d-bohn / fizzr

A tidy port of the PhysioScripts suite

fizzr: A tidy port of PhysioScripts

fizzr

DISCLAIMER Very much a work in progress!

The fizzr package (physiological analysis in r) is a tidy implementation/port of the PhysioScripts suite of functions, originally published by Christie and Gianaros (2013). Further documentation and relevant publication can be found online in the journal Behavioral Research Methods.

The original .RData file for the PhysioScripts package is included in the data folder. However, all of these original functions have been transferred to the package as “legacy” functions that are still callable (as of now).

Installing

devtools::install_github('d-bohn/fizzr')
library(fizzr)

Browsing

Each function is being re-evaluated for ‘tidiness,’ and then ported (slowly) to this package. If you would like to browse the original PhysioScripts suite of functions that will be available once this package is finished, you can do so by:

library(fizzr)
ls('package:fizzr')

Timeline/Steps

1). Import all of the PhysioScripts functions “as is” into the fizzr package (complete)

2). Tidy and re-write the functions

• Clean/format code according to style guide

• Add documentation to separate files

• Remove ability to use interactive file selection

  • Remove dependency for tcltk

  • Remove unnecessary interactive functions (e.g., confirm.file.replace)

• Replace . with _ to emulate tidy logic

• Alter code to allow “pipe-able” chains

3). Deploy version 1.0 with example data

Currently ‘tidied’ Functions

So far, the following functions are be called directly from the package without having to load the PhysioScripts suite:

• process_ecg()

Motivation and Rationale

The current motivation for porting these functions to a new package is three-fold. First, PhysioScripts is no longer maintained by the original author(s). Providing appropriate documentation and package structure will allow for the community to easily add to, modify, and continue development on this package.

Second, R lacks an intuitive and straightforward package to do physiological data pre-processing. While the functions provided by PhysioScripts are meant to be an end-to-end solution, most of the currently ported functions are used for pre-processing heart rate data. The pre-processed data can then passed to other packages that require cleaned data (e.g., RHRV).

Finally, porting code that you did not write yourself is an excellent learning opportunity for working with second-hand code and package building. Creating a package of functions from original code that both maintains its functionality while also fitting into a larger ecosystem (e.g., the tidyverse) is a great learning exercise. One immediate change from PhysioScripts that I made was renaming functions to fit the “underscore” naming convention often used in the tidy ecosystem. This convention is desirable over the “period” naming convention for a number of reasons, one of which being that periods are used in python to denote “within,” much like the $ operator in R.

References

Christie, Israel C, and Peter J Gianaros. 2013. “PhysioScripts: An Extensible, Open Source Platform for the Processing of Physiological Data.” Behavior Research Methods 45 (1): 125–31. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3767411/.