Skip to content
Projects
Groups
Snippets
Help
Loading...
Sign in
Toggle navigation
F
ffm-baseline
Project
Project
Details
Activity
Cycle Analytics
Repository
Repository
Files
Commits
Branches
Tags
Contributors
Graph
Compare
Charts
Issues
0
Issues
0
List
Board
Labels
Milestones
Merge Requests
0
Merge Requests
0
CI / CD
CI / CD
Pipelines
Jobs
Schedules
Charts
Wiki
Wiki
Members
Members
Collapse sidebar
Close sidebar
Activity
Graph
Charts
Create a new issue
Jobs
Commits
Issue Boards
Open sidebar
ML
ffm-baseline
Commits
578d8211
Commit
578d8211
authored
Sep 03, 2019
by
高雅喆
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
优化
parent
73ebb97e
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
14 additions
and
23 deletions
+14
-23
dist_update_user_portrait.py
eda/smart_rank/dist_update_user_portrait.py
+14
-23
No files found.
eda/smart_rank/dist_update_user_portrait.py
View file @
578d8211
...
@@ -97,13 +97,20 @@ def tag_list2dict(lst, size):
...
@@ -97,13 +97,20 @@ def tag_list2dict(lst, size):
return
result
[:
size
]
return
result
[:
size
]
def
get_user_tag_score
(
cl_id
,
all_log_df
,
size
=
10
):
def
get_user_tag_score
(
cl_id
,
size
=
10
):
try
:
try
:
db_jerry_test
=
pymysql
.
connect
(
host
=
'172.16.40.158'
,
port
=
4000
,
user
=
'root'
,
passwd
=
'3SYz54LS9#^9sBvC'
,
db_jerry_test
=
pymysql
.
connect
(
host
=
'172.16.40.158'
,
port
=
4000
,
user
=
'root'
,
passwd
=
'3SYz54LS9#^9sBvC'
,
db
=
'jerry_test'
,
charset
=
'utf8'
)
db
=
'jerry_test'
,
charset
=
'utf8'
)
cur_jerry_test
=
db_jerry_test
.
cursor
()
cur_jerry_test
=
db_jerry_test
.
cursor
()
user_log_df
=
all_log_df
.
loc
[
all_log_df
[
'cl_id'
]
==
cl_id
]
# 获取该用户的所有行为
sql_user_log
=
"select time,cl_id,score_type,tag_id,tag_referrer,action from user_new_tag_log where cl_id ='{}'"
.
format
(
cl_id
)
cur_jerry_test
.
execute
(
sql_user_log
)
user_log
=
cur_jerry_test
.
fetchall
()
user_log_df
=
pd
.
DataFrame
(
list
(
user_log
))
user_log_df
.
columns
=
[
"time"
,
"cl_id"
,
"score_type"
,
"tag_id"
,
"tag_referrer"
,
"action"
]
if
not
user_log_df
.
empty
:
if
not
user_log_df
.
empty
:
user_log_df
[
"tag_id"
]
=
np
.
where
(
user_log_df
[
"action"
]
==
"do_search"
,
user_log_df
[
"tag_referrer"
],
user_log_df
[
"tag_id"
])
user_log_df
[
"tag_id"
]
=
np
.
where
(
user_log_df
[
"action"
]
==
"do_search"
,
user_log_df
[
"tag_referrer"
],
user_log_df
[
"tag_id"
])
user_log_df
[
"days_diff_now"
]
=
round
((
int
(
time
.
time
())
-
user_log_df
[
"time"
])
/
(
24
*
60
*
60
))
user_log_df
[
"days_diff_now"
]
=
round
((
int
(
time
.
time
())
-
user_log_df
[
"time"
])
/
(
24
*
60
*
60
))
...
@@ -122,12 +129,14 @@ def get_user_tag_score(cl_id, all_log_df, size=10):
...
@@ -122,12 +129,14 @@ def get_user_tag_score(cl_id, all_log_df, size=10):
tag_id_list_json
=
json
.
dumps
(
tag_id_list
)
tag_id_list_json
=
json
.
dumps
(
tag_id_list
)
gm_kv_cli
.
set
(
cl_id_portrait_key
,
tag_id_list_json
)
gm_kv_cli
.
set
(
cl_id_portrait_key
,
tag_id_list_json
)
# 写tidb
# 写tidb
stat_date
=
datetime
.
datetime
.
today
()
.
strftime
(
'
%
Y-
%
m-
%
d'
)
replace_sql
=
"""replace into user_portrait_tags (stat_date, cl_id, tag_list) values("{stat_date}","{cl_id}","{tag_list}")"""
\
replace_sql
=
"""replace into user_portrait_tags (stat_date, cl_id, tag_list) values("{stat_date}","{cl_id}","{tag_list}")"""
\
.
