Metadata-Version: 2.1
Name: tsfresh-ppi
Version: 2021.8.29
Summary: Add peak-to-peak interval features to tsfresh
Home-page: https://github.com/atpage/tsfresh_ppi
Author: Alex Page
Author-email: a.t.page@gmail.com
License: UNKNOWN
Project-URL: Bug Tracker, https://github.com/atpage/tsfresh_ppi/issues
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Description-Content-Type: text/markdown
License-File: LICENSE

# tsfresh_ppi

This package provides some peak-to-peak interval (PPI) variability features to augment `tsfresh` [1].
Currently, `tsfresh` will find and count peaks using a couple of different methods.
However, it does not measure the variability in timing between those peaks.

## Installation

Option 1: Clone this repo.  Install prereqs from `requirements.txt`.  Then install the cloned directory, e.g. with `pip3 install -e [repo path]`.

Option 2: Install directly from PyPI (`tsfresh-ppi`).

## Usage

    from tsfresh import extract_features
    from tsfresh_ppi import get_fc_parameters

    my_signal = ...  # some pandas DataFrame

    # The default way to extract features with tsfresh looks something like this:
    features = extract_features(
        my_signal,
        ...
    )

    # This is how to get all of the default tsfresh features, plus default PPI features:
    fc_params_with_ppi = get_fc_parameters()
    features = extract_features(
        my_signal,
        default_fc_parameters = fc_params_with_ppi,
        ...
    )

## References
[1] https://github.com/blue-yonder/tsfresh


