def __get_info(self):
"""Retrieve information
"""
start_date = utils.get_date_str(self.date)
if self.duration < 0:
search_date = ['>=' + start_date]
else:
end_date = self.date + timedelta(self.duration)
today = utils.get_date_ymd('today')
if end_date > today:
search_date = ['>=' + start_date]
else:
search_date = [
'>=' + start_date, '<' + utils.get_date_str(end_date)
]
nb_hits = []
socorro.SuperSearch(
params={
'product': self.product,
'signature': '=' + self.signature,
'date': search_date,
'release_channel': self.channel,
'_results_number': 0
},
handler=lambda json: nb_hits.append(json['total'])).wait()
if nb_hits[0] > 1000:
nb_hits[0] = 1000
self.search = socorro.SuperSearch(params={
'product':
self.product,
'signature':
'=' + self.signature,
'date':
search_date,
'release_channel':
self.channel,
'_sort':
'build_id',
'_columns': ['uuid', 'topmost_filenames'],
'_facets': [
'platform_pretty_version', 'build_id', 'version',
'system_memory_use_percentage', 'cpu_name', 'cpu_info',
'reason', 'addons', 'uptime', 'url'
],
'_facets_size':
nb_hits[0],
'_results_number':
nb_hits[0]
},
handler=self.__handler)
def __get_info(self):
header = { 'Auth-Token': self.__get_apikey() }
self.results.append(self.session.get(self.SUPERSEARCH_URL,
params = { 'product': 'Firefox',
'signature': '=' + self.signature,
'date': ['>=' + utils.get_date_str(self.date),
'<' + utils.get_date_str(self.date + timedelta(self.day_delta))],
'release_channel': 'nightly',
'_sort': 'build_id',
'_columns': ['uuid', 'topmost_filenames'],
'_facets': ['platform_pretty_version', 'build_id', 'version', 'release_channel', 'system_memory_use_percentage', 'addons'],
'_results_number': 100,
},
headers = header,
timeout = self.TIMEOUT,
background_callback = self.__info_cb))
def get_feature_path(exp_name, dataset, ckpt_folder, nbit=32, date_str=None):
if date_str is None:
exp_path = "../../results/features/{}_{}".format(
exp_name, utils.get_date_str())
else:
exp_path = "../../results/features/{}_{}".format(exp_name, date_str)
ckpt_name = ckpt_folder.split("/")[-1]
folder = exp_path + "/{}/nbit_{}/{}".format(dataset, nbit, ckpt_name)
return folder
def main(argv=None):
if not FLAGS.save_name:
save_dir = utils.get_date_str()
else:
save_dir = FLAGS.save_name
save_locations = tf_easy_dir.tf_easy_dir(save_dir=save_dir)
if FLAGS.rewrite:
save_locations.clear_save_name()
def __get_info(self):
"""Retrieve information
"""
start_date = utils.get_date_str(self.date)
if self.duration < 0:
search_date = ['>=' + start_date]
else:
end_date = self.date + timedelta(self.duration)
today = utils.get_date_ymd('today')
if end_date > today:
search_date = ['>=' + start_date]
else:
search_date = ['>=' + start_date, '<' + utils.get_date_str(end_date)]
nb_hits = []
socorro.SuperSearch(params={'product': self.product,
'signature': '=' + self.signature,
'date': search_date,
'release_channel': self.channel,
'_results_number': 0},
handler=lambda json: nb_hits.append(json['total'])).wait()
if nb_hits[0] > 1000:
nb_hits[0] = 1000
self.search = socorro.SuperSearch(params={'product': self.product,
'signature': '=' + self.signature,
'date': search_date,
'release_channel': self.channel,
'_sort': 'build_id',
'_columns': ['uuid', 'topmost_filenames'],
'_facets': ['platform_pretty_version',
'build_id',
'version',
'system_memory_use_percentage',
'cpu_name',
'cpu_info',
'reason',
'addons',
'uptime',
'url'],
'_facets_size': nb_hits[0],
'_results_number': nb_hits[0]},
handler=self.__handler)
def main(argv=None):
if not FLAGS.save_name:
save_dir = os.path.join('Save', utils.get_date_str())
else:
