": {"type": "long"},
+    "platform_id": {"type": "long"},
+
+    "drop_score":{"type": "double"},  // 人工降分
+    "sort_score":{"type": "double"},  // 排序分
+
+    "pictorial_id":{"type": "long"}, //所在组ID
+    "pictorial_name":{  // 所在组名称
+        "type": "text",
+        "analyzer": "gm_default_index",
+        "search_analyzer": "gm_default_index"
+    }
+
+  }
+}

trans2es/mapping/topic.json

@@ -5,12 +5,14 @@
     "is_online":{"type":"boolean"},//上线
     "is_deleted":{"type":"boolean"},
     "vote_num":{"type":"long"},
+    "total_vote_num":{"type":"long"},
     "reply_num":{"type":"long"},
     "name":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
     "description":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
     "content":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},
     "content_level":{"type":"text"},
     "user_id":{"type":"long"},
+    "user_nick_name":{"type":"text","analyzer":"gm_default_index","search_analyzer":"gm_default_index"},//帖子用户名
     "group_id":{"type":"long"}, //所在组ID
     "tag_list":{"type":"long"},//标签属性
     "edit_tag_list":{"type":"long"},//编辑标签

trans2es/models/group.py

@@ -27,15 +27,15 @@ class Group(models.Model):
     create_time = models.DateTimeField(verbose_name=u'创建时间',default=datetime.datetime.fromtimestamp(0))
     update_time = models.DateTimeField(verbose_name=u'更新时间', default=datetime.datetime.fromtimestamp(0))
 
-    # #获取前一天4，5星发帖数
-    # def get_high_quality_topic_num(self):
-    #     yesterday = datetime.datetime.now()-datetime.timedelta(days=1)
-    #     yesterday_begin_time = "%s-%s-%s 00:00:00" % (yesterday.year, yesterday.month, yesterday.day)
-    #     yesterday_end_time = "%s-%s-%s 23:59:59" % (yesterday.year, yesterday.month, yesterday.day)
-    #
-    #     topic_num = self.group_topics.filter(content_level__in=("4","5"),create_time__gte=yesterday_begin_time,create_time__lte=yesterday_end_time).count()
-    #
-    #     return topic_num
+    #获取前一天4，5星发帖数
+    def get_high_quality_topic_num(self):
+        yesterday = datetime.datetime.now()-datetime.timedelta(days=1)
+        yesterday_begin_time = "%s-%s-%s 00:00:00" % (yesterday.year, yesterday.month, yesterday.day)
+        yesterday_end_time = "%s-%s-%s 23:59:59" % (yesterday.year, yesterday.month, yesterday.day)
+
+        topic_num = self.group_topics.filter(content_level__in=("4","5"),create_time__gte=yesterday_begin_time,create_time__lte=yesterday_end_time).count()
+
+        return topic_num
 
     def detail(self):
         result = {

trans2es/models/pictorial.py

-from django.db import models
 import datetime
+from django.db import models
 import logging
 import traceback
 
@@ -109,7 +109,6 @@ class Pictorial(models.Model):
             tags = Tag.objects.filter(id__in=tag_id, is_online=True).values_list("name", flat=True)
             for i in tags:
                 tag_name_list.append(i)
-                logging.info("get tags name i:%s" % i)
 
             return tag_name_list
 

trans2es/models/topic.py

@@ -52,9 +52,9 @@ class Topic(models.Model):
     id = models.IntegerField(verbose_name=u'日记ID', primary_key=True)
     name = models.CharField(verbose_name=u'日记名称', max_length=100)
     # group_id = models.IntegerField(verbose_name='用户所在组ID',default=-1)
-    # group = models.ForeignKey(
-    #     Group, verbose_name=u"关联的小组", related_name=u"group_topics", null=True, blank=True, default=None,
-    #     on_delete=models.CASCADE)
+    group = models.ForeignKey(
+        Group, verbose_name=u"关联的小组", related_name=u"group_topics", null=True, blank=True, default=None,
+        on_delete=models.CASCADE)
     user_id = models.IntegerField(verbose_name=u'用户ID')
     has_video = models.BooleanField(verbose_name=u'是否是视频日记')
     drop_score = models.IntegerField(verbose_name=u'人工赋分', default=0)
@@ -82,6 +82,14 @@ class Topic(models.Model):
     platform = models.IntegerField(verbose_name=u'平台来源', choices=GRAP_PLATFORM, default=GRAP_PLATFORM.ALPHA)
     platform_id = models.BigIntegerField(verbose_name='用平台ID', null=True)
 
