fslr implicitly is partnered with the nifti format from oro.nifti. All functions use the system command to implement the FSL functions. All functions will check to see whether the file passed in a character path to a filename. If the object is a nifti object, the programs will create a temporary file using tempfile() and then return the name of this file with the checkimg command. If a character path is passed to the file argument into an fslr command, then this will run in FSL assuming that file exists.
If R is run using a GUI, then fsl.path must be set using options(fsl.path='/path/to/fsl'). have.fsl() checks to see if FSL is in the PATH or the options are correctly set and returns TRUE and returns FALSE otherwise.
For example, you can wrap example functions with a logical check for have.fsl() to have the execute if FSL exists:
if (have.fsl()) {
print(fsl_version())
}## [1] "6.0.4"In functions where an image is the end result, the outfile and retimg arguments are present. If outfile is specified, the user wants a file to be saved to disk. If outfile is not specified, a temporary file (using tempfile) will be created for the result and deleted at the end of the R session. If retimg=TRUE, then the user specified that an object of class nifti will be returned to R. If both retimg=TRUE and outfile is specified, the image will be written to disk, read into R and the nifti object will be returned. If retimg=FALSE and outfile is not specified, the function will fail as fslr would be calculating an image and no result would be able to be obtained since the outfile is deleted.
The main functions implemented are fslmaths and fslstats. fslmaths usually returns images as the result and fslstats usually returns a vector of information.
Functions functions such as fslmask, fslerode, fslsmooth, and fslfill are implemented wrappers for ease of use, but are essentially calling fslmaths with specific options. Functions such as fslrange is an example of a wrapper for fslstats. As fslstats outputs are variable in their format, the results are returned simply as a character vector which needs to be parsed. This GitHub repository will be used for development and issues.