def __init__(self, arguments, args):
self._arguments = arguments
if args.show_commands:
line()
log.debug(
f'Running the following command:\n{" ".join(self._arguments)}')
line()
def prepare_data(self):
# setup training set
if "timit" in self.hp.data:
train, val, test = TrainTestDataset.get_datasets(
path=self.hp.timit_path)
elif "buckeye" in self.hp.data:
train, val, test = TrainValTestDataset.get_datasets(
path=self.hp.buckeye_path, percent=self.hp.buckeye_percent)
else:
raise Exception("no such training data!")
if "libri" in self.hp.data:
libri_train = LibriSpeechDataset(path=self.hp.libri_path,
subset=self.hp.libri_subset,
percent=self.hp.libri_percent)
train = ConcatDataset([train, libri_train])
train.path = "\n\t+".join(
[dataset.path for dataset in train.datasets])
print(f"added libri ({len(libri_train)} examples)")
self.train_dataset = train
self.valid_dataset = val
self.test_dataset = test
line()
print("DATA:")
print(f"train: {self.train_dataset.path} ({len(self.train_dataset)})")
print(f"valid: {self.valid_dataset.path} ({len(self.valid_dataset)})")
print(f"test: {self.test_dataset.path} ({len(self.test_dataset)})")
line()
def configure_optimizers(self):
parameters = filter(lambda p: p.requires_grad, self.parameters())
if self.hp.optimizer == "sgd":
self.opt = optim.SGD(parameters,
lr=self.hparams.lr,
momentum=0.9,
weight_decay=5e-4)
elif self.hp.optimizer == "adam":
self.opt = optim.Adam(parameters,
lr=self.hparams.lr,
weight_decay=5e-4)
elif self.hp.optimizer == "ranger":
self.opt = optim_extra.Ranger(parameters,
lr=self.hparams.lr,
alpha=0.5,
k=6,
N_sma_threshhold=5,
betas=(.95, 0.999),
eps=1e-5,
weight_decay=0)
else:
raise Exception("unknown optimizer")
print(f"optimizer: {self.opt}")
line()
self.scheduler = optim.lr_scheduler.StepLR(
self.opt,
step_size=self.hp.lr_anneal_step,
gamma=self.hp.lr_anneal_gamma)
return [self.opt]
def get_user_by_username(ep, username, level):
utils.line()
print(SPACE * level + 'Get user={user}'.format(user=username))
print(
ep.get_doc_by_query(
'users',
{'username': {
'$regex': '^' + username + '$',
'$options': 'i'
}}))
position_id = ep.get_doc_by_query('users', {
'username': {
'$regex': '^' + username + '$',
'$options': 'i'
}
}).get('position')
group_id = ep.get_doc_by_query('userpositions', {
'_id': position_id
}).get('group')
account_id = ep.get_doc_by_query('usergroups', {
'_id': group_id
}).get('account')
organization_id = ep.get_doc_by_query('accounts', {
'_id': account_id
}).get('organization')
print(
'position_id={position_id}, group_id={group_id}, account_id={account_id}, organization={organization_id}'
.format(position_id=position_id,
group_id=group_id,
account_id=account_id,
organization_id=organization_id))
def delete_users_by_account_id(ep, account_id, level):
utils.line()
print(SPACE * level +
'Get users by account_id={account_id}'.format(account_id=account_id))
print(ep.get_doc_by_query('accounts', {'_id': account_id}))
organization_id = ep.get_doc_by_query('accounts', {
'_id': account_id
}).get('organization')
group_ids = ep.get_ids_by_query('usergroups', {'account': account_id},
'_id')
position_ids = []
for group_id in group_ids:
ids = ep.get_ids_by_query('userpositions', {'group': group_id}, '_id')
position_ids.append((group_id, ids))
# print('[(group, positions)]', position_ids)
users = []
for (group_id, positions) in position_ids:
for position_id in positions:
user = ep.get_doc_by_query('users', {'position': position_id})
username = user.get('username')
group_flag = ep.get_doc_by_query('usergroups', {
'_id': group_id
}).get('default')
user_id = user.get('_id')
isSuperAdmin = user.get('isSuperAdmin')
users.append((user_id, username, isSuperAdmin, position_id,