+    def get_virtual_vote_num(self):
+
+        try:
+            topic_extra = TopicExtra.object.get(topic_id=self.id)
+            return topic_extra.virtual_vote_num
+        except:
+            return 0
+
     def get_pictorial_id(self):
         try:
             pictorial_id_list =[]
@@ -106,8 +114,7 @@ class Topic(models.Model):
         try:
             has_image = False
 
-            query_list = TopicImage.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=self.id, is_deleted=False,
-                                                                                 is_online=True)
+            query_list = TopicImage.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=self.id, is_deleted=False, is_online=True)
             if len(query_list) > 0:
                 has_image = True
 
@@ -133,13 +140,12 @@ class Topic(models.Model):
             topic_tag_id_list = list()
             edit_tag_id_list = list()
 
-            tag_id_list = TopicTag.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=self.id).values_list("tag_id",
-                                                                                                              flat=True)
+            tag_id_list = TopicTag.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=self.id).values_list("tag_id", flat=True)
             tag_query_results = Tag.objects.using(settings.SLAVE_DB_NAME).filter(id__in=tag_id_list)
             for tag_item in tag_query_results:
-                is_online = tag_item.is_online
-                is_deleted = tag_item.is_deleted
-                collection = tag_item.collection
+                is_online=tag_item.is_online
+                is_deleted=tag_item.is_deleted
+                collection=tag_item.collection
 
                 if is_online and not is_deleted:
                     topic_tag_id_list.append(tag_item.id)
@@ -149,7 +155,7 @@ class Topic(models.Model):
             return (topic_tag_id_list, edit_tag_id_list)
         except:
             logging.error("catch exception,err_msg:%s" % traceback.format_exc())
-            return ([], [])
+            return ([],[])
 
     def get_tag_name_list(self, tag_id_list):
         try:
@@ -188,12 +194,9 @@ class Topic(models.Model):
             elif self.content_level == '3':
                 offline_score += 2.0
 
-            exposure_count = ActionSumAboutTopic.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=self.id,
-                                                                                              data_type=1).count()
-            click_count = ActionSumAboutTopic.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=self.id,
-                                                                                           data_type=2).count()
-            uv_num = ActionSumAboutTopic.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=self.id,
-                                                                                      data_type=3).count()
+            exposure_count = ActionSumAboutTopic.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=self.id, data_type=1).count()
+            click_count = ActionSumAboutTopic.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=self.id, data_type=2).count()
+            uv_num = ActionSumAboutTopic.objects.using(settings.SLAVE_DB_NAME).filter(topic_id=self.id, data_type=3).count()
 
             if exposure_count > 0:
                 offline_score += click_count / exposure_count
@@ -224,8 +227,7 @@ class TopicComplaint(models.Model):
     id = models.IntegerField(verbose_name='日记图片ID', primary_key=True)
     user_id = models.BigIntegerField(verbose_name=u'用户ID', db_index=True)
     topic = models.ForeignKey(
-        Topic, verbose_name=u"关联的帖子", null=True, blank=True, default=None, on_delete=models.CASCADE,
-        related_name='complaints')
+        Topic, verbose_name=u"关联的帖子", null=True, blank=True, default=None, on_delete=models.CASCADE, related_name='complaints')
 
     is_online = models.BooleanField(verbose_name=u"是否有效", default=True)
 
