def test_load_old_checkpoint(cls):
dataset = AmazonReviewSentimentCrossLingualDataset()
sha1sum_id = "4ba096cdf6bd76c06386f2c27140db055e59c91b"
checkpoint_name = "mdeberta-v3-base-checkpoint"
save_path = os.path.join(get_home_dir(), "checkpoints")
file_path = os.path.join(save_path, f"{checkpoint_name}.zip")
checkpoint_path = os.path.join(get_home_dir(), "checkpoints",
checkpoint_name)
if os.path.exists(save_path):
shutil.rmtree(save_path)
download(
url=
f"s3://automl-mm-bench/unit-tests-0.4/checkpoints/{checkpoint_name}.zip",
path=file_path,
sha1_hash=sha1sum_id,
)
protected_zip_extraction(
file_path,
sha1_hash=sha1sum_id,
folder=save_path,
)
predictor = cls.load(checkpoint_path)
verify_predictor_save_load(predictor, dataset.test_df, cls=cls)
# continuous training
predictor.fit(
dataset.train_df,
presets="multilingual",
time_limit=10,
hyperparameters={"optimization.top_k_average_method": "uniform_soup"},
)
verify_predictor_save_load(predictor, dataset.test_df, cls=cls)
def _get_config_files(self):
if string.find(sys.platform, 'sunos') > -1:
return [
"/opt/redhat/rhn/solaris/etc/sysconfig/rhn/%s.conf" % self.section,
os.path.join(utils.get_home_dir(), self._local_config_file_name),
self._local_config_file_name,
]
else:
return [
"/etc/sysconfig/rhn/%s.conf" % self.section,
os.path.join(utils.get_home_dir(), self._local_config_file_name),
self._local_config_file_name,
]
def _get_config_files(self):
if string.find(sys.platform, 'sunos') > -1:
return [
"/opt/redhat/rhn/solaris/etc/sysconfig/rhn/%s.conf" %
self.section,
os.path.join(utils.get_home_dir(),
self._local_config_file_name),
self._local_config_file_name,
]
else:
return [
"/etc/sysconfig/rhn/%s.conf" % self.section,
os.path.join(utils.get_home_dir(),
self._local_config_file_name),
self._local_config_file_name,
]
def slot_playlist(self):
dialog = QtGui.QFileDialog()
dialog.setFileMode(QtGui.QFileDialog.ExistingFile)
if config.post_download_playlist_path:
dialog.setDirectory(os.path.dirname(config.post_download_playlist_path))
else:
# The default playlist directory
dialog.setDirectory(r'%s\My Documents\My Music\My Playlists' % utils.get_home_dir())
f = unicode(dialog.getOpenFileName(caption=tr("Open Playlist"), filter=tr("Supported Playlist Files") + " (*.m3u *.wpl)"))
f = f.replace('/','\\')
if f:
try:
log.debug("Adding %s to the %s playlist..." % (self.audio_path, f))
utils.add_item_to_playlist(f, self.audio_path)
self.statusBar_append = tr("; Saved to playlist")
except (IOError, RuntimeError), e:
log.error(str(e))
QtGui.QMessageBox.critical(self, tr("Error"), str(e), QtGui.QMessageBox.Ok)
return
config.post_download_playlist_path = f
if self.saveSelection_CheckBox.isChecked():
config.post_download_action = 'addPlaylist'
self.close()
def slot_post_download_action_changed(self, i):
val = config.post_download_action_dict.keys()[i]
if val == 'addItunes' and not config.is_itunes_installed:
QtGui.QMessageBox.critical(self, tr("Error"), tr("iTunes is not installed on this system."), QtGui.QMessageBox.Ok)
self.post_download_action.setCurrentIndex(config.post_download_action_dict.keys().index(config.post_download_action))
elif val == 'addPlaylist':
dialog = QtGui.QFileDialog()
dialog.setFileMode(QtGui.QFileDialog.ExistingFile)
if config.post_download_playlist_path:
dialog.setDirectory(os.path.dirname(config.post_download_playlist_path))
else:
dialog.setDirectory(r'%s\My Documents\My Music\My Playlists' % utils.get_home_dir()) # The default playlist directory
f = unicode(dialog.getOpenFileName(caption=tr("Open Playlist"), filter=tr("Supported Playlist Files") + " (*.m3u *.wpl)"))
f = f.replace('/','\\')
if f:
config.post_download_playlist_path = f
elif not config.post_download_playlist_path:
# if new playlist is not choosed, and there is NO playlist path in the config file, we should revert the post_download_action to the last one.
