def main(args):
logger = logging.getLogger()
file_log_handler = logging.FileHandler('Train.log')
logger.addHandler(file_log_handler)
logger.setLevel('INFO')
formatter = logging.Formatter()
file_log_handler.setFormatter(formatter)
stderr_log_handler = logging.StreamHandler(sys.stdout)
logger.addHandler(stderr_log_handler)
stderr_log_handler.setFormatter(formatter)
# class_dic : {'class name ': instance number of test/train set}
_, test_class_dic, _, _ = class_counter(args, 'test')
_, train_class_dic, _, _ = class_counter(args, 'train')
#print(train_class_dic)
kwargs = {'num_workers':4, 'pin_memory':True}
# the class order of loaded ptcloud is consistent with train/test_class_dic's key
ptcloud = np.load(args.ptcloud_path)
train_set = ptcloud['train']
test_set = ptcloud['test']
assert train_set.shape[0] == sum(train_class_dic.values())
assert test_set.shape[0] == sum(test_class_dic.values())
train_set = torch.from_numpy(train_set).to(args.device)
test_set = torch.from_numpy(test_set).to(args.device)
model = Oracle_NN(args.device, logger)
train_idx = 0 # front index of a train class
test_idx = 0 # front index of a test class
train_slicer = 0 # back index of a train class
test_slicer = 0 # back index of a test class
class_loss = {}
for idx, key in enumerate(train_class_dic):
train_slicer += train_class_dic[key]
test_slicer += test_class_dic[key]
logger.info('key:{},train_idx:{},train_slicer:{},num:{}'.format(key, train_idx, train_slicer, train_class_dic[key]))
logger.info('key:{},test_idx:{},test_slicer:{},num:{}'.format(key, test_idx, test_slicer, test_class_dic[key]))
# input specific class train set and test set to model.forward
avg_loss, class_prediction = model.forward(train_set[train_idx:train_slicer], test_set[test_idx:test_slicer])
if idx == 0:
prediction = class_prediction
else:
prediction = torch.cat((prediction, class_prediction),0)
train_idx += train_class_dic[key]
test_idx += test_class_dic[key]
class_loss[key] = avg_loss.item()
logger.info('key:{}, avg loss:{}'.format(key, class_loss[key]))
logger.info('final avg loss: {}'.format(avg))
prediction = prediction.cpu().numpy()
np.save('prediction.npy', prediction)
np.save('class_loss.npy', class_loss)
def encode_sentences(self, sentences):
"""Encode a list of sentences
Args:
sentences: the list of sentences
Returns:
an (N, d) numpy matrix of sentence encodings.
"""
encodings = []
glog.setLevel("ERROR")
for i in tqdm(range(0, len(sentences), self._batch_size),
"encoding sentence batches"):
encodings.append(
self._encoder_client.encode(sentences[i:i + self._batch_size]))
glog.setLevel("INFO")
print('DEBUG SBERT SIZE:', np.vstack(encodings).shape)
return l2_normalize(np.vstack(encodings))
def __init__(self, model_dir=None):
""" Init the parser given the model folder.
:param model_dir: Path to the folder storing the model, downloaded by
RerankingParser.fetch_and_load() method.
:type model_dir: str
"""
if model_dir is None:
# Use 'model/GENIA+PubMed' under current folder as default.
filepath = os.path.realpath(__file__)
dirpath = os.path.dirname(filepath)
model_dir = os.path.join(dirpath, 'model/GENIA+PubMed')
glog.setLevel(glog.INFO)
glog.info('Loading model from ' + model_dir)
