-
Notifications
You must be signed in to change notification settings - Fork 14
/
chemOS.py
executable file
·254 lines (189 loc) · 8.45 KB
/
chemOS.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
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
#!/usr/bin/env python
__author__ = 'Florian Hase'
#========================================================================
import sys
import copy
import time
import numpy as np
from threading import Thread
from Analyzer.analyzer import Analyzer
from BotManager.bot_manager import BotManager
from Communicator.communicator import Communicator
from DatabaseHandler.feedback_handler import FeedbackHandler
from DatabaseHandler.request_handler import RequestHandler
from DatabaseHandler.results_handler import ResultsHandler
from ParamGenerator.param_generator import ParamGenerator
from Utilities import defaults
from Utilities.misc import Printer
#========================================================================
np.set_printoptions(precision = 4)
#========================================================================
class ChemOS(Printer):
SETTINGS = defaults._DEFAULT_SETTINGS
num_submitted_jobs = 0
num_evaluated_jobs = 0
def __init__(self, verbose = True):
Printer.__init__(self, 'CHEM_OS', color = 'green')
self.verbose = verbose
# initialize bot manager
if self.verbose: self._print('initializing bot manager')
self.bot_manager = BotManager(settings = self.SETTINGS)
# FIXME: THE LINE BELOW DO NOT ALIGN WITH THE IDEAL WORKFLOW
self.bot_manager.boot_bots()
# initialize communicator
if self.verbose: self._print('initializing communicator')
self.communicator = Communicator(settings = self.SETTINGS)
# initialize feedback handler
if self.verbose: self._print('initializing feedback handler')
self.feedback_handler = FeedbackHandler(settings = self.SETTINGS)
# initialize request handler
if self.verbose: self._print('initializing request handler')
self.request_handler = RequestHandler(settings = self.SETTINGS)
# initialize results handler
if self.verbose: self._print('initializing results handler')
self.results_handler = ResultsHandler(settings = self.SETTINGS)
# initialize parameter generator
if self.verbose: self._print('initializing parameter generator')
self.param_generator = ParamGenerator(settings = self.SETTINGS)
# initialize analyzer
if self.verbose: self._print('initializing analyzer')
self.analyzer = Analyzer(settings = self.SETTINGS)
def start_communication_stream(self):
self._print('starting communication stream')
self.stream_thread = Thread(target = self.communicator.stream)
self.stream_thread.start()
def purge(self, exp_identifier):
# delete running parameter generations
self.param_generator.kill_running_instances(exp_identifier)
# delete running robots
self.bot_manager.kill_running_robots(exp_identifier)
# delete parameters in parameter database
self.param_generator.remove_parameters(exp_identifier)
# remove pending requests
self.request_handler.remove_requests(exp_identifier)
# remove pending results
self.results_handler.remove_results(exp_identifier)
# remove collected feedback
self.feedback_handler.remove_feedback(exp_identifier)
def _run(self):
# search for new requests
new_requests = copy.deepcopy(self.communicator.RECEIVED_REQUESTS)
for request in new_requests:
if request['kind'] == 'start':
self.request_handler.process_request(request)
elif request['kind'] == 'restart':
self.purge(request['exp_identifier'])
self.request_handler.process_request(request)
elif request['kind'] == 'stop':
self.purge(request['exp_identifier'])
elif request['kind'] == 'progress':
exp_identifier = request['exp_identifier']
observations = self.feedback_handler.get_observations(exp_identifier)
self.analyzer.analyze(request, observations)
for index in range(len(new_requests)):
self.communicator.RECEIVED_REQUESTS.pop()
# update list of new requests
new_requests = self.request_handler.get_pending_request()
if new_requests:
# check if there are bots available for each new request
for request in new_requests:
available_bot = self.bot_manager.get_available(request['exp_identifier'])
if available_bot:
# get parameters for this experiment
parameters, wait, retrain = self.param_generator.select_parameters(request['exp_identifier'])
if wait: continue
request['parameters'] = parameters
# now we have request, bot and parameters
self.bot_manager.submit(available_bot, request)
self.num_submitted_jobs += 1
# change status of parameters
self.request_handler.dump_parameters(request, parameters)
self.request_handler.label_processing(request)
# if the parameter database is exhausted we need to regenerate parameters
if retrain:
if self.verbose: self._print('parameter database exhausted for %s' % request['exp_identifier'])
if not self.param_generator.BUSY[request['exp_identifier']]:
# get observations for this experiment
observations = self.feedback_handler.get_observations(request['exp_identifier'])
if self.verbose: self._print('starting parameter generation process for %s' % request['exp_identifier'])
self.param_generator.TARGET_SPECS[request['exp_identifier']] = observations
self.param_generator.generate_new_parameters(request['exp_identifier'])
# check for completed experiments
processed_jobs = copy.deepcopy(self.bot_manager.PROCESSED_JOBS)
for job_id in processed_jobs:
info_dict = getattr(self.bot_manager, 'info_dict_%s' % job_id)
self.results_handler.process_results(info_dict)
self.communicator.notify_user(info_dict)
self.bot_manager.PROCESSED_JOBS.remove(job_id)
