def resample_batch(self, batch: DataFrame,
resample_config: dict) -> DataFrame:
if not str2bool(get_recursive_config(resample_config, 'enabled')):
return batch
target_sampling_rate = get_recursive_config(resample_config,
'target_sampling_rate')
method = get_recursive_config(resample_config, 'method')
resampled_batch = batch.resample(target_sampling_rate)
if method == 'ffill':
return resampled_batch.ffill()
elif method == 'bfill':
return resampled_batch.bfill()
elif method == 'mean':
interpolation_method = get_recursive_config(
resample_config, 'interpolation_method')
return resampled_batch.mean().interpolate(
method=interpolation_method)
elif method == 'fill_value':
fill_value = get_recursive_config(resample_config, 'fill_value')
return resampled_batch.asfreq(fill_value=fill_value)
elif method == 'interpolate':
interpolation_method = get_recursive_config(
resample_config, 'interpolation_method')
return resampled_batch.interpolate(method=interpolation_method)
else:
return resampled_batch
def __init__(self, config):
super().__init__(config)
self.show_columns_progress = self.get_config('show_columns_progress',
default=[])
self.show_data_frame = str2bool(
self.get_config('show_data_frame', default=False))
def __init__(self, config):
super().__init__(config)
self.format = self.get_config('format', default='svg')
self.dpi = self.get_config('dpi', default=None)
self.show_plots = str2bool(self.get_config('show_plots', default=True))
self.save_to_image = str2bool(self.get_config('save_to_image', default=False))
self.save_to_pickle = str2bool(self.get_config('save_to_pickle', default=False))
self.save_path = self.get_config('save_path', default='./out/')
self.columns_to_plot = []
self.accumulated_data_frame = None
self.cli_arguments = get_cli_arguments()
for figure_config in self.get_config('figures'):
for plot_config in figure_config['plots']:
self.columns_to_plot.append(plot_config['column'])
def __init__(self, config: dict):
super().__init__(config)
self.verbose = str2bool(self.get_config('verbose', default=False))
self.batch_size = self.get_config('batch_size', default=1)
self.epochs = self.get_config('epochs', default=1)
self.save_best_model = self.get_config('save_best_model', default=False)
self.load_model = self.get_config('load_model', default=None)
self.shuffle = self.get_config('shuffle', default=True)
self.next_batch = []
self.callbacks = []
self.autoencoder = None
if self.save_best_model:
self.callbacks.append(ModelCheckpoint(get_current_out_path('best_model.hdf5'), save_best_only=True, monitor='loss'))
def __init__(self, config: dict):
super().__init__()
self.config = config
self.inverted = str2bool(self.get_config('inverted', default=False))