group_id, group_flag))
for i, (user_id, username, isSuperAdmin, position_id, group_id,
group_flag) in enumerate(users, 1):
print(ep.delete_one_by_query('users', {'username': username}))
print(ep.delete_one_by_query('userpositions', {'_id': position_id}))
print(ep.delete_one_by_query('usergroups', {'_id': group_id}))
print(ep.delete_one_by_query('accounts', {'_id': account_id}))
print(ep.delete_one_by_query('organizations', {'_id': organization_id}))
def generic_eval_end(self, outputs, mode):
metrics = {}
data = self.hp.data
for k, v in self.stats.items():
metrics[f"train_{k}"] = self.stats[k]["train"].get_stats()
metrics[f"{mode}_{k}"] = self.stats[k][mode].get_stats()
epoch = self.current_epoch + 1
metrics['epoch'] = epoch
metrics['current_lr'] = self.opt.param_groups[0]['lr']
line()
for pred_type in self.pr.keys():
if mode == "val":
(precision, recall, f1, rval), (width, prominence, distance) = self.pr[pred_type][mode].get_stats()
if rval > self.best_rval[pred_type][mode][0]:
self.best_rval[pred_type][mode] = rval, self.current_epoch
self.peak_detection_params[pred_type]["width"] = width
self.peak_detection_params[pred_type]["prominence"] = prominence
self.peak_detection_params[pred_type]["distance"] = distance
self.peak_detection_params[pred_type]["epoch"] = self.current_epoch
print(f"saving for test - {pred_type} - {self.peak_detection_params[pred_type]}")
else:
print(f"using pre-defined peak detection values - {pred_type} - {self.peak_detection_params[pred_type]}")
(precision, recall, f1, rval), _ = self.pr[pred_type][mode].get_stats(
width=self.peak_detection_params[pred_type]["width"],
prominence=self.peak_detection_params[pred_type]["prominence"],
distance=self.peak_detection_params[pred_type]["distance"],
)
# test has only one epoch so set it as best
# this is to get the overall best pred_type later
self.best_rval[pred_type][mode] = rval, self.current_epoch
metrics[f'{data}_{mode}_{pred_type}_f1'] = f1
metrics[f'{data}_{mode}_{pred_type}_precision'] = precision
metrics[f'{data}_{mode}_{pred_type}_recall'] = recall
metrics[f'{data}_{mode}_{pred_type}_rval'] = rval
metrics[f"{data}_{mode}_{pred_type}_max_rval"] = self.best_rval[pred_type][mode][0]
metrics[f"{data}_{mode}_{pred_type}_max_rval_epoch"] = self.best_rval[pred_type][mode][1]
# get best rval from all rval types and all epochs
best_overall_rval = -float("inf")
for pred_type, rval in self.best_rval.items():
if rval[mode][0] > best_overall_rval:
best_overall_rval = rval[mode][0]
metrics[f'{mode}_max_rval'] = best_overall_rval
for k, v in metrics.items():
print(f"\t{k:<30} -- {v}")
line()
# wandb.log(metrics)
output = OrderedDict({
'log': metrics
})
return output
def clear_ship_by_ship_id(ep, ship_id, level):
utils.line()
print(SPACE * level + 'Remove ship_id={ship_id}'.format(ship_id=ship_id))
voyage_ids = ep.get_ids_by_query('voyagesegments', {'ship': ship_id},
'_id')
level += 1
for i, voyage_id in enumerate(voyage_ids, 1):
utils.line()
print(SPACE * level + 'voyage {number}'.format(number=str(i)))
clear_voyage_by_voyage_id(ep, voyage_id, level)
def parse_post_details(self, post_detail_html):
soup = self.get_soup(post_detail_html)
post_header = soup.select_one('div#post-header>h1').text
post_content = soup.select_one('div.post-section')
print('\n')
print(Console.yellow(line(post_header)))
print(post_header)
print(Console.yellow(line(post_header)))
print('\n')
post_detail = post_content.text.strip()
print(post_detail)
def create_clips(video_path, output_folder, interval_seconds, clip_length):
# The output folder for the clips.
output_folder = os.path.join(output_folder, 'clips')
if not os.path.exists(video_path):
raise ClipError('The specified video file does not exist.')
if not os.path.exists(output_folder):
os.mkdir(output_folder)
provider = VideoInfoProvider(video_path)
duration = int(float(provider.get_duration()))
if interval_seconds > duration:
raise ClipError(
f'The interval ({interval_seconds}s) may not be longer than the video ({duration}s).'
)
number_steps = math.trunc(duration / interval_seconds)
txt_file_path = f'{output_folder}/clips.txt'
# Create the file.
open(txt_file_path, 'w').close()
print('Overview mode activated.')
print(
f'Creating a {clip_length} second clip every {interval_seconds} seconds from {video_path}...'
)
line()
try:
for step in range(1, number_steps):
clip_name = f'clip{step}.mkv'
with open(txt_file_path, 'a') as f:
f.write(f"file '{clip_name}'\n")
clip_output_path = os.path.join(output_folder, clip_name)
clip_offset = step_to_movie_timestamp(step * interval_seconds)
print(f'Creating clip {step} which starts at {clip_offset}...')
subprocess_cut_args = [
"ffmpeg", "-loglevel", "warning", "-stats", "-y", "-ss",
clip_offset, "-i", video_path, "-map", "0:V", "-t",
clip_length, "-c:v", "libx264", "-crf", "0", "-preset",
"ultrafast", clip_output_path
]
subprocess.run(subprocess_cut_args)
except Exception as error:
print('An error occurred while trying to create the clips.')
exit_program(error)
else:
return txt_file_path
def linkGal2(self, obj1, obj2, intens):
### Attention
###!!!
### pour l'instant ne gere pas plusieurs link
self.maskPoint = utils.line(self.listCenter[obj1],
self.listCenter[obj2])
maskPointIm = np.zeros_like(self.maskAll, dtype=bool)
for k in self.maskPoint:
maskPointIm[k] = True
maskPointIm = binary_dilation(maskPointIm, np.ones((3, 3)))
dist = float((self.listCenter[obj1][0] - self.listCenter[obj2][0])**2 +
(self.listCenter[obj1][1] - self.listCenter[obj2][1])**2)
listDist = [(self.listCenter[obj1][0] - k[0])**2 +
(self.listCenter[obj1][1] - k[1])**2
for k in self.maskPoint]
listLmbda = [
self.spectraSourcesLmbda[obj1] +
(self.spectraSourcesLmbda[obj2] - self.spectraSourcesLmbda[obj1]) *
k / dist for k in listDist
]
listSpectra = [
createSpectra(i, self.shape[0], width=5) for i in listLmbda
]
listPoints = np.nonzero(maskPointIm)
for i, j in zip(listPoints[0], listPoints[1]):
self.data[:, i, j] = listSpectra[np.random.randint(
len(listSpectra))] * intens
self.maskAll = self.maskAll + maskPointIm
def updata_img(self):
if self.type == "curve":
self.img = polynomial(self.raw_img, self.X, self.Y)
self.show_img()
self.show_curve()
elif self.type == "line":
self.img = line(self.raw_img, self.a, self.b)
self.show_line()
self.show_img()
elif self.type == "broken":
self.img = polynomial(self.raw_img, self.X, self.Y)
self.show_img()
self.show_curve()
elif self.type == "histeq":
self.img, self.cdf = histeq(self.raw_img)
self.show_img()
self.show_histeq()
elif self.type == "gamma":
self.img = gamma_trans(self.raw_img, self.gamma)
self.show_img()
self.show_gamma()
def test():
L, X = iline()
S, = line()
def solve():
now = '1'
compressed = reduce(mul, S, '1')
now = '1'
i = 0
need = 'i', 'j', 'k'
for L in repeat(S, X):
if need:
for c in S:
now = mul(now, c)
i += 1
if need and i > L:
print 'NO'
return
if need and now == need[0]:
need = need[1:]
now = '1'
i = 0
else:
now = mul(now, compressed)
if need or now != '1':
print 'NO'
else:
print 'YES'
return solve
def on_touch_down(self,touch):
if touch.is_double_tap:
x1, y1 = self.outlet.to_window(*self.outlet.pos)
x2, y2 = self.inlet.to_window(*self.inlet.pos)
y, x = touch.y, utils.line(x1+10,y1+7,x2+10,y2+7)(touch.x)
if(fabs(x - y) < self.line_width/2):
self.remove()
def get_users_by_account_id(ep, account_id, level):
utils.line()
print(SPACE * level +