save_dir = os.path.join('Save', FLAGS.save_name)
save_locations = tf_easy_dir.tf_easy_dir(save_dir=save_dir)
if FLAGS.rewrite:
save_locations.clear_save_name()
train(save_locations)
def get_pred_path_from_feature_path(exp_name,
dataset,
feat_folder,
nbit=32,
date_str=None):
if date_str is None:
exp_path = "../../results/predictions/{}_{}".format(
exp_name, utils.get_date_str())
else:
exp_path = "../../results/predictions/{}_{}".format(exp_name, date_str)
feat_name = feat_folder.split("/")[-1]
folder = exp_path + "/{}/nbit_{}/{}".format(dataset, nbit, feat_name)
return os.path.abspath(folder)
def loop_recompute(argo_exec, date_range, tenant, job, log):
"""
For a specific time period, loop and execute recomputations for each day
:param argo_exec: path to argo bin directory
:param date_range: list with all available dates included in the period
:param tenant: tenant name
:param job_set: list of tenant's available jobs
:param log: logger reference
"""
for dt in date_range:
date_arg = get_date_str(dt)
do_recompute(argo_exec, date_arg, tenant, job, log)
def generate_rtx_cardinfo(jokes_json, ori_url, today=True):
if None == jokes_json:
print("jokes_json is None")
return None
articles = []
if today:
articles.append({
'title': '如何正确地吐槽',
'url': ori_url,
'picurl': jokes_json['cover'],
'description': '今天是 ' + utils.get_date_str()
})
else:
articles.append({
'title': '往期沙雕',
'url': ori_url,
'picurl': jokes_json['cover']
})
data = {}
data['msgtype'] = 'news'
data['news'] = {}
data['news']['articles'] = articles
print(data)
return data
def main(args=None):
# default paths
fn_ar_cfg = "/etc/ar-compute-engine.conf"
arcomp_conf = "/etc/ar-compute/"
arcomp_exec = "/usr/libexec/ar-compute/"
stdl_exec = "/usr/libexec/ar-compute/bin"
pig_script_path = "/usr/libexec/ar-compute/pig/"
one_day_ago = utils.get_actual_date(args.date) - timedelta(days=1)
prev_date = utils.get_date_str(one_day_ago)
prev_date_under = utils.get_date_under(prev_date)
date_under = utils.get_date_under(args.date)
# Init configuration
cfg = utils.ArgoConfiguration(fn_ar_cfg)
cfg.load_tenant_db_conf(os.path.join(arcomp_conf, args.tenant + "_db_conf.json"))
# Init logging
log = init_log(cfg.log_mode, cfg.log_file, cfg.log_level, 'argo.job_status_detail')
local_cfg_path = arcomp_conf
# open job configuration file
json_cfg_file = open(
local_cfg_path + args.tenant + "_" + args.job + "_cfg.json")
json_cfg = json.load(json_cfg_file)
# Inform the user in wether argo runs locally or distributed
if cfg.mode == 'local':
log.info("ARGO compute engine runs in LOCAL mode")
log.info("computation job will be run locally")
else:
log.info("ARGO compute engine runs in CLUSTER mode")
log.info("computation job will be submitted to the hadoop cluster")
# Proposed hdfs pathways
hdfs_mdata_path = './' + args.tenant + "/mdata/"
hdfs_sync_path = './scratch/sync/' + args.tenant + \
"/" + args.job + "/" + date_under + "/"
# Proposed local pathways
local_mdata_path = '/tmp/' + args.tenant + "/mdata/"
local_sync_path = '/tmp/scratch/sync/' + args.tenant + \
'/' + args.job + '/' + date_under + '/'
local_cfg_path = arcomp_conf
if cfg.mode == 'cluster':
mode = 'cache'
mdata_path = hdfs_mdata_path
sync_path = hdfs_sync_path
cfg_path = hdfs_sync_path
else:
mode = 'local'
mdata_path = local_mdata_path
sync_path = local_sync_path
cfg_path = local_cfg_path
# dictionary with necessary pig parameters
pig_params = {}
pig_params['mdata'] = mdata_path + 'prefilter_' + date_under + '.avro'
pig_params['p_mdata'] = mdata_path + \
'prefilter_' + prev_date_under + '.avro'
pig_params['egs'] = sync_path + 'group_endpoints.avro'
