-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsim_run_experiment.py
204 lines (180 loc) · 10.5 KB
/
sim_run_experiment.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
import pickle
from importlib import reload
import logging
from pandas import DataFrame
import constants
from bandits.experiment_report import ExpReport
from database.config_test_run import ConfigRunner
from database.dta_test_run_v2 import DTARunner
from shared import configs_v2 as configs, helper
def select_driver():
import pyodbc
"""Find least version of: ODBC Driver for SQL Server."""
#print(pyodbc.drivers())
drv = sorted([drv for drv in pyodbc.drivers() if "ODBC Driver " in drv and " for SQL Server" in drv])
if len(drv) == 0:
raise Exception("No 'ODBC Driver XX for SQL Server' found.")
return drv[-1]
print(select_driver()) # ODBC Driver 17 for SQL Server
# Define Experiment ID list that we need to run
exp_id_list = ["tpc_h_highshift"]#["tpc_h_skew_static_10_MAB"]
# Comparing components
OPTIMAL = constants.COMPONENT_OPTIMAL in configs.components
TA_OPTIMAL = constants.COMPONENT_TA_OPTIMAL in configs.components
TA_FULL = constants.COMPONENT_TA_FULL in configs.components
TA_CURRENT = constants.COMPONENT_TA_CURRENT in configs.components
TA_SCHEDULE = constants.COMPONENT_TA_SCHEDULE in configs.components
MAB = constants.COMPONENT_MAB in configs.components
NO_INDEX = constants.COMPONENT_NO_INDEX in configs.components
RL = constants.COMPONENT_DDQN in configs.components
# Generate form saved reports
FROM_FILE = False
SEPARATE_EXPERIMENTS = False
PLOT_LOG_Y = False
PLOT_MEASURE = (constants.MEASURE_BATCH_TIME, constants.MEASURE_QUERY_EXECUTION_COST,
constants.MEASURE_INDEX_CREATION_COST, constants.MEASURE_INDEX_USAGE, constants.MEASURE_INDEX_USAGE_ROWS)
UNIFORM = False
exp_report_list = []
for i in range(len(exp_id_list)):
if SEPARATE_EXPERIMENTS:
exp_report_list = []
experiment_folder_path = helper.get_experiment_folder_path(exp_id_list[i])
helper.change_experiment(exp_id_list[i])
reload(configs)
reload(logging)
OPTIMAL = constants.COMPONENT_OPTIMAL in configs.components
TA_OPTIMAL = constants.COMPONENT_TA_OPTIMAL in configs.components
TA_FULL = constants.COMPONENT_TA_FULL in configs.components
TA_CURRENT = constants.COMPONENT_TA_CURRENT in configs.components
TA_SCHEDULE = constants.COMPONENT_TA_SCHEDULE in configs.components
MAB = constants.COMPONENT_MAB in configs.components
NO_INDEX = constants.COMPONENT_NO_INDEX in configs.components
DDQN = constants.COMPONENT_DDQN in configs.components
# configuring the logger
if not FROM_FILE:
logging.basicConfig(
filename=experiment_folder_path + configs.experiment_id + '.log',
filemode='w', format='%(asctime)s - %(levelname)s - %(message)s')
logging.getLogger().setLevel(constants.LOGGING_LEVEL)
if FROM_FILE:
with open(experiment_folder_path + "reports.pickle", "rb") as f:
exp_report_list = exp_report_list + pickle.load(f)
else:
print("Currently running: ", exp_id_list[i])
# Running MAB
if MAB:
Simulators = {}
for mab_version in configs.mab_versions:
Simulators[mab_version] = (getattr(__import__(mab_version, fromlist=['Simulator']), 'Simulator'))
for version, Simulator in Simulators.items():
#version_number = version.split("_v", 1)[1]
exp_report_mab = ExpReport(configs.experiment_id,
constants.COMPONENT_MAB + version + exp_id_list[i], configs.reps,
configs.rounds)
for r in range(configs.reps):
simulator = Simulator()
results, total_workload_time = simulator.run(exp_report_list, version, exp_id_list[i])
temp = DataFrame(results, columns=[constants.DF_COL_BATCH, constants.DF_COL_MEASURE_NAME,
constants.DF_COL_MEASURE_VALUE])
temp.append([-1, constants.MEASURE_TOTAL_WORKLOAD_TIME, total_workload_time])
temp[constants.DF_COL_REP] = r
exp_report_mab.add_data_list(temp)
exp_report_list.append(exp_report_mab)
# Running No Index
if NO_INDEX:
exp_report_no_index = ExpReport(configs.experiment_id, constants.COMPONENT_NO_INDEX + exp_id_list[i], configs.reps,
configs.rounds)
for r in range(configs.reps):
results, total_workload_time = ConfigRunner.run("no_index.sql", uniform=UNIFORM)
temp = DataFrame(results, columns=[constants.DF_COL_BATCH, constants.DF_COL_MEASURE_NAME,
constants.DF_COL_MEASURE_VALUE])
temp.append([-1, constants.MEASURE_TOTAL_WORKLOAD_TIME, total_workload_time])
temp[constants.DF_COL_REP] = r
exp_report_no_index.add_data_list(temp)