self.post_download_action.setCurrentIndex(config.post_download_action_dict.keys().index(config.post_download_action))
return
config.post_download_action = val
self.slot_changed_checkbox()
def __init__(self, host=None):
self.host = host
ssl_dir = os.path.join(get_home_dir(), "ssl")
self.key_path = os.path.join(ssl_dir, "dodo.key")
self.cert_path = os.path.join(ssl_dir, "dodo.crt")
if not os.path.exists(ssl_dir):
os.makedirs(ssl_dir)
def __init__(self,host=None):
self.host = host
ssl_dir = os.path.join(get_home_dir(), "ssl")
self.key_path = os.path.join(ssl_dir, "dodo.key")
self.cert_path = os.path.join(ssl_dir, "dodo.crt")
if not os.path.exists(ssl_dir):
os.makedirs(ssl_dir)
def test_move_most_recent_files(self):
self.data1 = pd.DataFrame({'data': [1, 2, 3, 4, 5]})
self.data1.to_csv(get_home_dir() + '/Downloads/test1.csv')
download_start = datetime.now().timestamp()
self.data2 = pd.DataFrame({'data': [1, 2, 3, 4, 5]})
self.data2.to_csv(
get_home_dir() +
'/Downloads/Britain Coronavirus Disease Prevention Map Mar 06 2020 Id Id Colocation Map_2020-03-31.csv'
)
self.assertTrue(os.path.exists(get_home_dir() +
'/Downloads/test1.csv'))
self.assertTrue(
os.path.exists(
get_home_dir() +
'/Downloads/Britain Coronavirus Disease Prevention Map Mar 06 2020 Id Id Colocation Map_2020-03-31.csv'
))
move_most_recent_files('./tmp1', ['url'], download_start)
self.assertTrue(os.path.exists(get_home_dir() +
'/Downloads/test1.csv'))
self.assertTrue(os.path.exists('./tmp1/Britain_2020_03_31.csv'))
os.remove(get_home_dir() + '/Downloads/test1.csv')
def __init__(self, optionparser, store_true_list):
sysdir = '/etc/sysconfig/rhn'
homedir = utils.get_home_dir()
default = 'rhnpushrc'
regular = '.rhnpushrc'
deffile = os.path.join(sysdir, default)
regfile = os.path.join(homedir, regular)
cwdfile = os.path.join(os.getcwd(), regular)
self.cfgFileList = [deffile, regfile, cwdfile]
self.defaultconfig = rhnpush_config.rhnpushConfigParser(ensure_consistency=True)
# Get a reference to the object containing command-line options
self.cmdconfig = optionparser
self.store_true_list = store_true_list
def __init__(self, optionparser, store_true_list):
sysdir = '/etc/sysconfig/rhn'
homedir = utils.get_home_dir()
default = 'rhnpushrc'
regular = '.rhnpushrc'
deffile = os.path.join(sysdir, default)
regfile = os.path.join(homedir, regular)
cwdfile = os.path.join(os.getcwd(), regular)
self.cfgFileList = [deffile, regfile, cwdfile]
self.defaultconfig = rhnpush_config.rhnpushConfigParser(ensure_consistency=True)
# Get a reference to the object containing command-line options
self.cmdconfig = optionparser
self.store_true_list = store_true_list
def test_hpo(searcher, scheduler):
dataset = PetFinderDataset()
hyperparameters = {
"optimization.learning_rate": tune.uniform(0.0001, 0.01),
"optimization.max_epochs": 1,
"model.names": ["numerical_mlp", "categorical_mlp", "fusion_mlp"],
"data.categorical.convert_to_text": False,
"data.numerical.convert_to_text": False,
"env.num_workers": 0,
"env.num_workers_evaluation": 0,
}
hyperparameter_tune_kwargs = {
"searcher": searcher,
"scheduler": scheduler,
"num_trials": 2,
}
predictor = MultiModalPredictor(
label=dataset.label_columns[0],
problem_type=dataset.problem_type,
eval_metric=dataset.metric,
)
save_path = os.path.join(get_home_dir(), "hpo", f"_{searcher}",
f"_{scheduler}")
if os.path.exists(save_path):
shutil.rmtree(save_path)
predictor = predictor.fit(
train_data=dataset.train_df,
hyperparameters=hyperparameters,
time_limit=60,
save_path=save_path,
hyperparameter_tune_kwargs=hyperparameter_tune_kwargs,
)
score = predictor.evaluate(dataset.test_df)
verify_predictor_save_load(predictor, dataset.test_df)
# test for continuous training
predictor = predictor.fit(
train_data=dataset.train_df,
hyperparameters=hyperparameters,
time_limit=60,
hyperparameter_tune_kwargs=hyperparameter_tune_kwargs,
)
def slot_post_download_action_changed(self, i):
val = config.post_download_action_dict.keys()[i]
if val == 'addItunes' and not config.is_itunes_installed:
QtGui.QMessageBox.critical(
self, tr("Error"),
tr("iTunes is not installed on this system."),
QtGui.QMessageBox.Ok)
self.post_download_action.setCurrentIndex(
config.post_download_action_dict.keys().index(
config.post_download_action))
elif val == 'addPlaylist':
dialog = QtGui.QFileDialog()
dialog.setFileMode(QtGui.QFileDialog.ExistingFile)
if config.post_download_playlist_path:
dialog.setDirectory(
os.path.dirname(config.post_download_playlist_path))
else:
dialog.setDirectory(
r'%s\My Documents\My Music\My Playlists' %
utils.get_home_dir()) # The default playlist directory
f = unicode(
dialog.getOpenFileName(caption=tr("Open Playlist"),
filter=tr("Supported Playlist Files") +
" (*.m3u *.wpl)"))
f = f.replace('/', '\\')
if f:
config.post_download_playlist_path = f
elif not config.post_download_playlist_path:
# if new playlist is not choosed, and there is NO playlist path in the config file, we should revert the post_download_action to the last one.
self.post_download_action.setCurrentIndex(
config.post_download_action_dict.keys().index(
config.post_download_action))
return
config.post_download_action = val
self.slot_changed_checkbox()
def test_standalone(): # test standalong feature in MultiModalPredictor.save()
from unittest import mock
import torch
requests_gag = mock.patch(
'requests.Session.request',
mock.Mock(side_effect=RuntimeError(
'Please use the `responses` library to mock HTTP in your tests.')))