pig_params['ggs'] = sync_path + 'group_groups.avro'
pig_params['mps'] = sync_path + 'poem_sync.avro'
pig_params['cfg'] = cfg_path + args.tenant + '_' + args.job + '_cfg.json'
pig_params['aps'] = cfg_path + args.tenant + '_' + args.job + '_ap.json'
pig_params['rec'] = cfg_path + 'recomputations_' + args.tenant + '_' + date_under + '.json'
pig_params['ops'] = cfg_path + args.tenant + '_ops.json'
pig_params['dt'] = args.date
pig_params['mode'] = mode
pig_params['flt'] = '1'
pig_params['mongo_status_metrics'] = cfg.get_mongo_uri('status', 'status_metrics')
pig_params['mongo_status_endpoints'] = cfg.get_mongo_uri('status', 'status_endpoints')
pig_params['mongo_status_services'] = cfg.get_mongo_uri('status', 'status_services')
pig_params['mongo_status_endpoint_groups'] = cfg.get_mongo_uri('status', 'status_endpoint_groups')
cmd_pig = []
# Append pig command
cmd_pig.append('pig')
# Append Pig local execution mode flag
if cfg.mode == "local":
cmd_pig.append('-x')
cmd_pig.append('local')
# Append Pig Parameters
for item in pig_params:
cmd_pig.append('-param')
cmd_pig.append(item + '=' + pig_params[item])
# Append Pig Executionable Script
cmd_pig.append('-f')
cmd_pig.append(pig_script_path + 'compute-status.pig')
# Command to clean a/r data from mongo
cmd_clean_mongo_status = [
os.path.join(stdl_exec, "mongo_clean_status.py"), '-d', args.date, '-t', args.tenant, '-r', json_cfg['id']]
# Command to upload sync data to hdfs
cmd_upload_sync = [os.path.join(
stdl_exec, "upload_sync.py"), '-d', args.date, '-t', args.tenant, '-j', args.job]
# Command to clean hdfs data
cmd_clean_sync = ['hadoop', 'fs', '-rm', '-r', '-f', hdfs_sync_path]
# Upload data to hdfs
log.info("Uploading sync data to hdfs...")
run_cmd(cmd_upload_sync, log)
# Clean data from mongo
log.info("Cleaning data from mongodb")
run_cmd(cmd_clean_mongo_status, log)
# Call pig
log.info("Submitting pig compute status detail job...")
run_cmd(cmd_pig, log)
# Cleaning hdfs sync data
if cfg.sync_clean == "true":
log.info("System configured to clean sync hdfs data after job")
run_cmd(cmd_clean_sync, log)
log.info("Execution of status job for tenant %s for date %s completed!",
args.tenant, args.date)
def get_folder_id(token):
folder_name = get_date_str()
folder_id = folder_exist(token, folder_name)
if folder_id:
return folder_id, folder_name
return add_folder(token, folder_name)
"Green laser cutter 32x18 in": "Green"
})
# group by cutters, then by start times within each cutter grouping
df = df.sort_values(by=["Cutter", "Start"])
# create dictionaries. Consider using default dicts later
# {string date: pandas.Timedelta total_time}
reds = {}
blues = {}
greens = {}
yellows = {}
for index, cutter, start, stop in df.itertuples():
print(start.day_name())
date = utils.get_date_str(start)
# Initialize dictionary with default time deltas of 0
if date not in reds or date not in blues or date not in greens or date not in yellows:
reds[date] = pandas.Timedelta(0)
blues[date] = pandas.Timedelta(0)
greens[date] = pandas.Timedelta(0)
yellows[date] = pandas.Timedelta(0)
# add time deltas to the respective dicts
if cutter == "Red":
reds[date] += stop - start
elif cutter == "Blue":
blues[date] += stop - start
elif cutter == "Green":
greens[date] += stop - start
elif cutter == "Yellow":
def main():
args = parse_args()
### SETTINGS #############################
n = 100000 # num images
frame_size = (64, 64)
patch_size = 18
# count_distrib = {1: 1}
count_distrib = {0: 1 / 3, 1: 1 / 3, 2: 1 / 3}
allow_overlap = True
##########################################
# Generate sprites and labels
print("generating sprites...")