'Get users by account_id={account_id}'.format(account_id=account_id))
print(ep.get_doc_by_query('accounts', {'_id': account_id}))
organization_id = ep.get_doc_by_query('accounts', {
'_id': account_id
}).get('organization')
group_ids = ep.get_ids_by_query('usergroups', {'account': account_id},
'_id')
position_ids = []
for group_id in group_ids:
ids = ep.get_ids_by_query('userpositions', {'group': group_id}, '_id')
position_ids.append((group_id, ids))
# print('[(group, positions)]', position_ids)
users = []
for (group_id, positions) in position_ids:
for position_id in positions:
user = ep.get_doc_by_query('users', {'position': position_id})
username = user.get('username')
user_group = user.get('group')
date_delete = user.get('dateDelete')
if date_delete is None:
date_delete = 'No'
else:
date_delete = 'Yes'
group_flag = ep.get_doc_by_query('usergroups', {
'_id': group_id
}).get('default')
user_id = user.get('_id')
isSuperAdmin = user.get('isSuperAdmin')
users.append((user_id, username, date_delete, isSuperAdmin,
position_id, group_id, group_flag, user_group))
for i, (user_id, username, date_delete, isSuperAdmin, position_id,
group_id, group_flag, user_group) in enumerate(users, 1):
print(
'{i} {user_id}, dateDelete={date_delete}, username={username}, isSuperAdmin={isSuperAdmin}, position={position_id}, group={group_id}({flag}), user.group={user_group}, organization={organization_id}'
.format(i=i,
user_id=user_id,
username=username,
date_delete=date_delete,
isSuperAdmin=isSuperAdmin,
position_id=position_id,
group_id=group_id,
flag=group_flag,
user_group=user_group,
organization_id=organization_id))
def clear_voyage_by_voyage_id(ep, voyage_id, level):
utils.line()
print(SPACE * level +
'Remove voyage_id={voyage_id}'.format(voyage_id=voyage_id))
print(SPACE * level + '{count} voyage(s) removed'.format(
count=ep.delete_many_by_query('voyagesegments', {'_id': voyage_id})))
print(SPACE * level + '{count} message(s) removed'.format(
count=ep.delete_many_by_query('messages', {
'voyage': voyage_id,
'type': 'Voyage'
})))
order_ids = ep.get_ids_by_query('orders', {'segment': voyage_id}, '_id')
level += 1
for i, order_id in enumerate(order_ids, 1):
utils.line()
print(SPACE * level + 'order {number}'.format(number=str(i)))
clear_order_by_order_id(ep, order_id, level)
def test():
_, row = line()
def solve():
print max(v - i for i, v in enumerate(
chain(*(repeat(i, n) for i, n in enumerate(imap(int, row))))))
return solve
def create_movie_overview(video_path, output_folder, interval_seconds, clip_length):
extension = Path(video_path).suffix
try:
txt_file_path = create_clips(video_path, output_folder, interval_seconds, clip_length)
output_file = concatenate_clips(txt_file_path, output_folder, extension, interval_seconds, clip_length)
result = True
except ClipError as err:
result = False
exit_program(err.args[0])
except ConcatenateError as err:
result = False
exit_program(err.args[0])
if result:
print(f'Overview Video: {clip_length}-{interval_seconds} (ClipLength-IntervalSeconds){extension}')
line()
return result, output_file
def markdown(self, run=True):
if run:
try:
return self.text.format(**self.variables)
except KeyError, name:
raise KeyError(utils.line("""
You referenced {name} in your literate document,
but no such variable was found. Did you enable
code evaluation?".format(name=name))
""", name=name))
def test():
s, = line()
yield
n = 0
for a, b in zip(s, s[1:] + '+'):
if a != b:
n += 1
print n
def find_file(self, file_name):
self.file_to_find = file_name[0]
Console.log(
f'Finding "{Console.green(self.file_to_find)}" in {Console.green(file_name[1][0])}. Please wait...'