# BLLIP parser doesn't support multi-thread parsing, and we will
# run it as one thread in one process (in both nlputils and spark),
# so this should be fine. It will raise a RunTimeError when initialize
# twice.
self.reranking_parser = \
bllipparser.RerankingParser.from_unified_model_dir(model_dir)
def get_logger(logPath='./logs', fileName=None, logLevel="INFO", is_Del=True):
surName = datetime.datetime.now().strftime('%m-%d-%y_%H-%M-%S')
if fileName is None:
fileName = surName
else:
fileName = fileName + '-' + surName
logFile = "{0}/{1}.log".format(logPath, fileName)
is_Exist_Path = os.path.exists(logPath)
is_Exist_File = os.path.exists(logFile)
if not is_Exist_Path:
os.system('mkdir ' + logPath)
if is_Del and is_Exist_File:
os.system('rm ' + logFile)
handler = glog.logging.FileHandler(logFile)
glog.logger.addHandler(handler)
glog.setLevel(logLevel)
glog.info("Happy Logging!")
return glog, logFile
steps = 0
total_reward = 0
while not done:
state = torch.FloatTensor(state).unsqueeze(0).to(device)
dist, _ = agent.model(state)
next_state, reward, done, _ = env.step(dist.sample().cpu().numpy()[0])
state = next_state
steps +=1
total_reward += reward
mean_reward = total_reward / steps
return [total_reward, mean_reward]
'''---------------------------------------------------------------'''
# Set to INFO for debugging
log.setLevel("WARN")
'''---------------Copy settings file to data folder---------------'''
path = os.path.dirname(real_lsd.__file__)
abs_path = path + '/envs/settings/landing/cpptest.json'
cp_path = '/media/scratch1/nasib/data/'
list_files = subprocess.run(["cp", abs_path, cp_path])
log.warn("The exit code was: %d" % list_files.returncode)
'''---------------------------------------------------------------'''
'''---------------Check cuda availability/set device--------------'''
use_cuda = torch.cuda.is_available()
device = torch.device("cuda" if use_cuda else "cpu")
'''---------------------------------------------------------------'''
# initialise environment
# Train binary classifiers.
self.binary_learner.learn(binary_class_data, binary_class_label, 10)
def learn(self, epoch=10):
for i in range(epoch):
glog.info('Epoch {}'.format(i))
self.e_step_result = []
for f in range(self.fold):
self.e_step(f)
for f in range(self.fold):
self.m_step(f)
if __name__ == '__main__':
glog.setLevel(glog.DEBUG)
train_file = sys.argv[1]
test_file = sys.argv[2]
eval_file = sys.argv[3]
# Used for training the binary classifier at entity pair level.
relation_name = sys.argv[4]
# Iterator for training data.
iter_train = MultiInstanceIterator('train')
iter_train.base_iter.set_min_feature_freq(2)
iter_train.base_iter.set_feature_set('miml')
iter_train.base_iter.log_setting()
iter_train.base_iter.load_instance(train_file)
iter_train.load_as_training_data(shuffle_entity_pair=True)
def _main():
parser = argparse.ArgumentParser()
parser.add_argument('--version',
'-V',
action='store_true',
help='show version')
parser.add_argument('--reset-config',
action='store_true',
help='reset qt config')
parser.add_argument(
'--logger-level',
default='info',
choices=['debug', 'info', 'warning', 'fatal', 'error'],
help='logger level',
)
parser.add_argument('filename', nargs='?', help='image or label filename')
parser.add_argument(
'--output',
'-O',
'-o',
help='output file or directory (if it ends with .json it is '
'recognized as file, else as directory)')
default_config_file = os.path.join(os.path.expanduser('~'), '.labelmerc')
parser.add_argument(
'--config',
dest='config_file',
help='config file (default: %s)' % default_config_file,
default=default_config_file,
)
# config for the gui
parser.add_argument(
'--nodata',
dest='store_data',
action='store_false',
help='stop storing image data to JSON file',
default=argparse.SUPPRESS,
)
parser.add_argument(
'--autosave',
dest='auto_save',
action='store_true',
help='auto save',
default=argparse.SUPPRESS,
)
parser.add_argument(
'--nosortlabels',
dest='sort_labels',
action='store_false',
help='stop sorting labels',
