Skip to content

C
cybertron
Commit 66ffbe75 authored by lixiaofang's avatar lixiaofang
Browse files
Options

add

parent 21f9e748
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 6 additions and 171 deletions
.idea/workspace.xml
View file @ 66ffbe75
This diff is collapsed.
index_contrast/pictorial.py
View file @ 66ffbe75
This diff is collapsed.
index_contrast/tag.py
View file @ 66ffbe75
...@@ -18,7 +18,6 @@ ES_INFO_LIST = [ ...@@ -18,7 +18,6 @@ ES_INFO_LIST = [
] ]
ES_INDEX_PREFIX = "gm-dbmw" ES_INDEX_PREFIX = "gm-dbmw"
class ESPerform(object): class ESPerform(object):
cli_obj = None cli_obj = None
cli_info_list = ES_INFO_LIST cli_info_list = ES_INFO_LIST
...@@ -190,8 +189,8 @@ class ESPerform(object): ...@@ -190,8 +189,8 @@ class ESPerform(object):
try: try:
assert (es_cli is not None) assert (es_cli is not None)
official_index_name = sub_index_name official_index_name = sub_index_name
if sub_index_name != "mv-alpha-tag-test-190711901": # if sub_index_name != "mv-alpha-user-test-190711901":
official_index_name = cls.get_official_index_name(sub_index_name, "read") # official_index_name = cls.get_official_index_name(sub_index_name, "read")
index_exists = es_cli.indices.exists(official_index_name) index_exists = es_cli.indices.exists(official_index_name)
if not index_exists: if not index_exists:
...@@ -236,13 +235,13 @@ class ESPerform(object): ...@@ -236,13 +235,13 @@ class ESPerform(object):
try: try:
assert (es_cli is not None) assert (es_cli is not None)
official_index_name = cls.get_official_index_name(sub_index_name, "read") # official_index_name = cls.get_official_index_name(sub_index_name, "read")
index_exists = es_cli.indices.exists(official_index_name) index_exists = es_cli.indices.exists(sub_index_name)
if not index_exists: if not index_exists:
logging.error("index:%s is not existing,get_search_results error!" % official_index_name) logging.error("index:%s is not existing,get_search_results error!" % sub_index_name)
return None return None
res = es_cli.indices.analyze(index=official_index_name, body=query_body) res = es_cli.indices.analyze(index=sub_index_name, body=query_body)
return res return res
except: except:
logging.error("catch exception,err_msg:%s" % traceback.format_exc()) logging.error("catch exception,err_msg:%s" % traceback.format_exc())
...@@ -284,170 +283,6 @@ class ESPerform(object): ...@@ -284,170 +283,6 @@ class ESPerform(object):
logging.error("catch exception,err_msg:%s" % traceback.format_exc()) logging.error("catch exception,err_msg:%s" % traceback.format_exc())
return True return True
@classmethod
def get_tag_topic_list(cls, tag_id, have_read_topic_id_list, size=100):
try:
functions_list = list()
# for id in tag_id:
# functions_list.append(
# {
# "filter": {"term": {"tag_list": id}},
# "weight": 1
# }
# )
functions_list += [
{
"filter": {
"constant_score": {
"filter": {
"term": {"content_level": 6}}
}
},
"weight": 60
},
{
"filter": {
"constant_score": {
"filter": {
"term": {"content_level": 5}}
}
},
"weight": 50
},
{
"filter": {
"constant_score": {
"filter": {
"term": {"content_level": 4}}
}
},
"weight": 40
}
]
q = {
"query": {
"function_score": {
"query": {
"bool": {
"must": [
{"range": {"content_level": {"gte": 4, "lte": 6}}},
{"term": {"is_online": True}},
{"term": {"is_deleted": False}},
{"terms": {"tag_list": tag_id}}
]
}
},
"boost_mode": "sum",
"score_mode": "sum",
"functions": functions_list
}
},
"_source": {
"include": ["id"]
},
"sort": [
{"_score": {"order": "desc"}},
{"create_time_val": {"order": "desc"}},
# {"language_type": {"order": "asc"}},
]
}
if len(have_read_topic_id_list) > 0:
q["query"]["function_score"]["query"]["bool"]["must_not"] = {
"terms": {
"id": have_read_topic_id_list
}
}
result_dict = ESPerform.get_search_results(ESPerform.get_cli(), sub_index_name="topic", query_body=q,
offset=0, size=size, routing="4,5,6")
topic_id_list = [item["_source"]["id"] for item in result_dict["hits"]]
logging.info("topic_id_list:%s" % str(topic_id_list))
return topic_id_list
except:
logging.error("catch exception,err_msg:%s" % traceback.format_exc())
return list()
@classmethod
def get_tag_topic_list_dict(cls, tag_id, have_read_topic_id_list, size=100):
try:
functions_list = list()
for id in tag_id:
functions_list.append(
{
"filter": {"term": {"tag_list": id}},
"weight": 1
}
)
# functions_list += [
# {
# "filter": {"term": {"content_level": 6}},
# "weight": 6000
# },
# {
# "filter": {"term": {"content_level": 5}},
# "weight": 5000
# },
# {
# "filter": {"term": {"content_level": 4}},
# "weight": 4000
# }
# ]
q = {
"query": {
"function_score": {
"query": {
"bool": {
"must": [
{"term": {"content_level": 6}},
{"term": {"is_online": True}},
{"term": {"is_deleted": False}},
{"terms": {"tag_list": tag_id}}
]
}
},
"boost_mode": "sum",
"score_mode": "sum",
"functions": functions_list
}
},
"_source": {
"include": ["id", "user_id"]
},
"sort": [
{"_score": {"order": "desc"}},
{"create_time_val": {"order": "desc"}},
{"language_type": {"order": "asc"}},
],
"collapse": {
"field": "user_id"
}
}
if len(have_read_topic_id_list) > 0:
q["query"]["function_score"]["query"]["bool"]["must_not"] = {
"terms": {
"id": have_read_topic_id_list
}
}
result_dict = ESPerform.get_search_results(ESPerform.get_cli(), sub_index_name="topic-high-star",
query_body=q,
offset=0, size=size, routing="6")
topic_id_list = [item["_source"]["id"] for item in result_dict["hits"]]
# logging.info("topic_id_list:%s" % str(topic_id_list))
# topic_id_dict = [{str(item["_source"]["id"]):item["_source"]["user_id"]} for item in result_dict["hits"]]
topic_id_dict = dict()
for item in result_dict["hits"]:
topic_id_dict[str(item["_source"]["id"])] = item["_source"]["user_id"]
logging.info("topic_id_list:%s" % str(topic_id_dict))
return topic_id_list, topic_id_dict
except:
logging.error("catch exception,err_msg:%s" % traceback.format_exc())
return list()
# 先获取一部分数据 # 先获取一部分数据
es_cli_obj = ESPerform.get_cli() es_cli_obj = ESPerform.get_cli()
......
index_contrast/user.py
View file @ 66ffbe75
This diff is collapsed.
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment