def compute(self):
sample_source = self.sample_source
r = self.rep
d = sample_source.dim()
job_func = self.job_func
logger.info("computing. %s. r=%d, d=%d"%(job_func.__name__, r, d))
# sample_size is a global variable
tst_data = sample_source.sample(sample_size, seed=r)
tr, te = tst_data.split_tr_te(tr_proportion=tr_proportion, seed=r+20 )
prob_label = self.prob_label
test_result = job_func(sample_source, tr, te, r)
# create ScalarResult instance
result = SingleResult(test_result)
# submit the result to my own aggregator
self.aggregator.submit_result(result)
logger.info("done. ex2: %s, r=%d, d=%d, "%(job_func.__name__, r, d))
# save result
func_name = job_func.__name__
fname = '%s-%s-J%d_r%d_d%d_a%.3f_trp%.2f.p' \
%(prob_label, func_name, J, r, d, alpha, tr_proportion)
glo.ex_save_result(ex, test_result, prob_label, fname)
def compute(self):
sample_source = self.sample_source
r = self.rep
ni = self.ni
n = self.n
job_func = self.job_func
logger.info("computing. %s. r=%d, n=%d" % (job_func.__name__, r, n))
tst_data = sample_source.sample(n, seed=r)
tr, te = tst_data.split_tr_te(tr_proportion=tr_proportion, seed=r + 20)
prob_label = self.prob_label
test_result = job_func(prob_label, tr, te, r, ni, n)
# create ScalarResult instance
result = SingleResult(test_result)
# submit the result to my own aggregator
self.aggregator.submit_result(result)
logger.info("done. ex1: %s, r=%d, n=%d, " % (job_func.__name__, r, n))
# save result
func_name = job_func.__name__
fname = '%s-%s-J%d_r%d_n%d_a%.3f_trp%.2f.p' \
%(prob_label, func_name, J, r, n, alpha, tr_proportion)
glo.ex_save_result(ex, test_result, prob_label, fname)
def compute(self):
sample_source = self.sample_source
r = self.rep
d = sample_source.dim()
job_func = self.job_func
logger.info("computing. %s. r=%d, d=%d"%(job_func.__name__, r, d))
# sample_size is a global variable
tst_data = sample_source.sample(sample_size, seed=r)
tr, te = tst_data.split_tr_te(tr_proportion=tr_proportion, seed=r+20 )
prob_label = self.prob_label
test_result = job_func(sample_source, tr, te, r, self.n_locs)
# create ScalarResult instance
result = SingleResult(test_result)
# submit the result to my own aggregator
self.aggregator.submit_result(result)
logger.info("done. ex2: %s, r=%d, d=%d, "%(job_func.__name__, r, d))
# save result
func_name = job_func.__name__
J = self.n_locs
fname = '%s-%s-J%d_n%d_r%d_a%.3f_trp%.2f.p' \
%(prob_label, func_name, J, sample_size, r, alpha, tr_proportion)
glo.ex_save_result(ex, test_result, prob_label, fname)
def compute(self):
sample_source = self.sample_source
r = self.rep
ni = self.ni
n = self.n
job_func = self.job_func
logger.info("computing. %s. r=%d, n=%d"%(job_func.__name__, r, n))
tst_data = sample_source.sample(n, seed=r)
tr, te = tst_data.split_tr_te(tr_proportion=tr_proportion, seed=r+20 )
prob_label = self.prob_label
test_result = job_func(prob_label, tr, te, r, ni, n)
# create ScalarResult instance
result = SingleResult(test_result)
# submit the result to my own aggregator
self.aggregator.submit_result(result)
logger.info("done. ex1: %s, r=%d, n=%d, "%(job_func.__name__, r, n))
# save result
func_name = job_func.__name__
fname = '%s-%s-J%d_r%d_n%d_a%.3f_trp%.2f.p' \
%(prob_label, func_name, J, r, n, alpha, tr_proportion)
glo.ex_save_result(ex, test_result, prob_label, fname)
def compute(self):
r = self.rep
sample_source, nmax = get_sample_source(self.prob_label)
d = sample_source.dim()
job_func = self.job_func
logger.info("computing. %s. r=%d "%(job_func.__name__, r ))
tst_data = sample_source.sample(self.n, seed=r)
tr, te = tst_data.split_tr_te(tr_proportion=tr_proportion, seed=r+20 )
prob_label = self.prob_label
job_result = job_func(sample_source, tr, te, r)
# create ScalarResult instance
result = SingleResult(job_result)
# submit the result to my own aggregator
self.aggregator.submit_result(result)
logger.info("done. ex2: %s, r=%d "%(job_func.__name__, r))
# save result
func_name = job_func.__name__
fname = '%s-%s-J%d_r%d_d%d_a%.3f_trp%.2f.p' \
%(prob_label, func_name, J, r, d, alpha, tr_proportion)
glo.ex_save_result(ex, job_result, prob_label, fname)
def run_dataset(prob_label):
"""Run the experiment"""
sample_source, n = get_sample_source(prob_label)
# /////// submit jobs //////////
# create folder name string
home = os.path.expanduser("~")
foldername = os.path.join(home, "freqopttest_slurm", 'e%d' % ex)
logger.info("Setting engine folder to %s" % foldername)
# create parameter instance that is needed for any batch computation engine
logger.info("Creating batch parameter instance")
batch_parameters = BatchClusterParameters(foldername=foldername,
job_name_base="e%d_" % ex,
parameter_prefix="")