default=argparse.SUPPRESS,
)
parser.add_argument(
'--flags',
help='comma separated list of flags OR file containing flags',
default=argparse.SUPPRESS,
)
parser.add_argument(
'--labels',
help='comma separated list of labels OR file containing labels',
default=argparse.SUPPRESS,
)
parser.add_argument(
'--validatelabel',
dest='validate_label',
choices=['exact', 'instance'],
help='label validation types',
default=argparse.SUPPRESS,
)
parser.add_argument(
'--keep-prev',
action='store_true',
help='keep annotation of previous frame',
default=argparse.SUPPRESS,
)
parser.add_argument(
'--epsilon',
type=float,
help='epsilon to find nearest vertex on canvas',
default=argparse.SUPPRESS,
)
args = parser.parse_args()
if args.version:
print('{0} {1}'.format(__appname__, __version__))
sys.exit(0)
log.setLevel(getattr(logging, args.logger_level.upper()))
if hasattr(args, 'flags'):
if os.path.isfile(args.flags):
with codecs.open(args.flags, 'r', encoding='utf-8') as f:
args.flags = [l.strip() for l in f if l.strip()]
else:
args.flags = [l for l in args.flags.split(',') if l]
if hasattr(args, 'labels'):
if os.path.isfile(args.labels):
with codecs.open(args.labels, 'r', encoding='utf-8') as f:
args.labels = [l.strip() for l in f if l.strip()]
else:
args.labels = [l for l in args.labels.split(',') if l]
config_from_args = args.__dict__
config_from_args.pop('version')
reset_config = config_from_args.pop('reset_config')
filename = config_from_args.pop('filename')
output = config_from_args.pop('output')
config_file = config_from_args.pop('config_file')
config = get_config(config_from_args, config_file)
if not config['labels'] and config['validate_label']:
log.error('--labels must be specified with --validatelabel or '
'validate_label: true in the config file '
'(ex. ~/.labelmerc).')
sys.exit(1)
output_file = None
output_dir = None
if output is not None:
if output.endswith('.json'):
output_file = output
else:
output_dir = output
app = QtWidgets.QApplication(sys.argv)
app.setApplicationName(__appname__)
app.setWindowIcon(newIcon('icon'))
win = MainWindow(
config=config,
filename=filename,
output_file=output_file,
output_dir=output_dir,
)
if reset_config:
log.info('Resetting Qt config: %s' % win.settings.fileName())
win.settings.clear()
sys.exit(0)
win.show()
win.raise_()
sys.exit(app.exec_())
def main():
parser = argparse.ArgumentParser(
description="Upload MLBF files to Kinto as records"
)
group = parser.add_mutually_exclusive_group(required=True)
group.add_argument("--crlite", action="store_true", help="Perform a CRLite update")
group.add_argument(
"--intermediates", action="store_true", help="Perform an Intermediate CA update"
)
crlite_group = parser.add_argument_group("crlite", "crlite upload arguments")
crlite_group.add_argument("--noop", action="store_true", help="Don't update Kinto")
crlite_group.add_argument(
"--download-path",
type=Path,
default=Path(tempfile.TemporaryDirectory().name),
help="Path to temporarily store CRLite downloaded artifacts",
)
int_group = parser.add_argument_group(
"intermediates", "intermediates upload arguments"
)
int_group.add_argument(
"--debug", action="store_true", help="Enter a debugger during processing"
)
int_group.add_argument(
"--delete", action="store_true", help="Delete entries that are now missing"
)
int_group.add_argument(
"--export", help="Export intermediate set inspection files to this folder"
)
parser.add_argument("--filter-bucket", default="crlite_filters")
parser.add_argument("--verbose", "-v", help="Be more verbose", action="store_true")
signer_group = parser.add_mutually_exclusive_group()
signer_group.add_argument(
"--request-review",
action="store_true",
help="Mark the Kinto collection for signature when done",
)
signer_group.add_argument(
"--approve-sign",
action="store_true",
help="Approve the Kinto collection for signing",
)
args = parser.parse_args()
if args.verbose:
log.setLevel("DEBUG")
if args.noop:
log.info("The --noop flag is set, will not make changes.")