# analyze experimental results
# TODO -- this should go on a separate thread!
job_ids = self.results_handler.get_new_results()
for job_id in job_ids:
self.results_handler.analyze_new_results(job_id)
self.results_handler.set_all_to_used(job_id)
# check for analyzed experiments
processed_jobs = copy.deepcopy(self.results_handler.PROCESSED_JOBS)
for job_id in processed_jobs:
info_dict = getattr(self.results_handler, 'info_dict_%s' % job_id)
self.feedback_handler.process_feedback(info_dict)
self.results_handler.PROCESSED_JOBS.remove(job_id)
# analyze experimental results
# processed_jobs = copy.deepcopy(self.bot_manager.PROCESSED_JOBS)
# for job_id in processed_jobs:
# info_dict = getattr(self.bot_manager, 'info_dict_%s' % job_id)
# self.feedback_handler.process_feedback(info_dict)
# self.communicator.notify_user(info_dict)
# self.bot_manager.PROCESSED_JOBS.remove(job_id)
# check for new feedback
exp_identifiers = self.feedback_handler.get_new_feedback()
for exp_identifier in exp_identifiers:
if not self.param_generator.BUSY[exp_identifier]:
# get observations
observations = self.feedback_handler.get_observations(exp_identifier)
if self.verbose: self._print('starting parameter generation process for %s with %d observations' % (exp_identifier, len(observations)))
# submit generation process
self.param_generator.TARGET_SPECS[exp_identifier] = observations
self.param_generator.generate_new_parameters(exp_identifier)
# label feedback as used
self.feedback_handler.set_all_to_used(exp_identifier)
# check for analyzed experiments
analyzed_experiments = self.analyzer.get_analyzed_experiments()
for analyzed in analyzed_experiments:
# we need to notify the user
self.communicator.send(kind = 'analysis', request_details = analyzed['request_details'], file_names = [analyzed['progress_file']])
print('*****************************************')
def run(self, max_iter = 10**12):
if max_iter: self.max_iter = max_iter
self.start_communication_stream()
while self.num_submitted_jobs < self.max_iter:
try:
self._run()
time.sleep(2)
except (KeyboardInterrupt, SystemExit):
self._print('shutting down ..')
self._print('... turning off robots ...')
self.bot_manager.shutdown()
self._print('... completed shutdown')
break
self.num_submitted_jobs = -10
else:
self._print('completed all experiments: ')
self._print('shutting down ..')
self._print('... turning off robots ...')
self.bot_manager.shutdown()
self._print('... completed shutdown')
self._print('good night!')
#========================================================================
'''
settings ...
--> algorithm: dictionary with optimization algorithm parameters
--> param_database: dictionary, needs path and type
'''
#========================================================================
if __name__ == '__main__':
try:
max_iter = int(sys.argv[1])
except IndexError:
max_iter = 10**12
chem_os = ChemOS()
chem_os.run(max_iter)