# Use the following line if Slurm queue is not used.
#engine = SerialComputationEngine()
engine = SlurmComputationEngine(batch_parameters, do_clean_up=True)
n_methods = len(method_job_funcs)
# repetitions x #methods
aggregators = np.empty((reps, n_methods), dtype=object)
d = sample_source.dim()
for r in range(reps):
for mi, f in enumerate(method_job_funcs):
# name used to save the result
func_name = f.__name__
fname = '%s-%s-J%d_r%d_d%d_a%.3f_trp%.2f.p' \
%(prob_label, func_name, J, r, d, alpha, tr_proportion)
if not is_rerun and glo.ex_file_exists(ex, prob_label, fname):
logger.info('%s exists. Load and return.' % fname)
test_result = glo.ex_load_result(ex, prob_label, fname)
sra = SingleResultAggregator()
sra.submit_result(SingleResult(test_result))
aggregators[r, mi] = sra
else:
# result not exists or rerun
job = Ex5Job(SingleResultAggregator(), prob_label, r, n, f)
agg = engine.submit_job(job)
aggregators[r, mi] = agg
# let the engine finish its business
logger.info("Wait for all call in engine")
engine.wait_for_all()
# ////// collect the results ///////////
logger.info("Collecting results")
test_results = np.empty((reps, n_methods), dtype=object)
for r in range(reps):
for mi, f in enumerate(method_job_funcs):
logger.info("Collecting result (%s, r=%d)" % (f.__name__, r))
# let the aggregator finalize things
aggregators[r, mi].finalize()
# aggregators[i].get_final_result() returns a SingleResult instance,
# which we need to extract the actual result
test_result = aggregators[r, mi].get_final_result().result
test_results[r, mi] = test_result
func_name = f.__name__
fname = '%s-%s-J%d_r%d_d%d_a%.3f_trp%.2f.p' \
%(prob_label, func_name, J, r, d, alpha, tr_proportion)
glo.ex_save_result(ex, test_result, prob_label, fname)
func_names = [f.__name__ for f in method_job_funcs]
func2labels = exglobal.get_func2label_map()
method_labels = [func2labels[f] for f in func_names if f in func2labels]
# save results
results = {
'results': test_results,
'n': n,
'data_fname': label2fname[prob_label],
'alpha': alpha,
'J': J,
'sample_source': sample_source,
'tr_proportion': tr_proportion,
'method_job_funcs': method_job_funcs,
'prob_label': prob_label,
'method_labels': method_labels
}
# class name
fname = 'ex%d-%s-me%d_J%d_rs%d_nma%d_d%d_a%.3f_trp%.2f.p' \
%(ex, prob_label, n_methods, J, reps, n, d, alpha, tr_proportion)
glo.ex_save_result(ex, results, fname)
logger.info('Saved aggregated results to %s' % fname)
def run_dataset(prob_label):
"""Run the experiment"""
sample_source, n = get_sample_source(prob_label)
# /////// submit jobs //////////
# create folder name string
home = os.path.expanduser("~")
foldername = os.path.join(home, "freqopttest_slurm", 'e%d'%ex)
logger.info("Setting engine folder to %s" % foldername)
# create parameter instance that is needed for any batch computation engine
logger.info("Creating batch parameter instance")
batch_parameters = BatchClusterParameters(
foldername=foldername, job_name_base="e%d_"%ex, parameter_prefix="")