auth = {}
try:
ensureNonBlank(["KINTO_AUTH_TOKEN"])
auth = BearerTokenAuth(settings.KINTO_AUTH_TOKEN)
log.info("Using authentication bearer token")
except Exception:
ensureNonBlank(["KINTO_AUTH_USER", "KINTO_AUTH_PASSWORD"])
auth = HTTPBasicAuth(settings.KINTO_AUTH_USER, settings.KINTO_AUTH_PASSWORD)
log.info(
"Using username/password authentication. Username={}".format(
settings.KINTO_AUTH_USER
)
)
log.info(
f"Connecting... RO={settings.KINTO_RO_SERVER_URL}, RW={settings.KINTO_RW_SERVER_URL}"
)
rw_client = PublisherClient(
server_url=settings.KINTO_RW_SERVER_URL,
auth=auth,
bucket=settings.KINTO_BUCKET,
retry=5,
)
ro_client = PublisherClient(
server_url=settings.KINTO_RO_SERVER_URL,
bucket=settings.KINTO_BUCKET,
retry=5,
)
try:
if args.approve_sign:
if args.crlite:
crlite_sign(args=args, rw_client=rw_client)
elif args.intermediates:
intermediates_sign(args=args, rw_client=rw_client)
else:
parser.print_help()
return
if args.crlite:
publish_crlite(args=args, rw_client=rw_client, ro_client=ro_client)
if not args.noop and args.request_review:
log.info("Set for review")
rw_client.request_review_of_collection(
collection=settings.KINTO_CRLITE_COLLECTION,
)
return
if args.intermediates:
publish_intermediates(args=args, rw_client=rw_client, ro_client=ro_client)
return
parser.print_help()
except KintoException as ke:
log.error("An exception at Kinto occurred: {}".format(ke))
raise ke
except Exception as e:
log.error("A general exception occurred: {}".format(e))
raise e
def main(args):
# load data
starter_time = time.time()
kwargs = {'num_workers': 16, 'pin_memory': True}
print("loading train data ...")
train_loader = torch.utils.data.DataLoader(
what3d_dataset_views(data_basedir=args.data_basedir,
ptcloud_path=args.ptcloud_path,
img_path=args.img_path,
label_path=args.label_path,
splits_path=args.splits_path,
split_name='train',
class_path=args.class_path,
sample_ratio=args.sample_ratio,
image_height=args.image_size,
image_width=args.image_size,
views=list(args.views),
read_view=args.read_view,
points_num=args.pts_num,
mode=args.mode),
batch_size=args.train_batch_size,
shuffle=True,
**kwargs)
print("loading test data ...")
test_loader = torch.utils.data.DataLoader(what3d_dataset_views(
data_basedir=args.data_basedir,
ptcloud_path=args.ptcloud_path,
img_path=args.img_path,
label_path=args.label_path,
splits_path=args.splits_path,
split_name='test',
class_path=args.class_path,
sample_ratio=args.sample_ratio,
image_height=args.image_size,
image_width=args.image_size,
views=list(args.views),
read_view=args.read_view,
points_num=args.pts_num,
mode=args.mode),
batch_size=args.test_batch_size,
shuffle=False,
**kwargs)
print("loading val data ...")