if args.dataset_type == 'dsprites':
sprites, labels = generate_dsprites(patch_size)
elif args.dataset_type == 'binary_mnist':
sprites, labels = generate_binary_mnist(patch_size)
else:
raise NotImplementedError
# Show sprites
show_img_grid(8,
sprites,
random_selection=True,
fname='gen_{}_sprites.png'.format(get_date_str()))
# Create dataset
print("generating dataset...")
ch = sprites[0].shape[-1]
img_shape = (*frame_size, ch)
dataset, n_obj, labels = generate_multiobject_dataset(
n,
img_shape,
sprites,
labels,
count_distrib=count_distrib,
allow_overlap=allow_overlap)
print("done")
print("shape:", dataset.shape)
# Number of objects is part of the labels
labels['n_obj'] = n_obj
# Save dataset
print("saving...")
root = os.path.join('generated', args.dataset_type)
os.makedirs(root, exist_ok=True)
file_str = get_date_str()
fname = 'multi_' + args.dataset_type + '_' + file_str
fname = os.path.join(root, fname)
np.savez_compressed(fname, x=dataset, labels=labels)
print('done')
# Show samples and print their attributes
print("\nAttributes of saved samples:")
show_img_grid(4,
dataset,
labels,
fname='gen_{}_images.png'.format(get_date_str()))
# Show distribution of number of objects per image
plt.figure()
plt.hist(n_obj, np.arange(min(n_obj) - 0.5, max(n_obj) + 0.5 + 1, 1))
plt.title("Distribution of num objects per image")
plt.xlabel("Number of objects")
plt.savefig('gen_{}_distribution.png'.format(get_date_str()))
def main():
# user input
if not g_args.type:
g_args.type = utils.read_string_input(msg="type",
init_value=DEFAULT_TEMPLATE)
if not g_args.output:
g_args.output = utils.read_path_input(msg="file path")
g_args.author = utils.read_string_input(msg="author",
init_value=g_args.author)
if not g_args.description:
g_args.description = utils.read_string_input(
msg="description", init_value=HEADER_DESCRIPTION)
g_args.description = format_description(g_args.description)
if g_args.type != "bash":
g_args.copy_utils = utils.confirm("Copy utilities?",
"y" if g_args.copy_utils else "n")
template_path = utils.join_paths_str(g_args.script_dir,
TEMPLATE_OPTIONS[g_args.type])
if utils.exists_dir(g_args.output):
print(f"Error: target path is a directory!")
sys.exit(0)
elif utils.exists_file(g_args.output):
if not utils.confirm_delete_file(g_args.output, "n"):
utils.exit("Aborted")
out_folder = utils.get_file_path(g_args.output)
out_file = utils.get_file_name(g_args.output)
if not utils.exists_dir(out_folder):
utils.make_dir(g_args.output)
# copy template
utils.copy_to(template_path, g_args.output)
print(f"Created file {g_args.output}")
if g_args.type == "class":
utils.replace_file_text(g_args.output, CLASS_NAME, out_file)
if g_args.type != "bash":
if g_args.copy_utils:
utils_folder = PY_UTILS_DIR
out_py_utils_dir = utils.join_paths(out_folder, utils_folder)
utils.make_dir(out_py_utils_dir)
utils.copy_to(PY_UTILS_FILE, out_py_utils_dir)
print(f"Created file {out_py_utils_dir}/{PY_UTILS_FILE}")
else:
print("""
Important: Please make sure that python utils are available, i.e. inside PYTHONPATH.