dataset = PetFinderDataset()
config = {
MODEL: f"fusion_mlp_image_text_tabular",
DATA: "default",
OPTIMIZATION: "adamw",
ENVIRONMENT: "default",
}
hyperparameters = {
"optimization.max_epochs":
1,
"model.names": [
"numerical_mlp", "categorical_mlp", "timm_image", "hf_text",
"clip", "fusion_mlp"
],
"model.hf_text.checkpoint_name":
"prajjwal1/bert-tiny",
"model.timm_image.checkpoint_name":
"swin_tiny_patch4_window7_224",
"env.num_workers":
0,
"env.num_workers_evaluation":
0,
}
predictor = MultiModalPredictor(
label=dataset.label_columns[0],
problem_type=dataset.problem_type,
eval_metric=dataset.metric,
)
save_path = os.path.join(get_home_dir(), "standalone", "false")
if os.path.exists(save_path):
shutil.rmtree(save_path)
predictor.fit(
train_data=dataset.train_df,
config=config,
hyperparameters=hyperparameters,
time_limit=30,
save_path=save_path,
)
save_path_standalone = os.path.join(get_home_dir(), "standalone", "true")
predictor.save(
path=save_path_standalone,
standalone=True,
)
del predictor
torch.cuda.empty_cache()
loaded_online_predictor = MultiModalPredictor.load(path=save_path)
online_predictions = loaded_online_predictor.predict(dataset.test_df,
as_pandas=False)
del loaded_online_predictor
# Check if the predictor can be loaded from an offline enivronment.
with requests_gag:
# No internet connection here. If any command require internet connection, a RuntimeError will be raised.
with tempfile.TemporaryDirectory() as tmpdirname:
torch.hub.set_dir(tmpdirname) # block reading files in `.cache`
loaded_offline_predictor = MultiModalPredictor.load(
path=save_path_standalone)
offline_predictions = loaded_offline_predictor.predict(dataset.test_df,
as_pandas=False)
del loaded_offline_predictor
# check if save with standalone=True coincide with standalone=False
npt.assert_equal(online_predictions, offline_predictions)
def sanity_check():
"Sanity Check for script."
config.count_application_runs += 1
_warnings = []
### LOCAL CHECKS ###
# Windows version check
winver = sys.getwindowsversion()
log.debug('Running iQuality v%s (r%d) on Windows %d.%d.%d %s' % (__version__, __rev__, winver.major, winver.minor, winver.build, winver.service_pack))
# Python version check
if sys.version_info < (2, 6) or sys.version_info >= (3, 0):
msg = "must use python 2.7"
log.critical(msg)
raise Exception(msg)
log.debug('CPython version is %d.%d.%d' % (sys.version_info.major, sys.version_info.minor, sys.version_info.micro))
log.debug('PyQt version is %s, Qt version is %s' % (QtCore.PYQT_VERSION_STR, QtCore.QT_VERSION_STR))
# youtube-dl check
try:
import youtube_dl
log.debug("youtube-dl version is %s" % youtube_dl.__version__)
except ImportError:
log.warning("Could not load the youtube-dl module")
# Phonon version check
try:
from PyQt4.phonon import Phonon
if Phonon.BackendCapabilities.isMimeTypeAvailable('video/x-flv'):
log.debug('Phonon version is %s. video/x-flv is supported.' % Phonon.phononVersion())
else:
log.warning('Phonon version is %s. video/x-flv is not supported.' % Phonon.phononVersion())
except ImportError:
log.warning("Could not load the phonon module")
# Free space check
freespace = utils.get_free_space(config.temp_dir)
if freespace < 200*1024**2: # 200 MB
drive = os.path.splitdrive(config.temp_dir)[0]
log.warning("There are less than 200MB available in drive %s (%.2fMB left)." % (drive, freespace/1024.0**2))
_warnings.append(NoSpaceWarning(drive, freespace))
# iTunes' availablity check
itunesPath = r'%s\My Documents\My Music\iTunes\iTunes Media\Automatically Add to iTunes' % utils.get_home_dir()
if not os.path.exists(itunesPath):
config.is_itunes_installed = False
if config.post_download_action == 'addItunes':
config.post_download_action = 'ask'
log.warning("iTunes Media not found. setting is_itunes_installed to False")
# Context Menu check
try: # IMPROVE: REMOVE THE TRY-EXCEPT BLOCK
if config.id3editor_in_context_menu and not utils.check_context_menu_status():
log.debug("Registering Context Menu Object...")
try:
utils.register_with_context_menu()
except WindowsError, e:
if e.winerror == 5: # Access is denied
log.debug("Access is denied. Setting id3editor_in_context_menu to False.")
config.id3editor_in_context_menu = False
else:
raise
if not config.id3editor_in_context_menu and utils.check_context_menu_status():
log.debug("Unregistering Context Menu Object...")
try:
utils.unregister_with_context_menu()
except WindowsError, e:
if e.winerror == 5: # Access is denied
log.debug("Access is denied. Setting id3editor_in_context_menu to True.")
config.id3editor_in_context_menu = True
else:
raise
def get_config_dir():
if not os.path.exists('%s/.config/' % get_home_dir() ):
os.makedirs( '%s/.config/' % get_home_dir() )
if not os.path.exists('%s/.config/mypyapp' % get_home_dir() ):
os.makedirs( '%s/.config/mypyapp' % get_home_dir() )
return '%s/.config/mypyapp' % get_home_dir()
def test_get_home_dir(self):
self.assertTrue('~' not in get_home_dir())
def merge_data(self):
add_data_filename = unicode(QFileDialog.getOpenFileName(self, "Open File", get_home_dir(), get_format_filter()))
self._merge_data(decode_file(add_data_filename))
def excel_import(self):
excel_filename = unicode(QFileDialog.getOpenFileName(self, "Open File", get_home_dir(), "Excel files (*.xls)"))
if excel_filename:
data_to_merge = import_from_excel(self.scheme, excel_filename)
self.parent()._merge_data(data_to_merge)
from concurrent.futures import ProcessPoolExecutor
# custom
from utils import get_home_dir
import dl_quandl_EOD as dq
stocks = dq.load_stocks()
DEFAULT_STORAGE = '/home/nate/eod_data/'
# get todays date for checking if files up-to-date
MTN = timezone('America/Denver')
TODAY = datetime.datetime.now(MTN)
WEEKDAY = TODAY.weekday()
HOUR = TODAY.hour
HOME_DIR = get_home_dir()
Q_KEY = os.environ.get('quandl_api')
quandl.ApiConfig.api_key = Q_KEY
spy_vix = {}
closes = {}
dates = {}
for i in range(1, 10):
print(i)
spy_vix[i] = quandl.get("CHRIS/CBOE_VX" + str(i))
# spy_vix[i].to_csv()
closes[i] = spy_vix[i]['Close']
dates[i] = spy_vix[i].index
import argparse
import os
import re
import shutil
import utils
utils.append_home_to_path(__file__)
from tomahawk.rsync import RsyncMain
from tomahawk.constants import TimeoutError
from tomahawk.expect import CommandWithExpect
TESTS_DIR = os.path.dirname(os.path.abspath(__file__))
TMP_DIR = os.path.join(utils.get_home_dir(__file__), 'tmp')
if not os.path.exists(TMP_DIR):
os.mkdir(TMP_DIR)
env = os.environ
if env.get('TOMAHAWK_ENV') != None:
del env['TOMAHAWK_ENV']
hello_file = os.path.join(TMP_DIR, 'hello')
hello_file_copied = os.path.join(TMP_DIR, 'hello.copied')
if os.path.exists(hello_file_copied):
os.remove(hello_file_copied)
handle = open(hello_file, 'w')
handle.write('hello world')
handle.close()
def test_00_run(monkeypatch):
EXPECTED = {
__author__ = 'CSPF'
from flask import Flask
from flask_sqlalchemy import SQLAlchemy
from os import path, makedirs, stat
from utils import get_home_dir
db_dir = path.join(get_home_dir(), "db")
if not path.exists(db_dir):
#make sure to create home directory and database directory
makedirs(db_dir)
db_path = path.join(db_dir, "Database.sqlite")
app = Flask("DodoBank")
app.config["SECRET_KEY"] = "1234"
app.config['SQLALCHEMY_DATABASE_URI'] = 'sqlite:///' + db_path
db = SQLAlchemy(app)
def __init__(self):
self.homedir = utils.get_home_dir()
def sanity_check():
"Sanity Check for script."
config.count_application_runs += 1
_warnings = []
### LOCAL CHECKS ###
# Windows version check
winver = sys.getwindowsversion()
log.debug('Running iQuality v%s (r%d) on Windows %d.%d.%d %s' %
(__version__, __rev__, winver.major, winver.minor, winver.build,
winver.service_pack))
# Python version check
if sys.version_info < (2, 6) or sys.version_info >= (3, 0):
msg = "must use python 2.7"
log.critical(msg)
raise Exception(msg)
log.debug('CPython version is %d.%d.%d' %
(sys.version_info.major, sys.version_info.minor,
sys.version_info.micro))
log.debug('PyQt version is %s, Qt version is %s' %
(QtCore.PYQT_VERSION_STR, QtCore.QT_VERSION_STR))
# youtube-dl check
try:
import youtube_dl
log.debug("youtube-dl version is %s" % youtube_dl.__version__)
except ImportError:
log.warning("Could not load the youtube-dl module")
# Phonon version check
try:
from PyQt4.phonon import Phonon
if Phonon.BackendCapabilities.isMimeTypeAvailable('video/x-flv'):
log.debug('Phonon version is %s. video/x-flv is supported.' %
Phonon.phononVersion())
else:
log.warning('Phonon version is %s. video/x-flv is not supported.' %
Phonon.phononVersion())
except ImportError:
log.warning("Could not load the phonon module")
# Free space check
freespace = utils.get_free_space(config.temp_dir)
if freespace < 200 * 1024**2: # 200 MB
drive = os.path.splitdrive(config.temp_dir)[0]
log.warning(
"There are less than 200MB available in drive %s (%.2fMB left)." %
(drive, freespace / 1024.0**2))
_warnings.append(NoSpaceWarning(drive, freespace))
# iTunes' availablity check
itunesPath = r'%s\My Documents\My Music\iTunes\iTunes Media\Automatically Add to iTunes' % utils.get_home_dir(
)
if not os.path.exists(itunesPath):
config.is_itunes_installed = False
if config.post_download_action == 'addItunes':
config.post_download_action = 'ask'
log.warning(
"iTunes Media not found. setting is_itunes_installed to False")
# Context Menu check
try: # IMPROVE: REMOVE THE TRY-EXCEPT BLOCK
if config.id3editor_in_context_menu and not utils.check_context_menu_status(
):
log.debug("Registering Context Menu Object...")
try:
utils.register_with_context_menu()
except WindowsError, e:
if e.winerror == 5: # Access is denied
log.debug(
"Access is denied. Setting id3editor_in_context_menu to False."