val_loader = torch.utils.data.DataLoader(what3d_dataset_views(
data_basedir=args.data_basedir,
ptcloud_path=args.ptcloud_path,
img_path=args.img_path,
label_path=args.label_path,
splits_path=args.splits_path,
split_name='val',
class_path=args.class_path,
sample_ratio=args.sample_ratio,
image_height=args.image_size,
image_width=args.image_size,
views=list(args.views),
read_view=args.read_view,
points_num=args.pts_num,
mode=args.mode),
batch_size=args.val_batch_size,
shuffle=False,
**kwargs)
print("Initialize cache={}".format(time.time() - starter_time))
netG = GeneratorVanilla(grid_dims=(32, 32, 1),
resgen_width=512,
resgen_depth=5,
resgen_codelength=512,
class_num=55,
read_view=args.read_view,
folding_twice=args.folding_twice)
print(str(netG))
#netG.to(args.device)
netG = torch.nn.DataParallel(netG, device_ids=[0, 1])
logger = logging.getLogger()
logger.info('Number of parameters={}'.format(
count_parameter_num(netG.parameters())))
check_exist_or_mkdirs(args.log_dir)
file_log_handler = logging.FileHandler(
os.path.join(args.log_dir, args.log_filename))
logger.addHandler(file_log_handler)
#stderr_log_handler = logging.StreamHandler(sys.stdout)
#logger.addHandler(stderr_log_handler)
logger.setLevel('INFO')
formatter = logging.Formatter()
file_log_handler.setFormatter(formatter)
#stderr_log_handler.setFormatter(formatter)
logger.info(args)
# set solver and loss function
criterion_G = ChamferDistance().to(args.device)
criterion_C = torch.nn.CrossEntropyLoss().to(args.device)
criterion_M = torch.nn.MSELoss(reduction='mean')
optmizer_G = torch.optim.Adam(
netG.parameters(),
lr=args.lr_G,
betas=(args.G_adam_beta1, 0.999),
weight_decay=args.weight_decay,
)
lr_scheduler_G = torch.optim.lr_scheduler.StepLR(
optmizer_G,
step_size=args.lr_decay_step,
gamma=args.lr_decay_rate,
)
# train and test#
runner = TrainTester(
netG=netG,
criterion_G=criterion_G,
criterion_C=criterion_C,
criterion_M=criterion_M,
optmizer_G=optmizer_G,
lr_scheduler_G=lr_scheduler_G,
alpha=args.alpha,
logger=logger,
args=args,
)
if args.train:
runner.run(train_loader=train_loader,
test_loader=test_loader,
val_loader=val_loader)
logger.info('Training Done!')
if args.test:
runner.netG.load_state_dict(
torch.load(os.path.join(args.snapshot_dir,
"model_train_best.pth")))
runner.optimizer_G.load_state_dict(
torch.load(os.path.join(args.snapshot_dir,
"solver_train_best.pth")))
runner.test(epoch=args.total_epochs + 1,
loader=test_loader,
type='test')
logger.info('Testing Done!')
def main():
ray.init(ignore_reinit_error=True)
log.setLevel("DEBUG")
test_job_light()
test_job_m()
args = parser.parse_args()
# Load options from envs
for name in ['MASTER_ADDR', 'MASTER_PORT']:
setattr(args, name.lower(), os.environ[name])
for name in [
'RANK', 'WORLD_SIZE', 'LOCAL_RANK', 'WORLD_LOCAL_SIZE',
'WORLD_NODE_RANK'
]:
setattr(args, name.lower(), int(os.environ[name]))
args.node_rank = args.world_node_rank
log.setLevel(args.verbosity)
my_log(args)
ss.init()
if not args.cpu:
torch.cuda.set_device(args.local_rank)
torch.backends.cudnn.benchmark = True
if args.deterministic:
torch.manual_seed(args.rank)
np.random.seed(args.rank)
torch.backends.cudnn.deterministic = True
torch.cuda.manual_seed(args.rank)
dist.init_process_group(backend=args.backend)
__all__ = ['CephReader']
import ceph
import glog
glog.setLevel(glog.logging.ERROR)
class CephReader:
def __call__(self, path):
s3client = ceph.S3Client()
content = s3client.Get(path)
return content
def main(args):
# load data
starter_time = time.time()
kwargs = {'num_workers': 4, 'pin_memory': True}
print("loading train data ...")
trainset = PointCloudDataset(args.train_json)
train_loader = torch.utils.data.DataLoader(
trainset,
batch_size=args.batch_size,
shuffle=True,
**kwargs,
)
print("loading test data ...")