)
for file_obj in os.walk(file_name[1][0]):
self.fnames = file_obj[2]
self.dir_path = file_obj[0]
multi_process_this(self.fnames, self.actual_finding_of_a_file)
print('\n')
print_out = f"found {len(self.found_files)} resluts of '{self.file_name}'"
time.sleep(.2)
Console.warn(line(print_out))
time.sleep(.2)
Console.log(print_out)
time.sleep(.2)
Console.warn(line(print_out, '-'))
if len(self.found_files):
self.choose_to_copy_files()
def create_movie_overview(video_path, output_folder, interval_seconds, clip_length):
extension = os.path.splitext(video_path)[-1][1:]
output_file = str()
try:
clips_file = create_clips(video_path, output_folder, interval_seconds, clip_length)
output_file = concatenate_clips(clips_file, output_folder, extension, interval_seconds, clip_length)
result = True
except ClipError as err:
result = False
print(err.args[0])
except ConcatenateError as err:
result = False
print(err.args[0])
if result:
print(f'Overview Video: {clip_length}-{interval_seconds} (ClipLength-IntervalSeconds).{extension}')
line()
return result, output_file
def concatenate_clips(txt_file_path, output_folder, extension, interval_seconds, clip_length):
if not os.path.exists(txt_file_path):
raise ConcatenateError(f'{txt_file_path} does not exist.')
overview_filename = f'{clip_length}-{interval_seconds} (ClipLength-IntervalSeconds){extension}'
concatenated_filepath = os.path.join(output_folder, overview_filename)
subprocess_concatenate_args = [
"ffmpeg", "-loglevel", "warning", "-stats", "-y",
"-f", "concat", "-safe", "0", "-i", txt_file_path, "-c", "copy", concatenated_filepath
]
line()
log.info('Concatenating the clips to create the overview video...')
result = subprocess.run(subprocess_concatenate_args)
log.info('Done!')
shutil.rmtree(os.path.join(output_folder, 'clips'))
log.info('The clips have been deleted as they are no longer needed.')
if result.returncode == 0:
return concatenated_filepath
def markdown(self, run=True):
if run:
try:
return self.text.format(**self.variables)
except KeyError, name:
raise KeyError(
utils.line("""
You referenced {name} in your literate document,
but no such variable was found. Did you enable
code evaluation?".format(name=name))
""",
name=name))
def create_clips(video_path, output_folder, interval_seconds, clip_length):
if not os.path.exists(video_path):
raise ClipError(f'The specified video file does not exist.')
if not os.path.exists(output_folder):
os.mkdir(output_folder)
duration = int(float(get_duration(video_path)))
if interval_seconds > duration:
raise ClipError(f'The interval ({interval_seconds}s) may not be longer than the video ({duration}s).')
number_steps = math.trunc(duration / interval_seconds)
output_clip_names = 'clips.txt'
output_file_path = f'{output_folder}/{output_clip_names}'
clip_file = open(output_file_path, 'w')
line()
print(f'Creating a {clip_length} second clip every {interval_seconds} seconds from {video_path}...')
line()
try:
for step in range(1, number_steps):
clip_name = f'clip{step}.mkv'
clip_file.write(f'file \'{clip_name}\'\n')
output_filename = os.path.join(output_folder, clip_name)
clip_offset = step_to_movie_timestamp(step * interval_seconds)
print(f'Creating clip {step} which starts at {clip_offset}...')