)
config.id3editor_in_context_menu = False
else:
raise
if not config.id3editor_in_context_menu and utils.check_context_menu_status(
):
log.debug("Unregistering Context Menu Object...")
try:
utils.unregister_with_context_menu()
except WindowsError, e:
if e.winerror == 5: # Access is denied
log.debug(
"Access is denied. Setting id3editor_in_context_menu to True."
)
config.id3editor_in_context_menu = True
else:
raise
def __init__(self):
self.homedir = utils.get_home_dir()
self.desktop = os.getenv('XDG_CURRENT_DESKTOP')
if self.desktop is None:
self.desktop = os.getenv('XDG_SESSION_DESKTOP')
def test_predictor(
dataset_name,
model_names,
text_backbone,
image_backbone,
top_k_average_method,
efficient_finetune,
loss_function,
):
dataset = ALL_DATASETS[dataset_name]()
metric_name = dataset.metric
predictor = MultiModalPredictor(
label=dataset.label_columns[0],
problem_type=dataset.problem_type,
eval_metric=metric_name,
)
config = {
MODEL: f"fusion_mlp_image_text_tabular",
DATA: "default",
OPTIMIZATION: "adamw",
ENVIRONMENT: "default",
}
hyperparameters = {
"optimization.max_epochs": 1,
"model.names": model_names,
"env.num_workers": 0,
"env.num_workers_evaluation": 0,
"optimization.top_k_average_method": top_k_average_method,
"optimization.efficient_finetune": efficient_finetune,
"optimization.loss_function": loss_function,
}
if text_backbone is not None:
hyperparameters.update({
"model.hf_text.checkpoint_name": text_backbone,
})
if image_backbone is not None:
hyperparameters.update({
"model.timm_image.checkpoint_name":
image_backbone,
})
save_path = os.path.join(get_home_dir(), "outputs", dataset_name)
if text_backbone is not None:
save_path = os.path.join(save_path, text_backbone)
if image_backbone is not None:
save_path = os.path.join(save_path, image_backbone)
if os.path.exists(save_path):
shutil.rmtree(save_path)
predictor.fit(
train_data=dataset.train_df,
config=config,
hyperparameters=hyperparameters,
time_limit=30,
save_path=save_path,
)
score = predictor.evaluate(dataset.test_df)
verify_predictor_save_load(predictor, dataset.test_df)
# Test for continuous fit
predictor.fit(
train_data=dataset.train_df,
config=config,
hyperparameters=hyperparameters,
time_limit=30,
)
verify_predictor_save_load(predictor, dataset.test_df)
# Saving to folder, loading the saved model and call fit again (continuous fit)
with tempfile.TemporaryDirectory() as root:
predictor.save(root)
predictor = MultiModalPredictor.load(root)
predictor.fit(
train_data=dataset.train_df,
config=config,
hyperparameters=hyperparameters,
time_limit=30,
)
def __init__(self):
self.homedir = utils.get_home_dir()
self.desktop = os.getenv('XDG_CURRENT_DESKTOP')
if self.desktop is None:
self.desktop = os.getenv('XDG_SESSION_DESKTOP')
def load_data(self):
data_filename = unicode(QFileDialog.getOpenFileName(self, "Open File", get_home_dir(), get_format_filter()))
if data_filename:
self.__load_data(data_filename)
import argparse
import os
import re
import shutil
import utils
utils.append_home_to_path(__file__)
from tomahawk.rsync import RsyncMain
from tomahawk.constants import TimeoutError
from tomahawk.expect import CommandWithExpect
TESTS_DIR = os.path.dirname(os.path.abspath(__file__))
TMP_DIR = os.path.join(utils.get_home_dir(__file__), 'tmp')
if not os.path.exists(TMP_DIR):
os.mkdir(TMP_DIR)
env = os.environ
if env.get('TOMAHAWK_ENV') != None:
del env['TOMAHAWK_ENV']
hello_file = os.path.join(TMP_DIR, 'hello')
hello_file_copied = os.path.join(TMP_DIR, 'hello.copied')
if os.path.exists(hello_file_copied):
os.remove(hello_file_copied)
handle = open(hello_file, 'w')
handle.write('hello world')
handle.close()
def test_00_run(monkeypatch):
def main(argv):
"""Main function"""
## Get user information ##
global op_system
op_system = os.name
if op_system != 'nt':
op_system = subprocess.Popen(['uname', '-s'],
stdout=subprocess.PIPE).communicate()[0]
op_system = op_system.rstrip('\n')
# change Cygwin to say 'CYGWIN' always
under_score_idx = op_system.find('_')
if under_score_idx != -1:
op_system = op_system[:under_score_idx]
if op_system != 'Linux' and op_system != 'Darwin':
print('Error: %s is not a supported operating system at this time.'