testset = PointCloudDataset(args.test_json)
test_loader = torch.utils.data.DataLoader(testset,
batch_size=args.test_batch_size,
shuffle=True,
**kwargs)
print("Initialize cache={}".format(time.time() - starter_time))
im_encoder = SegNet(input_channels=3, output_channels=3)
pointVAE = PointVAE(args=args)
#net = GeneratorVAE(
# encoder_dim=(3, 3),
# grid_dims=(32, 32, 1),
# Generate1_dims=259,
# Generate2_dims=1091,
# Generate3_dims=1219,
# args=args,
#)
net = GeneratorVAE(
im_encoder=im_encoder,
pointVAE=pointVAE,
encoder_dim=(3, 3),
grid_dims=(32, 32, 1),
Generate1_dims=259,
Generate2_dims=1091,
Generate3_dims=1219,
args=args,
)
#init_weights(net, init_type="xavier")
logger = logging.getLogger()
file_log_handler = logging.FileHandler(args.log_dir + args.log_filename)
logger.addHandler(file_log_handler)
stderr_log_handler = logging.StreamHandler(sys.stdout)
logger.addHandler(stderr_log_handler)
logger.setLevel('INFO')
formatter = logging.Formatter()
file_log_handler.setFormatter(formatter)
stderr_log_handler.setFormatter(formatter)
logger.info(args)
criterion_I = MaskedL1().to(args.device)
criterion_PTC = ChamfersDistance().to(args.device)
optimizer_image = torch.optim.Adam(
im_encoder.parameters(),
lr=args.lr_image,
betas=(args.adam_beta1, 0.999),
weight_decay=args.weight_decay,
)
optimizer_VAE = torch.optim.Adam(
pointVAE.parameters(),
lr=args.lr_vae,
betas=(args.adam_beta1, 0.999),
weight_decay=args.weight_decay,
)
optimizer = torch.optim.Adam(
net.parameters(),
lr=args.lr,
betas=(args.adam_beta1, 0.999),
weight_decay=args.weight_decay,
)
lr_scheduler = torch.optim.lr_scheduler.StepLR(
optimizer,
step_size=args.lr_decay_step,
gamma=args.lr_decay_rate,
)
# train and test
runner = TrainTester(
net=net,
criterion_I=criterion_I,
criterion_PTC=criterion_PTC,
optimizer=optimizer,
optimizer_image=optimizer_image,
optimizer_VAE=optimizer_VAE,
lr_scheduler=lr_scheduler,
logger=logger,
args=args,
)
if args.train:
runner.run(
train_loader=train_loader,
test_loader=test_loader,
)
logger.info('Training Done!')
if args.test:
runner.test(
epoch=args.total_epochs + 1,
loader=test_loader,
)
logger.info('Testing Done!')
action="store_true",
help="show plot window",
)
output_opts.add_argument("--save_plots",
action="store_true",
help="Save plots?")
output_opts.add_argument("--save_boxplots",
action="store_true",
help="Save boxplots?")
output_opts.add_argument("--save_results",
action="store_true",
help="Save results?")
main_parser = argparse.ArgumentParser(description="{}".format(basic_desc))
sub_parsers = main_parser.add_subparsers(dest="subcommand")
sub_parsers.required = True
return shared_parser
import argcomplete
import sys
if __name__ == '__main__':
log.setLevel('INFO')
parser = parser()
argcomplete.autocomplete(parser)
args = parser.parse_args()
if run(args):
sys.exit(os.EX_OK)
else:
raise Exception("Regression tests failed.")