format
(
stat_date
=
stat_date
,
cl_id
=
cl_id
,
tag_list
=
tag_id_list
)
.
format
(
stat_date
=
stat_date
,
cl_id
=
cl_id
,
tag_list
=
tag_id_list
)
cur_jerry_test
.
execute
(
replace_sql
)
cur_jerry_test
.
execute
(
replace_sql
)
db_jerry_test
.
commit
()
db_jerry_test
.
commit
()
db_jerry_test
.
close
()
db_jerry_test
.
close
()
return
"sucess"
return
"sucess"
db_jerry_test
.
close
()
except
Exception
as
e
:
except
Exception
as
e
:
return
'pass'
return
'pass'
...
@@ -148,16 +157,8 @@ if __name__ == '__main__':
...
@@ -148,16 +157,8 @@ if __name__ == '__main__':
# 获取所有用户的行为日志
# 获取所有用户的行为日志
# sql_all_log = "select time,cl_id,score_type,tag_id,tag_referrer,action from user_new_tag_log"
# sql_all_log = "select time,cl_id,score_type,tag_id,tag_referrer,action from user_new_tag_log"
# 获取最近30天内的用户的所有行为
sql_all_log
=
"select time,cl_id,score_type,tag_id,tag_referrer,action from user_new_tag_log where cl_id in "
\
"(select distinct cl_id from user_new_tag_log "
\
"where time > UNIX_TIMESTAMP(DATE_SUB(NOW(), INTERVAL 30 day)))"
cur_jerry_test
.
execute
(
sql_all_log
)
all_log
=
cur_jerry_test
.
fetchall
()
db_jerry_test
.
close
()
all_log_df
=
pd
.
DataFrame
(
list
(
all_log
))
all_log_df
.
columns
=
[
"time"
,
"cl_id"
,
"score_type"
,
"tag_id"
,
"tag_referrer"
,
"action"
]
stat_date
=
datetime
.
datetime
.
today
()
.
strftime
(
'
%
Y-
%
m-
%
d'
)
# rdd
# rdd
sparkConf
=
SparkConf
()
.
set
(
"spark.hive.mapred.supports.subdirectories"
,
"true"
)
\
sparkConf
=
SparkConf
()
.
set
(
"spark.hive.mapred.supports.subdirectories"
,
"true"
)
\
.
set
(
"spark.hadoop.mapreduce.input.fileinputformat.input.dir.recursive"
,
"true"
)
\
.
set
(
"spark.hadoop.mapreduce.input.fileinputformat.input.dir.recursive"
,
"true"
)
\
...
@@ -170,17 +171,8 @@ if __name__ == '__main__':
...
@@ -170,17 +171,8 @@ if __name__ == '__main__':
spark
=
SparkSession
.
builder
.
config
(
conf
=
sparkConf
)
.
enableHiveSupport
()
.
getOrCreate
()
spark
=
SparkSession
.
builder
.
config
(
conf
=
sparkConf
)
.
enableHiveSupport
()
.
getOrCreate
()
spark
.
sparkContext
.
setLogLevel
(
"WARN"
)
spark
.
sparkContext
.
setLogLevel
(
"WARN"
)
device_ids_lst_rdd
=
spark
.
sparkContext
.
parallelize
(
device_ids_lst
)
device_ids_lst_rdd
=
spark
.
sparkContext
.
parallelize
(
device_ids_lst
)
result
=
device_ids_lst_rdd
.
repartition
(
100
)
.
map
(
lambda
x
:
get_user_tag_score
(
x
,
all_log_df
))
result
=
device_ids_lst_rdd
.
repartition
(
100
)
.
map
(
lambda
x
:
get_user_tag_score
(
x
))
result
.
collect
()
result
.
collect
()
# result_last = result_rename.withColumn("stat_date", lit(stat_date))
# result_last.show()
# df = result_last.select("stat_date", "cl_id", concat_ws(',', 'tag_list').alias("tag_list"))
# df.show()
# df.write.jdbc(
# mode="overwrite",
# url="jdbc:mysql://172.16.40.158:4000/jerry_test?user=root&password=3SYz54LS9#^9sBvC&useSSL=true",
# table="user_portrait_tags",
# properties={"driver": 'com.mysql.jdbc.Driver'})
except
Exception
as
e
:
except
Exception
as
e
:
send_email
(
"dist_update_user_portrait"
,
"dist_update_user_portrait"
,
"dist_update_user_portrait"
)
send_email
(
"dist_update_user_portrait"
,
"dist_update_user_portrait"
,
"dist_update_user_portrait"
)
\ No newline at end of file
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment