mem_bech

olapy/core/mdx/executor/execute_config_file.py

 from __future__ import absolute_import, division, print_function
 
+from ..tools.mem_bench import memory_usage
 from ..tools.connection import MyDB
 import pandas.io.sql as psql
 import os
@@ -18,14 +19,17 @@ def _load_table_config_file(executer_instance, cube_obj):
 
     db = MyDB(db=executer_instance.cube)
 
+    memory_usage("1 - before executing query //// _load_table_config_file")
     for table in cube_obj.dimensions:
-        value = psql.read_sql_query("SELECT * FROM {0}".format(table.name),
-                                    db.engine)
+        with db.engine as connection:
+            value = psql.read_sql_query("SELECT * FROM {0}".format(table.name),
+                                        connection)
 
         tables[table.name] = value[[
             col for col in value.columns if col.lower()[-3:] != '_id'
         ]]
 
+    memory_usage("2 - after query, before fetchall  /////// _load_table_config_file")
     # update table display name
     for dimension in cube_obj.dimensions:
         if dimension.displayName and dimension.name and dimension.displayName != dimension.name:
@@ -47,17 +51,22 @@ def _construct_star_schema_config_file(executer_instance, cubes_obj):
     executer_instance.facts = cubes_obj.facts[0].table_name
     db = MyDB(db=executer_instance.cube)
     # load facts table
-    fusion = psql.read_sql_query(
-        "SELECT * FROM {0}".format(executer_instance.facts), db.engine)
 
-    for fact_key, dimension_and_key in cubes_obj.facts[0].keys.items():
-        df = psql.read_sql_query(
-            "SELECT * FROM {0}".format(dimension_and_key.split('.')[0]),
-            db.connection)
+    memory_usage("1 - before executing query //// _construct_star_schema_config_file")
+    with db.engine as connection:
+        fusion = psql.read_sql_query(
+            "SELECT * FROM {0}".format(executer_instance.facts), connection)
 
-        fusion = fusion.merge(
-            df, left_on=fact_key, right_on=dimension_and_key.split('.')[1])
+        for fact_key, dimension_and_key in cubes_obj.facts[0].keys.items():
+            df = psql.read_sql_query(
+                "SELECT * FROM {0}".format(dimension_and_key.split('.')[0]),
+                connection)
+
+            fusion = fusion.merge(
+                df, left_on=fact_key, right_on=dimension_and_key.split('.')[1])
 
+
+    memory_usage("2 - after query, before fetchall  /////// _construct_star_schema_config_file")
         # TODO CHOSE BETWEEN THOSES DF
         # if separated dimensions
         # fusion = fusion.merge(df, left_on=fact_key,right_on=dimension_and_key.split('.')[1])
@@ -90,10 +99,16 @@ def _construct_web_star_schema_config_file(executer_instance, cubes_obj):
     if cubes_obj.facts[0].columns:
         all_columns += cubes_obj.facts[0].columns
 
+
+    memory_usage("1 - before executing query //// 1111 _construct_web_star_schema_config_file ")
     fusion = psql.read_sql_query(
         "SELECT * FROM {0}".format(executer_instance.facts), db.engine)
 
+    memory_usage("2 - after query, before fetchall  /////// 222222222222 _construct_star_schema_config_file")
+
+
     tables = {}
+    memory_usage("1 - before executing query //// 3333333333 _construct_web_star_schema_config_file ")
     for table in cubes_obj.tables:
 
         tab = psql.read_sql_query("SELECT * FROM {0}".format(table.name),
@@ -118,11 +133,14 @@ def _construct_web_star_schema_config_file(executer_instance, cubes_obj):
         all_columns += list(tab.columns)
         tables.update({table.name: tab})
 
+    memory_usage("2 - after query, before fetchall  /////// 44444444 _construct_star_schema_config_file")
+
     # measures in config-file only
     if cubes_obj.facts[0].measures:
         executer_instance.measures = cubes_obj.facts[0].measures
         all_columns += cubes_obj.facts[0].measures
 
+    memory_usage("1 - before executing query //// 55555555 _construct_web_star_schema_config_file ")
     for fact_key, dimension_and_key in cubes_obj.facts[0].keys.items():
         dimension_name = dimension_and_key.split('.')[0]
         if dimension_name in tables.keys():
@@ -138,5 +156,7 @@ def _construct_web_star_schema_config_file(executer_instance, cubes_obj):
               # remove suffixe from dimension and keep the same column name for facts
               suffixes=('', '_y'))
 
+    memory_usage("2 - after query, before fetchall  /////// 6666666666 _construct_star_schema_config_file")
+
     return fusion[[column for column in all_columns if 'id' != column[-2:]]]
 

olapy/core/mdx/executor/execute_db.py

@@ -2,6 +2,7 @@ from __future__ import absolute_import, division, print_function
 
 from sqlalchemy import inspect
 
+from ..tools.mem_bench import memory_usage
 from ..tools.connection import MyDB
 import pandas.io.sql as psql
 
@@ -16,6 +17,7 @@ def _load_tables_db(executer_instance):
     db = MyDB(db_config_file_path=executer_instance.DATA_FOLDER,db=executer_instance.cube)
     inspector = inspect(db.engine)
 