Clone repository via: git clone https://github.com/amplejoe/py_utils.git
""")
# header information
date = utils.get_date_str()
utils.replace_file_text(g_args.output, HEADER_DATE, date)
utils.replace_file_text(g_args.output, HEADER_AUTHOR, g_args.author)
if g_args.description:
utils.replace_file_text(g_args.output, HEADER_DESCRIPTION,
g_args.description)
def train():
# import cifar10_simple_baseline as cifar10_model
cifar10_model = importlib.import_module(FLAGS.architecture)
if not FLAGS.save_name:
save_dir = os.path.join('Save',
utils.get_date_str() + FLAGS.architecture)
else:
save_dir = os.path.join('Save', FLAGS.save_name)
save_locations = tf_easy_dir.tf_easy_dir(save_dir=save_dir)
if FLAGS.rewrite:
save_locations.clear_save_name()
with tf.Graph().as_default() as graph:
global_step = tf.get_variable(name='gstep',
initializer=tf.constant(0),
trainable=False)
[batch_images, batch_labels] = cifar10_inputs.inputs(FLAGS.data_dir,
FLAGS.batch_size,
isTraining=True,
isRandom=True)
print 'size of image input: [{:s}]'.format(', '.join(
map(str,
batch_images.get_shape().as_list())))
print 'size of labels : [{:s}]'.format(', '.join(
map(str,
batch_labels.get_shape().as_list())))
print '-' * 32
sys.stdout.flush()
logits = cifar10_model.inference(batch_images, isTraining=True)
loss = cifar10_model.loss(logits=logits, labels=batch_labels)
train_op, lr = cifar10_model.train(loss, global_step)
# update_ops = tf.group(tf.get_collection(tf.GraphKeys.UPDATE_OPS))
# with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
# update_ops = tf.no_op()
correct_ones = cifar10_model.correct_ones(logits=logits,
labels=batch_labels)
# # debug:
# with tf.variable_scope("", reuse=True):
# moving_mean1 = tf.get_variable('inference/Conv/BatchNorm/moving_mean')
# moving_variance1 = tf.get_variable('inference/Conv/BatchNorm/moving_variance')
saver = tf.train.Saver(max_to_keep=None)
summary_op = tf.summary.merge_all()
config = tf_utils_inner.gpu_config(FLAGS.gpu_id)
with tf.Session(config=config) as sess:
sess.run(tf.variables_initializer(tf.global_variables()))
summary_writer = tf.summary.FileWriter(
logdir=save_locations.summary_save_dir, graph=sess.graph)
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
for i in range(FLAGS.max_steps):
_, loss_, correct_ones_, lr_ = sess.run(
[train_op, loss, correct_ones, lr])
assert not np.isnan(
loss_), 'Model diverged with loss = NaN, try again'
# if (i+1) % 10 == 0:
print '[{:s} -- {:08d}|{:08d}]\tloss : {:.3f}\t, l-rate: {:.6f}\tcorrect ones [{:d}|{:d}]'.format(
save_dir, i, FLAGS.max_steps, loss_, lr_, correct_ones_,
FLAGS.batch_size)
sys.stdout.flush()
# mm1, mv1 = sess.run([moving_mean1, moving_variance1])
# print 'Sum of moving mean: {:.6f} \t, moving variance: {:.06f}'.format(np.sum(mm1), np.sum(mv1))
if (i + 1 % 100) == 0:
summary_ = sess.run(summary_op)
summary_writer.add_summary(summary_,
global_step=global_step)
if (i + 1) % 2000 == 0:
save_path = os.path.join(save_locations.model_save_dir,
'model')
saver.save(sess=sess,
global_step=global_step,
save_path=save_path)
coord.request_stop()
coord.join(threads=threads)
print("spider failed")
myjokes = utils.get_qa_from_html(myhtml)
if None == myjokes:
print("parse failed")
rtx_md = push.generate_rtx_markdown(myjokes)
rtx_card = push.generate_rtx_cardinfo(myjokes, ori_url)
push.push_to_rtx(rtx_card)
push.push_to_rtx(rtx_md)
def timer_func():
global pushed
if utils.check_if_is_time(7, 30) and not pushed:
pushed = True
do_push()
import os
os._exit(0)
return True
return False
def timer_test_func():
print('timer')
if __name__ == "__main__":
while True:
print(utils.get_date_str() + " " + utils.get_time_str())
timer_func()
time.sleep(30)
def get_new_capture_folder(self):
rel_dir = self.sys_id + os.sep + get_date_str(
) + os.sep + self.session + os.sep + f"{self.segment:04d}"
folder = os.path.join(self.capture_dir_base, rel_dir)
os.makedirs(folder, exist_ok=True)
return folder
def main(argv=None):
if not FLAGS.save_name:
FLAGS.save_name = os.path.join('c3dSave', utils.get_date_str())
else:
FLAGS.save_name = os.path.join('c3dSave', FLAGS.save_name)
# print parameters
tf_utils.print_gflags(FLAGS=FLAGS)
NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN = len(
glob.glob(
os.path.join(FLAGS.data_dir,
'*.{:s}'.format(input_reader.TF_FORMAT))))
if NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN < 1:
print "Check file path"
return
steps_per_epoch = int(
math.ceil(NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN * 1.0 / FLAGS.batch_size))
print 'Each epoch consists {:d} Steps'.format(steps_per_epoch)
sys.stdout.flush()
lr_decay_every_n_step = int(steps_per_epoch * FLAGS.num_epoch_per_decay)
# save_dir = FLAGS.save_name
save_locations = tf_easy_dir.tf_easy_dir(save_dir=FLAGS.save_name)
if FLAGS.rewrite:
save_locations.clear_save_name()
with tf.Graph().as_default() as graph:
global_step = tf.get_variable(name='gstep',
initializer=tf.constant(0),
trainable=False)
batch_images, batch_labels, batch_filenames = input_reader.inputs(
FLAGS.data_dir, isTraining=True)
print 'size of image input: [{:s}]'.format(', '.join(
map(str,
batch_images.get_shape().as_list())))
print 'size of labels : [{:s}]'.format(', '.join(
map(str,
batch_labels.get_shape().as_list())))
print '-' * 32
sys.stdout.flush()
logits = c3d_model.inference_c3d(batch_images, isTraining=True)
loss = c3d_model.loss(logits=logits, labels=batch_labels)
train_op, lr = c3d_model.train(loss, global_step,
lr_decay_every_n_step)
correct_ones = c3d_model.correct_ones(logits=logits,
labels=batch_labels)
saver = tf.train.Saver(max_to_keep=None)
summary_op = tf.summary.merge_all()
config = tf_utils.gpu_config(FLAGS.gpu_id)
with tf.Session(config=config) as sess:
sess.run(tf.variables_initializer(tf.global_variables()))
summary_writer = tf.summary.FileWriter(
logdir=save_locations.summary_save_dir, graph=sess.graph)
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
print 'Training Start!'
sys.stdout.flush()
cum_loss = 0
cum_correct = 0
for i in range(FLAGS.max_steps):
_, loss_, correct_ones_ = sess.run(
[train_op, loss, correct_ones])
assert not np.isnan(
loss_), 'Model diverged with loss = NaN, try again'
cum_loss += loss_
cum_correct += correct_ones_
# update: print loss every epoch
if (i + 1) % steps_per_epoch == 0:
lr_ = sess.run(lr)
print '[{:s} -- {:08d}|{:08d}]\tloss : {:.3f}\t, correct ones [{:d}|{:d}], l-rate:{:.06f}'.format(
save_dir, i, FLAGS.max_steps,
cum_loss / steps_per_epoch, cum_correct,
NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN, lr_)
sys.stdout.flush()
cum_loss = 0
cum_correct = 0
if (i + 1 % 100) == 0:
summary_ = sess.run(summary_op)
summary_writer.add_summary(summary_,
global_step=global_step)
if (i + 1) % 2000 == 0:
save_path = os.path.join(save_locations.model_save_dir,
'model')
saver.save(sess=sess,
global_step=global_step,
save_path=save_path)
coord.request_stop()
coord.join(threads=threads)