exp_report_list.append(exp_report_no_index)
# Running Optimal
if OPTIMAL:
exp_report_optimal = ExpReport(configs.experiment_id, constants.COMPONENT_OPTIMAL + exp_id_list[i], configs.reps, configs.rounds)
for r in range(configs.reps):
results, total_workload_time = ConfigRunner.run("optimal_config.sql", uniform=UNIFORM)
temp = DataFrame(results, columns=[constants.DF_COL_BATCH, constants.DF_COL_MEASURE_NAME,
constants.DF_COL_MEASURE_VALUE])
temp.append([-1, constants.MEASURE_TOTAL_WORKLOAD_TIME, total_workload_time])
temp[constants.DF_COL_REP] = r
exp_report_optimal.add_data_list(temp)
exp_report_list.append(exp_report_optimal)
# Running DTA Optimal
if TA_OPTIMAL:
exp_report_ta = ExpReport(configs.experiment_id, constants.COMPONENT_TA_OPTIMAL + exp_id_list[i], configs.reps, configs.rounds)
for r in range(configs.reps):
dta_runner = DTARunner(configs.ta_runs, workload_type=constants.TA_WORKLOAD_TYPE_OPTIMAL)
results, total_workload_time = dta_runner.run()
temp = DataFrame(results, columns=[constants.DF_COL_BATCH, constants.DF_COL_MEASURE_NAME,
constants.DF_COL_MEASURE_VALUE])
temp.append([-1, constants.MEASURE_TOTAL_WORKLOAD_TIME, total_workload_time])
temp[constants.DF_COL_REP] = r
exp_report_ta.add_data_list(temp)
exp_report_list.append(exp_report_ta)
# Running DTA Full
if TA_FULL:
exp_report_ta = ExpReport(configs.experiment_id, constants.COMPONENT_TA_FULL + exp_id_list[i], configs.reps,
configs.rounds)
for r in range(configs.reps):
dta_runner = DTARunner([0], workload_type=constants.TA_WORKLOAD_TYPE_FULL)
results, total_workload_time = dta_runner.run()
temp = DataFrame(results, columns=[constants.DF_COL_BATCH, constants.DF_COL_MEASURE_NAME,
constants.DF_COL_MEASURE_VALUE])
temp.append([-1, constants.MEASURE_TOTAL_WORKLOAD_TIME, total_workload_time])
temp[constants.DF_COL_REP] = r
exp_report_ta.add_data_list(temp)
exp_report_list.append(exp_report_ta)
# Running DTA Current
if TA_CURRENT:
exp_report_ta = ExpReport(configs.experiment_id, constants.COMPONENT_TA_CURRENT + exp_id_list[i],
configs.reps, configs.rounds)
for r in range(configs.reps):
dta_runner = DTARunner(configs.ta_runs, workload_type=constants.TA_WORKLOAD_TYPE_CURRENT)
results, total_workload_time = dta_runner.run()
temp = DataFrame(results, columns=[constants.DF_COL_BATCH, constants.DF_COL_MEASURE_NAME,
constants.DF_COL_MEASURE_VALUE])
temp.append([-1, constants.MEASURE_TOTAL_WORKLOAD_TIME, total_workload_time])
temp[constants.DF_COL_REP] = r
exp_report_ta.add_data_list(temp)
exp_report_list.append(exp_report_ta)
# Running DTA Schedule (everything from last run)
if TA_SCHEDULE:
exp_report_ta = ExpReport(configs.experiment_id, constants.COMPONENT_TA_SCHEDULE + exp_id_list[i],
configs.reps, configs.rounds)
for r in range(configs.reps):
dta_runner = DTARunner(configs.ta_runs, workload_type=constants.TA_WORKLOAD_TYPE_SCHEDULE)
results, total_workload_time = dta_runner.run()
temp = DataFrame(results, columns=[constants.DF_COL_BATCH, constants.DF_COL_MEASURE_NAME,
constants.DF_COL_MEASURE_VALUE])
temp.append([-1, constants.MEASURE_TOTAL_WORKLOAD_TIME, total_workload_time])
temp[constants.DF_COL_REP] = r
exp_report_ta.add_data_list(temp)
exp_report_list.append(exp_report_ta)
# Running DDQN
if DDQN:
from simulation.sim_ddqn_v3 import Simulator as DDQNSimulator
exp_report_mab = ExpReport(configs.experiment_id, constants.COMPONENT_MAB + exp_id_list[i],
configs.reps, configs.rounds)
for r in range(configs.reps):
simulator = DDQNSimulator()
results, total_workload_time = simulator.run()
temp = DataFrame(results, columns=[constants.DF_COL_BATCH, constants.DF_COL_MEASURE_NAME,
constants.DF_COL_MEASURE_VALUE])
temp.append([-1, constants.MEASURE_TOTAL_WORKLOAD_TIME, total_workload_time])
temp[constants.DF_COL_REP] = r
exp_report_mab.add_data_list(temp)
exp_report_list.append(exp_report_mab)
# Save results
with open(experiment_folder_path + "reports.pickle", "wb") as f:
pickle.dump(exp_report_list, f)
if SEPARATE_EXPERIMENTS:
helper.plot_exp_report(configs.experiment_id, exp_report_list, PLOT_MEASURE, PLOT_LOG_Y)
helper.create_comparison_tables(configs.experiment_id, exp_report_list)
# plot line graphs
if not SEPARATE_EXPERIMENTS:
helper.plot_exp_report(configs.experiment_id, exp_report_list, PLOT_MEASURE, PLOT_LOG_Y)
helper.create_comparison_tables(configs.experiment_id, exp_report_list)