% op_system)
exit(5)
global data_dir
data_dir = os.path.join(utils.get_home_dir(), '.pimessage/')
opt1 = grab_opt(argv, 1)
if opt1 == 'uninstall':
exit(uninstall())
script_name = argv[0]
if not os.path.isdir(data_dir):
install(script_name, data_dir)
dir_files = subprocess.Popen(['ls', '-A', data_dir],
stdout=subprocess.PIPE).communicate()[0]
# must be in correct ls -A order
correct_dir_files = """contacts
conversations
daemonError.log
editor
"""
alt_dir_files = """contacts
conversations
editor
"""
if dir_files != correct_dir_files and dir_files != alt_dir_files:
install(script_name, data_dir)
# get user's chosen editor
edit_cmd = open(data_dir+'editor', 'r').read().rstrip('\n')
if not subprocess.Popen(['which', edit_cmd],
stdout=subprocess.PIPE).communicate()[0]:
print('Error: %s is not a valid editor. Please adjust your editor '
'value' % edit_cmd)
exit(2)
global host_ip
host_ip = ip.get_host_ip()
#if host_ip == IP_FAILURE:
# print 'Error: your IP address could not be correctly retrieved.'
# exit(2)
## Option parsing ##
# figure out which option was called
parse_opts(argv, edit_cmd)
return 0
# core
import time
import sys
from concurrent.futures import ProcessPoolExecutor
# custom
sys.path.append('../code')
import dl_quandl_EOD as dq
import data_processing as dp
import short_squeeze_eda as sse
from utils import get_home_dir
HOME_DIR = get_home_dir(repo_name='scrape_stocks')
# installed
import numpy as np
import pandas as pd
from tqdm import tqdm
# plotting
import matplotlib.pyplot as plt
%matplotlib inline
from plotly.offline import init_notebook_mode, iplot
from plotly.graph_objs import Scatter, Scattergl, Figure, Layout
def calc_vol(st, mean_vol):
"""
takes dataframe of stock data (st) and calculates tp, 50d-mva, and volatility
also takes dictionary (mean_vol) as arg
"""
st['typical_price'] = st[['Adj_High', 'Adj_Low', 'Adj_Close']].mean(axis=1)
st['50d_mva'] = talib.SMA(st['typical_price'].values, timeperiod=50)
__author__ = 'CSPF'
from flask import Flask
from flask_sqlalchemy import SQLAlchemy
from os import path,makedirs,stat
from utils import get_home_dir
db_dir = path.join(get_home_dir(),"db")
if not path.exists(db_dir):
#make sure to create home directory and database directory
makedirs(db_dir)
db_path = path.join(db_dir,"Database.sqlite")
app = Flask("DodoBank")
app.config["SECRET_KEY"]="1234"
app.config['SQLALCHEMY_DATABASE_URI'] = 'sqlite:///'+db_path
db = SQLAlchemy(app)
def test_distillation():
dataset = PetFinderDataset()
hyperparameters = {
"optimization.max_epochs": 1,
"model.names": ["hf_text", "timm_image", "fusion_mlp"],
"model.hf_text.checkpoint_name": "prajjwal1/bert-tiny",
"model.timm_image.checkpoint_name": "swin_tiny_patch4_window7_224",
"env.num_workers": 0,
"env.num_workers_evaluation": 0,
}
teacher_predictor = MultiModalPredictor(
label=dataset.label_columns[0],
problem_type=dataset.problem_type,
eval_metric=dataset.metric,
)
teacher_save_path = os.path.join(get_home_dir(), "petfinder", "teacher")
if os.path.exists(teacher_save_path):
shutil.rmtree(teacher_save_path)
teacher_predictor = teacher_predictor.fit(
train_data=dataset.train_df,
hyperparameters=hyperparameters,
time_limit=30,
save_path=teacher_save_path,
)
# test for distillation
predictor = MultiModalPredictor(
label=dataset.label_columns[0],
problem_type=dataset.problem_type,
eval_metric=dataset.metric,
)
student_save_path = os.path.join(get_home_dir(), "petfinder", "student")
if os.path.exists(student_save_path):
shutil.rmtree(student_save_path)
predictor = predictor.fit(
train_data=dataset.train_df,
teacher_predictor=teacher_predictor,
hyperparameters=hyperparameters,
time_limit=30,
save_path=student_save_path,
)
verify_predictor_save_load(predictor, dataset.test_df)
# test for distillation with teacher predictor path
predictor = MultiModalPredictor(
label=dataset.label_columns[0],
problem_type=dataset.problem_type,
eval_metric=dataset.metric,
)
student_save_path = os.path.join(get_home_dir(), "petfinder", "student")
if os.path.exists(student_save_path):
shutil.rmtree(student_save_path)
predictor = predictor.fit(
train_data=dataset.train_df,
teacher_predictor=teacher_predictor.path,
hyperparameters=hyperparameters,
time_limit=30,
save_path=student_save_path,
)
verify_predictor_save_load(predictor, dataset.test_df)
def __init__(self):
self.homedir = utils.get_home_dir()
def __init__(self):
self.location = os.path.join(utils.get_home_dir(), ".rhnpushcache")
self.session = None
def install(script_name, homedir):
"""Installs the project"""
print 'Initializing new user'
# Let's make sure everyone knows for sure that this is being installed
print """
Do you grant piMessage permission to install on your computer?