# Use the following line if Slurm queue is not used.
#engine = SerialComputationEngine()
engine = SlurmComputationEngine(batch_parameters, do_clean_up=True)
n_methods = len(method_job_funcs)
# repetitions x #methods
aggregators = np.empty((reps, n_methods ), dtype=object)
d = sample_source.dim()
for r in range(reps):
for mi, f in enumerate(method_job_funcs):
# name used to save the result
func_name = f.__name__
fname = '%s-%s-J%d_r%d_d%d_a%.3f_trp%.2f.p' \
%(prob_label, func_name, J, r, d, alpha, tr_proportion)
if not is_rerun and glo.ex_file_exists(ex, prob_label, fname):
logger.info('%s exists. Load and return.'%fname)
test_result = glo.ex_load_result(ex, prob_label, fname)
sra = SingleResultAggregator()
sra.submit_result(SingleResult(test_result))
aggregators[r, mi] = sra
else:
# result not exists or rerun
job = Ex5Job(SingleResultAggregator(), prob_label, r, n, f)
agg = engine.submit_job(job)
aggregators[r, mi] = agg
# let the engine finish its business
logger.info("Wait for all call in engine")
engine.wait_for_all()
# ////// collect the results ///////////
logger.info("Collecting results")
test_results = np.empty((reps, n_methods), dtype=object)
for r in range(reps):
for mi, f in enumerate(method_job_funcs):
logger.info("Collecting result (%s, r=%d)" % (f.__name__, r ))
# let the aggregator finalize things
aggregators[r, mi].finalize()
# aggregators[i].get_final_result() returns a SingleResult instance,
# which we need to extract the actual result
test_result = aggregators[r, mi].get_final_result().result
test_results[r, mi] = test_result
func_name = f.__name__
fname = '%s-%s-J%d_r%d_d%d_a%.3f_trp%.2f.p' \
%(prob_label, func_name, J, r, d, alpha, tr_proportion)
glo.ex_save_result(ex, test_result, prob_label, fname)
func_names = [f.__name__ for f in method_job_funcs]
func2labels = exglobal.get_func2label_map()
method_labels = [func2labels[f] for f in func_names if f in func2labels]
# save results
results = {'results': test_results, 'n': n, 'data_fname':label2fname[prob_label],
'alpha': alpha, 'J': J, 'sample_source': sample_source,
'tr_proportion': 0.5, 'method_job_funcs': method_job_funcs,
'prob_label': prob_label, 'method_labels': method_labels}
# class name
fname = 'ex%d-%s-me%d_J%d_rs%d_nma%d_d%d_a%.3f_trp%.2f.p' \
%(ex, prob_label, n_methods, J, reps, n, d, alpha, tr_proportion)
glo.ex_save_result(ex, results, fname)
logger.info('Saved aggregated results to %s'%fname)
# aggregators[i].get_final_result() returns a SingleResult instance,
# which we need to extract the actual result
test_result = aggregators[r, mi].get_final_result().result
if isinstance(test_result, SingleResult):
test_result = test_result.result
if isinstance(test_result, SingleResult):
test_result = test_result.result
if isinstance(test_result, SingleResult):
test_result = test_result.result
test_results[r, mi] = test_result
func_name = f.__name__
fname = '%s-%s-J%d_r%d_d%d_a%.3f_trp%.2f.p' \
%(prob_label, func_name, J, r, d, alpha, tr_proportion)
glo.ex_save_result(ex, test_result, prob_label, fname)
func_names = [f.__name__ for f in method_job_funcs]
func2labels = exglobal.get_func2label_map()
method_labels = [func2labels[f] for f in func_names if f in func2labels]
# save results
results = {'results': test_results, 'n': n, 'data_fname':label2fname[prob_label],
'alpha': alpha, 'J': J, 'sample_source': sample_source,
'tr_proportion': tr_proportion, 'method_job_funcs': method_job_funcs,
'prob_label': prob_label, 'method_labels': method_labels}
# class name
fname = 'ex%d-%s-me%d_J%d_rs%d_nma%d_d%d_a%.3f_trp%.2f.p' \
%(ex, prob_label, n_methods, J, reps, n, d, alpha, tr_proportion)
glo.ex_save_result(ex, results, fname)
logger.info('Saved aggregated results to %s'%fname)