subprocess_cut_args = [
"ffmpeg", "-loglevel", "warning", "-stats", "-y",
"-ss", str(clip_offset), "-i", video_path,
"-map", "0", "-t", str(clip_length),
"-c:v", "libx264", "-crf", "0", "-preset", "ultrafast",
"-an", "-sn", output_filename
]
subprocess.run(subprocess_cut_args)
finally:
clip_file.close()
return output_file_path
def controller():
utils.showProgramName()
utils.line()
# ---------------
stop = False
while stop == False:
utils.menuOptions()
shuffleDatas()
optionUser = str(input('Selecione uma ou mais opções:'))
optionUser = optionUser.split(',')
for option in optionUser:
if option == '1':
DATAS_USERS.append(NAMES[0])
elif option == '2':
DATAS_USERS.append(EMAILS[0])
elif option == '3':
DATAS_USERS.append(CITY[0])
elif option == '4':
DATAS_USERS.append(STATE[0])
elif option == '5':
DATAS_USERS.append(PHONE[0])
elif option == 'parar':
stop = True
else:
print('Comando inválido')
if len(DATAS_USERS) != 0:
utils.line()
for datas in DATAS_USERS:
print(datas)
utils.line()
while True:
makeTxt = str(
input(
'Deseja salvar as informações em um arquivo txt? (s/n)'
)).lower()[0]
if makeTxt == 's':
with open('dados.txt', 'a') as attach:
for datas in DATAS_USERS:
attach.write(datas + '\n')
attach.write('-' * 10 + '\n')
break
elif makeTxt == 'n':
break
DATAS_USERS.clear()
utils.line()
def clear_order_by_order_number(ep, order_number):
order_id = ep.get_doc_id_by_query('orders', {'order_number': order_number})
if order_id is not None:
utils.line()
print('Remove order_number={order_number}'.format(
order_number=order_number))
print('{count} order(s) removed'.format(
count=ep.delete_many_by_query('orders', {'_id': order_id})))
print('{count} orderentry(s) removed'.format(
count=ep.delete_many_by_query('orderentries',
{'order': order_id})))
print('{count} productconfig(s) removed'.format(
count=ep.delete_many_by_query('productconfigs',
{'order': order_id})))
print('{count} event(s) removed'.format(
count=ep.delete_many_by_query('events', {'order': order_id})))
print('{count} feedback(s) removed'.format(
count=ep.delete_many_by_query('feedbacks', {'order': order_id})))
print('{count} accessclaim(s) removed'.format(
count=ep.delete_many_by_query('accessclaims',
{'resourceId': str(order_id)})))
print('{count} message(s) removed'.format(
count=ep.delete_many_by_query('messages', {'order': order_id})))
def clear_user_by_username(ep, username, level):
utils.line()
print(SPACE * level + 'Remove user={user}'.format(user=username))
print(ep.get_doc_by_query('users', {'username': username}))
position_id = ep.get_doc_by_query('users', {
'username': username
}).get('position')
group_id = ep.get_doc_by_query('userpositions', {
'_id': position_id
}).get('group')
account_id = ep.get_doc_by_query('usergroups', {
'_id': group_id
}).get('account')
organization_id = ep.get_doc_by_query('accounts', {
'_id': account_id
}).get('organization')
print(position_id, group_id, account_id, organization_id)
ep.delete_one_by_query('users', {'username': username})
ep.delete_one_by_query('userpositions', {'_id': position_id})
ep.delete_one_by_query('usergroups', {'_id': group_id})
ep.delete_one_by_query('accounts', {'_id': account_id})
ep.delete_one_by_query('organizations', {'_id': organization_id})
ep.delete_one_by_query('accessclaims', {'gid': str(group_id)})