import gns_pod_tool as tool
import sys
import glog
import gns_sum_orbdif as sum_dif
parser = argparse.ArgumentParser(description="Demo of argparse")
parser.add_argument('--year', '-yr', '-yyyy', default='2018')
parser.add_argument('--idoy', default='001')
parser.add_argument('--bin', default='001')
parser.add_argument('--xml', default='001')
parser.add_argument('--dir', default='001')
args = parser.parse_args()
print(args)
glog.setLevel("ERROR")
doy_xml = ""
xml_dir = args.xml # xml file for each day, format is yyyydoy.xml
log_dir = args.dir
bin_dir = args.bin
grt_bin = "/project/jdhuang/GNSS_software/GREAT/great_pco_L3/build/Bin/"
scp_dir = "/workfs/jdhuang/great_projects/e_all/scripts/great" # scripts dir
mod_xml = "/workfs/jdhuang/great_projects/e_all/scripts/great/gnss_tb_ge_3_freq.xml" # model xml
python_version = "python3.7" # python version
year = '%04s' % args.year
idoy = '%03s' % args.idoy
yeardoy = "%04s%03s" % (year, idoy)
def __call__(self, force=False):
f = inspect.currentframe().f_back
if not force and not f.f_globals['__name__'] == '__main__':
return
if 'main' in f.f_globals:
parser = None
parser = argparse.ArgumentParser(
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('--verbosity', type=str, default='ERROR')
parser.add_argument('--log_file', type=str)
parser.add_argument('--runid', type=str, default='default')
want_cache = 'cache' in f.f_globals and not is_python2
cache = None
if want_cache:
cache_argparse(parser)
if 'args' in f.f_globals:
args_func = f.f_globals['args']
args_func(parser)
parser.add_argument('other_args',
nargs=argparse.REMAINDER,
default=['--'])
if not is_python2:
argcomplete.autocomplete(parser)
flags = parser.parse_args()
if flags.other_args and flags.other_args[0] == '--':
flags.other_args = flags.other_args[1:]
self.flags = flags
for k, v in self.override_flags.items():
setattr(self.flags, k, v)
glog.setLevel(flags.verbosity)
if flags.log_file:
glog.logger.addHandler(logging.FileHandler(flags.log_file))
if 'flags' in f.f_globals:
f.f_globals['flags'] = flags
if want_cache:
cache = FileCacheDB.load_from_argparse(flags)
f.f_globals['cache'] = cache
main_func = f.f_globals['main']
if is_python2:
main_func()
else:
with ExitStack() as stack:
self.stack = stack
stack.enter_context(self.global_context)
script_name = sys.argv[0]
plog_filename = '/tmp/opa_plog_{}_{}.log'.format(
os.path.basename(script_name), flags.runid)
plog_file = open(plog_filename, 'w')
stack.enter_context(plog_file)
self.plog_file = plog_file
if cache:
stack.enter_context(cache)
main_func()
from chdrft.emu.structures import StructBuilder, CodeStructExtractor, Structure, StructBackend, g_data, MemBufAccessor from chdrft.emu.func_call import AsyncMachineCaller, FuncCallWrapperGen, FunctionCaller, AsyncFunctionCaller, SimpleBufGen import unicorn as uc import unicorn.x86_const as ucx86 from chdrft.emu.elf import ElfUtils from chdrft.emu.trace import TraceEvent, WatchedMem, WatchedRegs, Display import chdrft.emu.trace as trace import traceback as tb import numpy as np import binascii import struct from chdrft.tube.fifo import ManagedBidirectionalFifo from chdrft.tube.process import Process import curses.ascii import time glog.setLevel(glog.INFO) import itertools import os import ctypes import ctypes.util from chdrft.utils.swig import swig_unsafe from collections import defaultdict import pefile import yaml import chdrft.emu.structures as Structures import subprocess as sp from chdrft.dbg.gdbdebugger import GdbDebugger, launch_gdb from concurrent.futures import wait import tempfile import re import multiprocessing