@@ -244,3 +246,15 @@ class PictorialTopic(models.Model):
     is_online = models.BooleanField(verbose_name=u"是否有效", default=True)
     is_online = models.BooleanField(verbose_name=u'是否上线')
     is_deleted = models.BooleanField(verbose_name=u'是否删除')
+
+
+class TopicExtra(BaseModel):
+    """帖子相关额外信息"""
+
+    class Meta:
+        verbose_name = '帖子额外信息'
+        app_label = 'community'
+        db_table = 'topic_extra'
+
+    topic_id = models.IntegerField(verbose_name=u"帖子ID",db_index=True)
+    virtual_vote_num = models.IntegerField(verbose_name="帖子虚拟点赞")

trans2es/models/user.py

@@ -34,6 +34,16 @@ class User(models.Model):
     create_time = models.DateTimeField(verbose_name=u'创建时间', default=datetime.datetime.fromtimestamp(0))
     update_time = models.DateTimeField(verbose_name=u'更新时间', default=datetime.datetime.fromtimestamp(0))
 
+    @classmethod
+    def get_user_nick_name(cls,user_id):
+        try:
+            nick_name = User.objects.using(settings.SLAVE_DB_NAME).filter(user_id=user_id).values_list("nick_name").first()
+
+            return nick_name[0]
+        except:
+            logging.error("catch exception,err_msg:%s" % traceback.format_exc())
+            return ""
+
     def get_is_recommend_flag(self):
         is_shadow = False
         is_recommend = False

trans2es/models/user_extra.py

@@ -15,7 +15,8 @@ class UserExtra(models.Model):
         db_table="user_extra"
 
     id = models.IntegerField(verbose_name="主键ID",primary_key=True)
-    user_id = models.BigIntegerField(verbose_name=u"用户ID")
+    user_id = models.CharField(verbose_name=u"用户ID",max_length=100)
+
     is_shadow = models.BooleanField(verbose_name=u"是否是马甲账户")
     is_online = models.BooleanField(verbose_name=u"是否上线")
     is_recommend = models.BooleanField(verbose_name=u"是否推荐")

trans2es/type_info.py

@@ -81,6 +81,9 @@ class TypeInfo(object):
 
     def bulk_get_data(self, instance_iterable):
         data_list = []
+        # 4，5星帖子单独索引
+        topic_data_high_star_list = list()
+
         if self.batch_get_data_func:
             _pk_list = [getattr(instance, 'pk', None) for instance in instance_iterable]
             not_found_pk_list = []
@@ -137,7 +140,10 @@ class TypeInfo(object):
                     ))
                 else:
                     data_list.append(data)
-        return data_list
+                    if self.type=="topic" and instance.content_level and int(instance.content_level)>=4:
+                        topic_data_high_star_list.append(data)
+
+        return (data_list,topic_data_high_star_list)
 
     def elasticsearch_bulk_insert_data(self, sub_index_name, data_list, es=None):
 
@@ -162,10 +168,10 @@ class TypeInfo(object):
         #             traceback.print_exc()
         #             es_result = 'error'
 
-        return ESPerform.es_helpers_bulk(es, data_list, sub_index_name, True)
+        return ESPerform.es_helpers_bulk(es, data_list, sub_index_name)
 
     def elasticsearch_bulk_insert(self, sub_index_name, instance_iterable, es=None):
-        data_list = self.bulk_get_data(instance_iterable)
+        data_list, topic_data_high_star_list = self.bulk_get_data(instance_iterable)
         return self.elasticsearch_bulk_insert_data(
             sub_index_name=sub_index_name,
             data_list=data_list,
@@ -188,7 +194,7 @@ class TypeInfo(object):
         time1=end-begin
 
         begin = time.time()
-        data_list = self.bulk_get_data(instance_list)
+        data_list, topic_data_high_star_list = self.bulk_get_data(instance_list)
         end = time.time()
         time2=end-begin
 
@@ -201,6 +207,14 @@ class TypeInfo(object):
             data_list=data_list,
             es=es,
         )
+        # 同时写4星及以上的帖子
+        if len(topic_data_high_star_list)>0:
+            self.elasticsearch_bulk_insert_data(
+                sub_index_name="topic-high-star",
+                data_list=topic_data_high_star_list,
+                es=es,
+            )
+
         end = time.time()
         time3=end-begin
         logging.info("duan add,insert_table_by_pk_list time cost:%ds,%ds,%ds,%ds" % (time0,time1,time2,time3))
@@ -215,7 +229,7 @@ class TypeInfo(object):
 
             stage_1_time = time.time()
 