def clear_order_by_order_id(ep, order_id, level):
utils.line()
print(SPACE * level +
'Remove order_id={order_id}'.format(order_id=order_id))
print(SPACE * level + '{count} order removed'.format(
count=ep.delete_many_by_query('orders', {'_id': order_id})))
print(SPACE * level + '{count} orderentry(s) removed'.format(
count=ep.delete_many_by_query('orderentries', {'order': order_id})))
print(SPACE * level + '{count} productconfig(s) removed'.format(
count=ep.delete_many_by_query('productconfigs', {'order': order_id})))
print(SPACE * level + '{count} event(s) removed'.format(
count=ep.delete_many_by_query('events', {'order': order_id})))
print(SPACE * level + '{count} feedback(s) removed'.format(
count=ep.delete_many_by_query('feedbacks', {'order': order_id})))
print(SPACE * level + '{count} accessclaim(s) removed'.format(
count=ep.delete_many_by_query('accessclaims',
{'resourceId': str(order_id)})))
print(SPACE * level + '{count} message(s) removed'.format(
count=ep.delete_many_by_query('messages', {
'order': order_id,
'type': 'Order'
})))
print(SPACE * level + '{count} inquiryorder(s) removed'.format(
count=ep.delete_many_by_query('inquiryorders', {'order': order_id})))
def test():
N, = iline()
sentences = [set(line()) for i in xrange(N)]
yield
answer = 1e100
for bits in xrange(1 << (N - 2)):
words = [set(), set()]
bits = 4 * bits + 2
for i, s in enumerate(sentences):
language = 1 if (bits & (1 << i)) else 0
words[language] |= s
answer = min(answer, len(words[0] & words[1]))
print answer
def test():
word, = line()
yield
nums = {}
for c in word:
nums[c] = nums.get(c, 0) + 1
ans = []
for v, c in answer:
n = nums.get(c, 0)
if n:
ans += [v] * n
for c in D[v]:
nums[c] -= n
ans = map(str, sorted(ans))
print ''.join(ans)
def show_curve(self):
if self.type == "curve":
x = np.arange(0, 255, 1)
y = polynomial(x, self.X, self.Y)
self.axes.plot(x, y)
self.axes.plot(self.X, self.Y, "ob")
elif self.type == "line":
x = np.arange(0, 255, 1)
y = line(x, self.X, self.Y)
self.axes.plot(x, y)
elif self.type == "broken":
x = np.arange(0, 255, 1)
y = broken(x, self.X, self.Y)
self.axes.plot(x, y)
self.axes.plot(self.X, self.Y, "ob")
elif self.type == "gamma":
x = np.arange(0, 255, 1)
y = gamma_trans(x, self.X)
self.axes.plot(x, y)
def add_intersection():
""" Add intersection with the Box border for begining or end."""
# Generate line equations Ax + By = C from two points and edges
line = utils.line(prev_pt, pt)
top = {'a': 0, 'b': 1, 'c': self.y1}
bottom = {'a': 0, 'b': 1, 'c': self.y}
left = {'a': 1, 'b': 0, 'c': self.x}
right = {'a': 1, 'b': 0, 'c': self.x1}
# Check intersection of segment with edges
pt_begin = None
for edge in [top, bottom, left, right]:
intersection = utils.line_intersection(line, edge)
if intersection is None:
continue
if self._is_in_range(intersection, pt, prev_pt):
pt_begin = intersection
if pt_begin is not None:
result.append(((pt_begin[0] - self.x) * px_size[0],
(pt_begin[1] - self.y) * px_size[1]))
# 1. Generate ground-truth trajectories
x1 = trajectory_simulation(0, 0, 300, 0, 0, 0, 200)
x2 = trajectory_simulation(x1[len(x1)-1][0], x1[len(x1)-1][2], 300, 0, 2, 2, 100)
x3 = trajectory_simulation(x2[len(x2)-1][0], x2[len(x2)-1][2], 0, 300, 0, 0, 200)
x1_n = trajectory_simulation_noise(0, 0, 300, 0, 0, 0, 200)