import argparse
import os
import gns_pod_tool as tool
import gns_sum_orbdif as sum_dif
import sys
import glog
parser = argparse.ArgumentParser(description="Demo of argparse")
parser.add_argument('--year', '-yr', '-yyyy', default='2018')
parser.add_argument('--idoy', default='001')
parser.add_argument('--ilen', default='001') # 只测试过1天解
args = parser.parse_args()
print(args) # 输出传入参数
glog.setLevel("ERROR") # 只输出错误信息和提示信息
doy_xml = ""
mod_xml = "/workfs/jdhuang/great_projects/e_all/scripts/great/gnss_pod_e_157_if_ecom2_merge.xml" # model xml
scp_dir = "/workfs/jdhuang/great_projects/e_all/scripts/great" # scripts dir
prj_dir = "/workfs/jdhuang/great_projects/e_all/if_157_ecom2_merge" # project dir
xml_dir = "/workfs/jdhuang/great_projects/e_all/if_157_ecom2_merge/xml" # xml file for each day, format is yyyydoy.xml
prj_log = "/workfs/jdhuang/great_projects/e_all/if_15_ecom2_new" # 周跳探测的log文件
prj_log13 = "/workfs/jdhuang/great_projects/e_all/if_17_ecom2_merge" # 周跳探测的log文件
bin_dir = "/project/jdhuang/GNSS_software/GREAT/pod_3_freq_IF/build/Bin" # bin dir for great or other app
prd_dir = "./result" # result dir for saving products
log_dir = "./logInfo" # log dir for app and cmd log file
python_version = "python3.7" # python version
year = '%04s' % args.year # 打印为yyyy的格式
idoy = '%03s' % args.idoy # 打印为doy的格式,右对齐补零
WARN = WARNING
INFO = 20
DEBUG = 10
NOTSET = 0
""")
ap.add_argument("-f",
"--file",
required=True,
help="path to the input xlsx file to generate configs to")
ap.add_argument("-t",
"--template",
required=True,
help="path to the template to use to generate config files")
args = vars(ap.parse_args())
log.setLevel(args['verbosity'])
log.info("Logs initialized successfully.")
general_params.load_params(args)
log.info("General params initialized successfully.")
def main():
log.info("main")
from xls import get_config_data
from config import cfg_writer
for row in get_config_data():
cfg_writer.write(row, f"{row['hostname']}.txt")
log.info('main -> done')
trainGraphs.extend(graphFolds[i])
hist = trainThenValid(trainGraphs, validGraphs, '%d-%d' % (runId, f))
cvMetrics.append(hist)
avgMetrics = averageMetrics(cvMetrics)
df = pd.DataFrame.from_dict(avgMetrics)
histFile = open(
'%sGpu%sRun%s.csv' % (cmd_args.data, cmd_args.gpu_id, runId), 'w')
histFile.write("# %s\n" % str(gHP))
df.to_csv(histFile, index_label='Epoch', float_format='%.6f')
histFile.close()
if __name__ == '__main__':
log.setLevel("INFO")
random.seed(cmd_args.seed)
np.random.seed(cmd_args.seed)
torch.manual_seed(cmd_args.seed)
startTime = time.process_time()
graphs = loadGraphsMayCache(cmd_args.train_dir)
normalizeFeatures(graphs, isTestSet=False, operation=cmd_args.norm_op)
trainGraphs = filterOutNoEdgeGraphs(trainGraphs)
dataReadyTime = time.process_time() - startTime
log.info('Dataset ready takes %.2fs' % dataReadyTime)
for (id, hp) in enumerate(HyperParameterIterator(cmd_args.hp_path)):
for (key, val) in hp.items():
gHP[key] = val
_encode_data((g, mode, sr)) for g in group if g
]
for data, duration in data_duration_tuple:
if duration < 0.7:
# skip data with less than 800ms
continue
assert not np.isnan(
data).any(), 'Invalid data generated:: %s' % group
all_data.append(data)
total_duration_sec += duration
return total_duration_sec, all_data
p = mp.Pool(POOL_SIZE)
if __name__ == '__main__':
import fire
if os.environ.get('DEBUG'):
glog.setLevel('DEBUG')
sample_file = os.path.join(VCTK_ROOT, 'p225', 'p225_001.wav')
fire.Fire({
't':
lambda: process_wav(sample_file, mode=DataMode.mfcc_delta).all() and
'fin',
})