PiMessage will have access to:
- your IP address
- any contacts you enter into PiMessage
- all of your conversations through PiMessage
PiMessage will create on your computer:
- a directory to save the following PiMessage information
- a file to save your contacts and their saved IP addresses
- your conversations with contacts
- your preferred text editor
"""
decision = raw_input('Do you want to install PiMessage? (y/n): ')
if decision != 'y':
print 'Not installing. Terminating PiMessage.'
exit(1)
# make the directory
if not os.path.isdir(homedir):
try:
os.mkdir(homedir)
except OSError:
exit(DIR_FAILURE)
# generate user's keys
# todo
if os.system('which vim >/dev/null 2>&1') == 0:
my_editor = 'vim'
else:
my_editor = 'nano'
os.system('clear')
for k in [0, 1, 2, 3]:
if k == 3:
print 'Error: too many tries for editor.'
exit(1)
# Get user input
prompt = '\nWhat is your preferred text editor? Press enter to '
prompt += 'default to '+my_editor+", enter 'cancel' to cancel the "
prompt += 'installation. '
resp = raw_input(prompt)
if resp == 'cancel' or resp == "'cancel'":
# cancel the installation now
exit(1)
elif not resp:
print 'Selecting', my_editor
break
else:
# check if their editor is a valid command
if os.system('which '+resp+' >/dev/null 2>&1') == 0:
my_editor = resp
break
else:
print resp, 'is not a recognized command.'
# write info to files
with open(homedir+'editor', 'w') as fname:
fname.write(my_editor) # doesn't terminate in newline
if not os.path.isdir(homedir + 'conversations'):
try:
os.mkdir(homedir + 'conversations', 0700)
except OSError:
exit(DIR_FAILURE)
with open(homedir+'contacts', 'w') as fname:
fname.write('') # doesn't terminate in newline
dir_path = os.path.abspath(os.path.dirname(sys.argv[0]))
script_name = dir_path+'/pimessage.py'
# alias `pimessage' to point to this script
_bashrc = os.path.join(utils.get_home_dir(), '.bashrc')
grep_results = subprocess.Popen(['grep', '^alias \\+pimessage='+script_name,
_bashrc],
stdout=subprocess.PIPE).communicate()[0]
if grep_results == '':
# must append alias command
try:
with open(_bashrc, 'a') as fname:
fname.write('\n# For PiMessage -- do not delete\n')
fname.write('alias pimessage='+script_name+'\n')
except IOError:
print 'Error applying shell alias for pimessage'
# start pmdaemon at startup
_profile = os.path.join(utils.get_home_dir(), '.profile')
grep_daemon_cmd = ['grep', '^'+dir_path+'/pmdaemon.py', _profile]
grep_results = subprocess.Popen(grep_daemon_cmd,
stdout=subprocess.PIPE).communicate()[0]
if not grep_results:
# must append alias command
start_daemon_cmd = dir_path+'/pmdaemon.py &'
flag_cmd = dir_path+'/pmdaemon.py -f'
try:
with open(_profile, 'a') as fname:
fname.write('\n'.join(['#start pimessage daemon',
start_daemon_cmd, flag_cmd, '']))
except IOError:
print 'Error loading PiMessage daemon in .profile'
# start the daemon manually this time
os.system(start_daemon_cmd)
print 'PiMessage has been successfully installed.'
exit(0)
def load_scheme(self):
scheme_filename = unicode(QFileDialog.getOpenFileName(self, "Open File", get_home_dir(), get_format_filter()))
if scheme_filename:
self.__load_scheme(scheme_filename)
def uninstall():
"""Uninstaller for pimessage"""
status = 0
try:
shutil.rmtree(data_dir, ignore_errors=True)
except OSError:
print 'Error in removing ~/.pimessage'
return 1
# Remove daemon from .profile
try:
_profile = os.path.join(utils.get_home_dir(), '.profile')
with open(_profile, 'r') as fname:
buf = fname.read()
# process buffer
lines = buf.split('\n')
dir_path = os.path.abspath(os.path.dirname(sys.argv[0]))
daemon_line0 = '#start pimessage daemon'
daemon_line1 = dir_path+'/pmdaemon.py &'
daemon_line2 = dir_path+'/pmdaemon.py -f'
lines_to_append = []
for line in lines:
if (line != daemon_line0 and line != daemon_line1 and
line != daemon_line2):
lines_to_append.append(line)
buf = '\n'.join(lines_to_append)
with open(_profile, 'w') as fname:
fname.write(buf)
except Exception as err:
print 'Error in handling ~/.profile'
print '%s' % str(err)
status = 1
# Remove pimessage alias from .bashrc
try:
_bashrc = os.path.join(utils.get_home_dir(), '.bashrc')
with open(_bashrc, 'r') as fname:
buf = fname.read()
# process buffer
lines = buf.split('\n')
alias_line0 = '# For PiMessage -- do not delete'
alias_line1 = 'alias pimessage='+dir_path+'/pimessage.py'
lines_to_append = []
for line in lines:
if line != alias_line0 and line != alias_line1:
lines_to_append.append(line)
buf = '\n'.join(lines_to_append)
with open(_bashrc, 'w') as fname:
fname.write(buf)
except Exception as err:
print 'Error in handling ~/.bashrc'
print '%s' % str(err)
status = 1
if status != 0:
print 'Error removing PiMessage.'
else:
print 'PiMessage has been successfully uninstalled.'
