ep-bandpass-filter-selector 0.0.4

The EP Bandpass Filter Selector package cycles through various combinations of high- and low-pass cutoffs of the bandpass frequency filter within user-specified limits and selects the optimal filtering option. Optimization is carried out basing on the ratio of the EP interpeak amplitude to the EP curves variability coefficient.

ep-bandpass-filter-selector

The EP Bandpass Filter Selector package cycles through various combinations of high- and low-pass cutoffs of the bandpass frequency filter within user-specified limits and selects the optimal filtering option. Optimization is carried out basing on the ratio of the EP interpeak amplitude to the EP curves variability coefficient.

Requirements

Package require python >= 3.8. For development require python3-venv. Tested on Ubuntu 20.04.

Installation

For usual using:

$ pip install ep_bandpass_filter_selector

For development:

$ git clone https://github.com/yaricp/ep-bandpass-filter-selector.git
$ cd ep-bandpass-filter-selector/
$ cd scripts/
$ ./install.sh

install.sh will download and set up the necessary packages in the .venv folder.

Usage

This package can be imported as a separated module:

$ python3
>>> from eeg_filters.upload import prepare_data
>>> from ep_bandpass_filter_selector import PassbandSelector
>>> (
    sample_rate, list_times, list_ticks, list_out
) = prepare_data("file_data_path.dat")
>>> pbs = PassbandSelector(
    curves=list_out,
    tick_times=list_ticks,
    fsr=sample_rate,
    max_search_range: (0.016, 0.023),
    min_search_range: (0.022, 0.028)
)
>>> result = pbs.start()
>>> print(result[0])
20,200

In this example, all curves from the file_data_path.dat were used for selecting the passband filter. Also, additional parameters were used with the default values. Their values are listed here:

filter_low_limit_range = 1, 30 step_low_filter = 1 filter_high_limit_range = 100, 500 step_high_filter = 10

You can declare them when creating an item of PassbandSelector. For example:

>>> pbs = PassbandSelector(
    curves=list_out,
    tick_times=list_ticks,
    fsr=sample_rate,
    max_search_range=(0.016, 0.023),
    min_search_range=(0.022, 0.028),
    filter_low_limit_range=[5, 20],
    step_low_filter=5,
    filter_high_limit_range=[100, 300],
    step_high_filter=50,
    type_mean="average",
    cheb_filter_order=3,
    cheb_ripple=3
)

Another notation is also possible for declaring parameters. For example:

>>> pbs = PassbandSelector(
    curves=list_out,
    tick_times=list_ticks,
    fsr=sample_rate,
    max_search_range=(0.016, 0.023),
    min_search_range=(0.022, 0.028),
    fllr=[5, 20],
    slf=5,
    fhlr=[100, 300],
    shf=50,
    tm="average",
    chfo=3,
    chr=3
)

Export data

On this moment there is only export to CSV format.

>>> from ep_bandpass_filter_selector import export_data
>>> export_data(
    "filepath_for_export.csv",
    list(result[0]),
    result[1]
)