+    memory_usage("1 - before executing query //// _load_tables_db")
     for table_name in inspector.get_table_names():
         value = psql.read_sql_query(
             'SELECT * FROM "{0}"'.format(table_name), db.engine)
@@ -23,6 +25,7 @@ def _load_tables_db(executer_instance):
         tables[table_name] = value[[
             col for col in value.columns if col.lower()[-3:] != '_id'
         ]]
+    memory_usage("2 - after query, before fetchall  /////// _load_tables_db")
     return tables
 
 
@@ -34,20 +37,22 @@ def _construct_star_schema_db(executer_instance):
     :return: star schema DataFrame
     """
     db = MyDB(db=executer_instance.cube)
-
+    memory_usage("1 - before executing query //// _construct_star_schema_db")
     # load facts table
-    fusion = psql.read_sql_query(
-        'SELECT * FROM "{0}" '.format(executer_instance.facts), db.engine)
-
-    inspector = inspect(db.engine)
-
-    for db_table_name in inspector.get_table_names():
-        try:
-            fusion = fusion.merge(
-                psql.read_sql_query("SELECT * FROM {0}".format(
-                    db_table_name[0]), db.engine))
-        except:
-            print('No common column')
-            pass
-
+    with db.engine as connection:
+        fusion = psql.read_sql_query(
+            'SELECT * FROM "{0}" '.format(executer_instance.facts), connection)
+
+        inspector = inspect(connection)
+
+        for db_table_name in inspector.get_table_names():
+            try:
+                fusion = fusion.merge(
+                    psql.read_sql_query("SELECT * FROM {0}".format(
+                        db_table_name[0]), connection))
+            except:
+                print('No common column')
+                pass
+
+    memory_usage("2 - after query, before fetchall  /////// _construct_star_schema_db")
     return fusion

olapy/core/mdx/tools/bech_mem_sqlalchemy_without_opt.txt

+Memory summary:1 - before executing query //// _load_tables_db
+
+VM: 3160.52Mb
+
+Memory summary:2 - after query, before fetchall  /////// _load_tables_db
+
+VM: 3268.52Mb
+
+Memory summary:1 - before executing query //// 1111 _construct_web_star_schema_config_file 
+
+VM: 3268.52Mb
+
+Memory summary:2 - after query, before fetchall  /////// 222222222222 _construct_star_schema_config_file
+
+VM: 3269.02Mb
+
+Memory summary:1 - before executing query //// 3333333333 _construct_web_star_schema_config_file 
+
+VM: 3269.02Mb
+
+Memory summary:2 - after query, before fetchall  /////// 44444444 _construct_star_schema_config_file
+
+VM: 3271.02Mb
+
+Memory summary:1 - before executing query //// 55555555 _construct_web_star_schema_config_file 
+
+VM: 3271.02Mb
+
+Memory summary:2 - after query, before fetchall  /////// 6666666666 _construct_star_schema_config_file
+
+VM: 3271.02Mb
+
+Memory summary:1 - before executing query //// 1111 _construct_web_star_schema_config_file 
+
+VM: 3271.02Mb
+
+Memory summary:2 - after query, before fetchall  /////// 222222222222 _construct_star_schema_config_file
+
+VM: 3270.77Mb
+
+Memory summary:1 - before executing query //// 3333333333 _construct_web_star_schema_config_file 
+
+VM: 3270.77Mb
+
+Memory summary:2 - after query, before fetchall  /////// 44444444 _construct_star_schema_config_file
+
+VM: 3271.02Mb
+
+Memory summary:1 - before executing query //// 55555555 _construct_web_star_schema_config_file 
+
+VM: 3271.02Mb
+
+Memory summary:2 - after query, before fetchall  /////// 6666666666 _construct_star_schema_config_file
+
+VM: 3271.02Mb
+
+Memory summary:1 - before executing query //// 1111 _construct_web_star_schema_config_file 
+
+VM: 3271.02Mb
+
+Memory summary:2 - after query, before fetchall  /////// 222222222222 _construct_star_schema_config_file
+
+VM: 3271.02Mb
+
+Memory summary:1 - before executing query //// 3333333333 _construct_web_star_schema_config_file 
+
+VM: 3271.02Mb
+
+Memory summary:2 - after query, before fetchall  /////// 44444444 _construct_star_schema_config_file
+
+VM: 3271.02Mb
+
+Memory summary:1 - before executing query //// 55555555 _construct_web_star_schema_config_file 
+
+VM: 3271.02Mb
+
+Memory summary:2 - after query, before fetchall  /////// 6666666666 _construct_star_schema_config_file
+
+VM: 3271.02Mb
+

olapy/core/mdx/tools/mem_bench.py

+import os
+from os.path import expanduser
+
+from pympler import summary, muppy
+import psutil
+
+def get_virtual_memory_usage_kb():
+    """
+    The process's current virtual memory size in Kb, as a float.
+
+    """
+    return float(psutil.Process().memory_info_ex().vms) / 1024.0
+
+def memory_usage(where):
+    """
+    Print out a basic summary of memory usage.
+
+    """
+    with open(os.path.join(expanduser('~'), 'bech_mem.txt'), mode='a+') as file:
+        mem_summary = summary.summarize(muppy.get_objects())
+        file.write("Memory summary:" + where + '\n\n')
+        summary.print_(mem_summary, limit=2)
+        file.write("VM: %.2fMb" % (get_virtual_memory_usage_kb() / 1024.0) + '\n\n')