return status
def __init__(self):
self.location = os.path.join(utils.get_home_dir(), ".rhnpushcache")
self.session = None
def check_for_new_excel(driver):
"""
checks for new excel files to download, and if they aren't in the data folder,
downloads them
"""
while True:
try:
driver.get('http://shortsqueeze.com/ShortFiles.php')
break
except TimeoutException:
pass
years = get_years(driver)
# get currently downloaded files
dates_df = pd.read_excel(
get_home_dir(repo_name='scrape_stocks') +
'short_squeeze_release_dates.xlsx', None)
cal_dict = {v: k for k, v in enumerate(calendar.month_name)}
del cal_dict['']
rev_cal_dict = {v: k for k, v in cal_dict.items()}
bimonthly_files = glob.glob(HOME_DIR + 'short_squeeze.com/*.xlsx')
bimonthly_filenames = set([f.split('/')[-1] for f in bimonthly_files])
bimo_dates = [
parse_bimo_dates(f, dates_df, rev_cal_dict) for f in bimonthly_files
]
# if any dates none, it's because that year isn't in excel file of release dates yet
all_none = [d is None for d in bimo_dates]
if any(all_none):
print('latest release date info not in excel file; need to update it')
return
latest_date = max(bimo_dates).date()
latest_year = latest_date.year
check_years = years[years >= latest_year]
files_to_dl = []
filenames = []
for y in check_years:
driver.get('http://shortsqueeze.com/' + str(y) + '.php')
links = driver.find_elements_by_partial_link_text('Download')
for l in links:
link = l.get_attribute('href')
if link == 'http://shortsqueeze.com/ShortFiles.php':
continue
filename = link.split('/')[-1]
if filename in bimonthly_filenames:
continue
files_to_dl.append(link)
filenames.append(filename)
if len(files_to_dl) == 0:
print('no new files to download')
# seems to hang on download, so this will make it continue
driver.set_page_load_timeout(4)
for l in files_to_dl:
try:
print('downloading', l)
driver.get(l) # saves to downloads folder
except TimeoutException:
pass
for f in filenames:
full_fn = '/home/nate/Downloads/' + f
print(full_fn)
if os.path.exists(full_fn):
# os.rename(full_fn, HOME_DIR + 'short_squeeze.com/' + f)
shutil.copy(full_fn, HOME_DIR + 'short_squeeze.com/' + f)
os.remove(full_fn)
computer to receive PiMessages sent over LAN.
2014 Nate Fischer, Ivan De Dios
"""
import socket
import sys
import ip # local file
import utils
import datetime
import time
import os
ERR_FILE = os.path.join(utils.get_home_dir(), '.pimessage', 'daemonError.log')
FLAG_FILE = os.path.join(utils.get_home_dir(), '.pimessage', 'flag')
def err_log(msg):
"""log an error with its time followed by the message describing it"""
now = datetime.datetime.now().time()
formatted_stamp = now.strftime('%H:%M %m/%d/%Y')
err_msg = '\t'.join([formatted_stamp, msg])
# print >>sys.stderr, err_msg
with open(ERR_FILE, 'a') as fname:
fname.write(err_msg+'\n')
def check_startup():
"""
Checks if the daemon is starting in detached mode
def test_customizing_model_names(hyperparameters, ):
dataset = ALL_DATASETS["petfinder"]()
metric_name = dataset.metric
predictor = MultiModalPredictor(
label=dataset.label_columns[0],
problem_type=dataset.problem_type,
eval_metric=metric_name,
)
config = {
MODEL: f"fusion_mlp_image_text_tabular",
DATA: "default",
OPTIMIZATION: "adamw",
ENVIRONMENT: "default",
}
hyperparameters.update({
"env.num_workers": 0,
"env.num_workers_evaluation": 0,
})
hyperparameters_gt = copy.deepcopy(hyperparameters)
if isinstance(hyperparameters_gt["model.names"], str):
hyperparameters_gt["model.names"] = OmegaConf.from_dotlist(
[f'names={hyperparameters["model.names"]}']).names
save_path = os.path.join(get_home_dir(), "outputs", "petfinder")
if os.path.exists(save_path):
shutil.rmtree(save_path)
predictor.fit(
train_data=dataset.train_df,
config=config,
hyperparameters=hyperparameters,
time_limit=10,
save_path=save_path,
)
assert sorted(predictor._config.model.names) == sorted(
hyperparameters_gt["model.names"])
for per_name in hyperparameters_gt["model.names"]:
assert hasattr(predictor._config.model, per_name)
score = predictor.evaluate(dataset.test_df)
verify_predictor_save_load(predictor, dataset.test_df)
# Test for continuous fit
predictor.fit(
train_data=dataset.train_df,
config=config,
hyperparameters=hyperparameters,
time_limit=10,
)
assert sorted(predictor._config.model.names) == sorted(
hyperparameters_gt["model.names"])
for per_name in hyperparameters_gt["model.names"]:
assert hasattr(predictor._config.model, per_name)
verify_predictor_save_load(predictor, dataset.test_df)
# Saving to folder, loading the saved model and call fit again (continuous fit)
with tempfile.TemporaryDirectory() as root:
predictor.save(root)
predictor = MultiModalPredictor.load(root)
predictor.fit(
train_data=dataset.train_df,
config=config,
hyperparameters=hyperparameters,
time_limit=10,
)
assert sorted(predictor._config.model.names) == sorted(
hyperparameters_gt["model.names"])
for per_name in hyperparameters_gt["model.names"]:
assert hasattr(predictor._config.model, per_name)