x2_n = trajectory_simulation_noise(x1_n[len(x1_n)-1][0], x1_n[len(x1_n)-1][2], 300, 0, 2, 2, 100)
x3_n = trajectory_simulation_noise(x2_n[len(x2_n)-1][0], x2_n[len(x2_n)-1][2], 0, 300, 0, 0, 200)
x_f = x1+x2+x3
x_f = np.asmatrix(x_f)
x_f_n = x1_n+x2_n+x3_n
x_f_n = np.asmatrix(x_f_n)
# 2. Plot the trajectories
# with noise
utils.line(x_f[:,0], x_f[:, 2], "x", "y", "Plot of ground-truth trajectory")
# without noise
fig = plt.figure()
plt.title("Plot of ground-truth and measurement")
plt.ylabel("y")
plt.xlabel("x")
plt.plot(x_f[:, 0], x_f[:, 2])
plt.plot(x_f_n[:, 0], x_f_n[:, 2])
fig.savefig(os.path.join(baseDir, 'Figures/Plot of ground-truth and measurement.png'))
utils.line(np.arange(500), x_f_n[:,1], "time", "velocity", "velocity of x")
utils.line(np.arange(500), x_f_n[:,3], "time", "velocity", "velocity of y")
print 'The accuracy of the model on test data is ', accuracy
#==============================================
# 2l. plot cmc
#
cmc = []
for k in range(1,10):
accuracy = 0
for d in test_data:
predicted_y = naive_bayes.predict(train_lh, train_prior, utils.bags(d['summary']))
sorted_y = sorted(predicted_y.items(), key=operator.itemgetter(1), reverse=True)
for i in range(k):
if sorted_y[i][0] == d['year']:
accuracy += 1
break
cmc.append(accuracy/float(len(test_data)))
print 'The CMC is ', cmc
utils.line([1,2,3,4,5,6,7,8,9], cmc, 'K', 'Accuracy', 'Cumulative Match Curve')
#==============================================
# 2m. Plot confusion matrix
#
cmatrix = np.zeros(shape=(9,9))
for d in test_data:
predicted_y = naive_bayes.predict(train_lh, train_prior, utils.bags(d['summary']))
sorted_y = sorted(predicted_y.items(), key=operator.itemgetter(1), reverse=True)
predicted_year = sorted_y[0][0]
cmatrix[(d['year']-1930)/10][(predicted_year-1930)/10] += 1
utils.spectrom(cmatrix, 'Decade', 'Decade', 'Confusion Matrix')
plt.title("2A. Plot of estimated trajectory, raw measurement, and true trajectory")
plt.ylabel("y")
plt.xlabel("x")
plt.plot(estimate[0, :].T, estimate[2, :].T)
plt.plot(x_f[:, 0], x_f[:, 2])
plt.plot(x_f_n[:, 0], x_f_n[:, 2])
fig.savefig(os.path.join(baseDir, 'Figures/2A. Estimated measurement and true trajectory.png'))
# error rate
for t in range(500):
if t == 0:
error = x_f_n[t] - estimate.T[t]
else:
error = np.concatenate((error, x_f_n[t] - estimate.T[t]), axis=0)
utils.line(np.matrix(np.arange(500)).T, error[:, 0], "time", "error X", "2A. Error of Trajectory X")
utils.line(np.matrix(np.arange(500)).T, error[:, 1], "time", "error X velocity", "2A. Error of X Velocity")
utils.line(np.matrix(np.arange(500)).T, error[:, 2], "time", "error Y", "2A. Error of Trajectory Y")
utils.line(np.matrix(np.arange(500)).T, error[:, 3], "time", "error Y velocity", "2A. Error of Y Velocity")
fig = plt.figure()
plt.title("2A. Plot of estimated velocity x, raw measurement, and true velocity")
plt.ylabel("x velocity")
plt.xlabel("time")
plt.scatter(np.matrix(np.arange(500)).T, estimate[1, :].T, c='red')
plt.scatter(np.matrix(np.arange(500)).T, x_f[:, 1], c='blue')
plt.scatter(np.matrix(np.arange(500)).T, x_f_n[:, 1], c='green')
fig.savefig(os.path.join(baseDir, 'Figures/2A. Estimated velocity x and true velocity.png'))
fig = plt.figure()
plt.title("2A. Plot of estimated velocity y, raw measurement, and true velocity")