-            data_list = self.bulk_get_data(instance_list)
+            data_list, topic_data_high_star_list = self.bulk_get_data(instance_list)
 
             stage_2_time = time.time()
 
@@ -260,7 +274,26 @@ def get_type_info_map():
         return _get_type_info_map_result
 
     type_info_list = [
-
+        TypeInfo(
+            name='topic-star-routing',
+            type='topic-star-routing',
+            model=topic.Topic,
+            query_deferred=lambda: topic.Topic.objects.all().query,
+            get_data_func=TopicTransfer.get_topic_data,
+            bulk_insert_chunk_size=100,
+            round_insert_chunk_size=5,
+            round_insert_period=2,
+        ),
+        TypeInfo(
+            name='topic-high-star',  # >=4星日记
+            type='topic-high-star',
+            model=topic.Topic,
+            query_deferred=lambda: topic.Topic.objects.all().query,
+            get_data_func=TopicTransfer.get_topic_data,
+            bulk_insert_chunk_size=100,
+            round_insert_chunk_size=5,
+            round_insert_period=2,
+        ),
         TypeInfo(
             name='topic',  # 日记
             type='topic',

trans2es/utils/group_transfer.py

@@ -33,7 +33,7 @@ class GroupTransfer(object):
             update_time = instance.update_time
             tzlc_udpate_time = tzlc(update_time)
             res["update_time"] = tzlc_udpate_time
-            # res["high_quality_topic_num"] = instance.get_high_quality_topic_num()
+            res["high_quality_topic_num"] = instance.get_high_quality_topic_num()
 
             return res
         except:

trans2es/utils/topic_transfer.py

@@ -8,7 +8,7 @@ from libs.tools import tzlc
 import time
 import re
 import datetime
-
+from trans2es.models.user import User
 
 class TopicTransfer(object):
 
@@ -27,12 +27,13 @@ class TopicTransfer(object):
             res["content"] = instance.content
             res["content_level"] = instance.content_level
             res["user_id"] = instance.user_id
+            res["user_nick_name"] = User.get_user_nick_name(instance.user_id)
 
-            # if instance.group:
-            #     res["group_id"] = instance.group.id
-            # else:
-            #     res["group_id"] = -1
-
+            if instance.group:
+                res["group_id"] = instance.group.id
+            else:
+                res["group_id"] = -1
+            res["share_num"] = instance.share_num
 
             res["pictorial_id"] = instance.get_pictorial_id()
 
@@ -82,6 +83,20 @@ class TopicTransfer(object):
             # else:
             #     res["language_type"] = instance.language_type
 
+            res["is_shadow"] = instance.is_shadow
+            res["is_recommend"] = True if instance.is_recommend else False
+            res["is_complaint"] = instance.is_complaint
+
+            res["virtual_content_level"] = instance.virtual_content_level
+            res["like_num_crawl"] = instance.like_num_crawl
+            res["comment_num_crawl"]= instance.comment_num_crawl
+
+            res["is_crawl"] = instance.is_crawl
+            res["platform"] = instance.platform
+            res["platform_id"] = instance.platform_id
+
+            res["drop_score"] = instance.drop_score
+            res["sort_score"] = instance.sort_score
 
             create_time = instance.create_time
             tzlc_create_time = tzlc(create_time)
@@ -91,11 +106,13 @@ class TopicTransfer(object):
 
             update_time = instance.update_time
             tzlc_update_time = tzlc(update_time)
-            # res["update_time"] = tzlc_update_time
+            res["update_time"] = tzlc_update_time
             res["update_time_val"] = int(time.mktime(tzlc_update_time.timetuple()))
 
+            res["total_vote_num"] = instance.get_virtual_vote_num() + instance.vote_num
+
             logging.info("test topic transfer time cost,time0:%d,time1:%d,time2:%d,time3:%d,time4:%d" % (time0,time1,time2,time3,time4))
             return res
         except:
             logging.error("catch exception,err_msg:%s" % traceback.format_exc())
-            return None
\ No newline at end of file
+            return None