engram.procedural package

Submodules

engram.procedural.analyze module

engram.procedural.data module

engram.procedural.data.events(feature, time, settings, prev_len)

engram.procedural.data.select(feature, time, settings, prev_len=None)

engram.procedural.data.trials()

engram.procedural.events module

engram.procedural.events.Neurogenesis()

engram.procedural.events.RAM(reader)

engram.procedural.events.select(name, reader)

engram.procedural.features module

engram.procedural.features.LFP(trace, settings)

engram.procedural.features.STFT(trace, settings)

engram.procedural.features.multiscale(trace, settings)

engram.procedural.features.normalize(data, settings)

engram.procedural.features.select(name, trace, settings)

engram.procedural.features.spikes(trace, settings)

engram.procedural.filters module

engram.procedural.filters.butter_bandpass(lowcut, highcut, fs, order=5)

engram.procedural.filters.butter_bandpass_filter(data, lowcut, highcut, fs, order=5)

engram.procedural.filters.butter_lowpass(cutoff, fs, order=5)

engram.procedural.filters.butter_lowpass_filter(data, cutoff, fs, order=5)

engram.procedural.filters.select(filter, data, min=0, max=None, fs=2000, order=5)

engram.procedural.missingdata module

engram.procedural.missingdata.interpolate_nans(y)

Helper to handle indices and logical indices of NaNs.

Input:

• y, 1d numpy array with possible NaNs

Output:

• nans, logical indices of NaNs
• index, a function, with signature indices= index(logical_indices), to convert logical indices of NaNs to ‘equivalent’ indices

Example:

>>> # linear interpolation of NaNs
>>> nans, x= nan_helper(y)
>>> y[nans]= np.interp(x(nans), x(~nans), y[~nans])

engram.procedural.models module

engram.procedural.models.cnn(shape)

engram.procedural.models.custom(shape)

engram.procedural.models.lstm()

engram.procedural.models.md(shape)

engram.procedural.models.mimo(shape)

engram.procedural.models.select(model, shape)

engram.procedural.neo_handler module

engram.procedural.neo_handler.unpackNeo(reader)

engram.procedural.predict module

engram.procedural.predict.predict(model=None, mneme=None)

engram.procedural.train module

engram.procedural.train.create_dataset(features=None, labels_for_categories=None)

Load and parse dataset. Args:

filenames: list of image paths labels: numpy array of shape (BATCH_SIZE, N_LABELS) is_training: boolean to indicate training mode

engram.procedural.train.get_data(features=None, labels_for_categories=None)

engram.procedural.train.train(model_type='CNN', in_matrix=None, labels=None)

Module contents

:mod:’engram.procedural’ provides functions for processing Engram data structures and encoding them into models