Write neurons from a neuronlist object to individual files, or a zip archive

write.neurons(
  nl,
  dir,
  format = NULL,
  subdir = NULL,
  INDICES = names(nl),
  files = NULL,
  include.data.frame = FALSE,
  metadata = FALSE,
  Force = FALSE,
  cl = NULL,
  ...
)

Arguments

nl: neuronlist object
dir: directory to write neurons, or path to zip archive (see Details).
format: Unique abbreviation of one of the registered file formats for neurons including 'swc', 'hxlineset', 'hxskel' (skeletons) and 'ply', 'obj' (neuron meshes). If no format can be extracted from the filename or the ext parameter, then it defaults to 'swc' for skeletons and 'ply' for meshes.
subdir: String naming field in neuron that specifies a subdirectory OR expression to evaluate in the context of neuronlist's df attribute
INDICES: Character vector of the names of a subset of neurons in neuronlist to write.
files: Character vector or expression specifying output filenames. See examples and write.neuron for details.
include.data.frame: Whether to include the metadata when writing a zip file (it will be called "write.neurons.dataframe.rds").
metadata: Whether to encode some metadata in the header file (currently only supported for SWC format). Either a data.frame or TRUE to indicate that the attached data.frame should be written. Default FALSE.
Force: Whether to overwrite an existing file
cl: Either the integer number of cores to use for parallel writes (2 or 3 seem useful) or a cluster object created by makeCluster. See the cl argument of pbsapply for details.
...: Additional arguments passed to write.neuron

Value

the path to the output file(s), absolute when this is a zip file.

Details

See write.neuron for details of how to specify the file format/extension/name of the output files and how to establish what output file formats are available. A zip archive of files can be written by specifying a value of dir that ends in .zip. When rds files (R's binary data representation, which is compressed by default) are stored inside a zip file the are not further compressed (zip option 0).

Author

jefferis

Examples

if (FALSE) {
# write some neurons in swc format
write.neurons(Cell07PNs, dir="testwn", format='swc')
# write some neurons in swc format for picky software
write.neurons(Cell07PNs, dir="testwn", format='swc', normalise.ids=TRUE)
# write some neurons in swc format and zip them up
write.neurons(Cell07PNs, dir="testwn.zip", format='swc')

# write some neurons in R's native RDS format using 3 cores for
# parallel writes and then zip them up (storing rather than compressing)
write.neurons(Cell07PNs, dir="testwn.zip", format='rds', cl=3)

# write some neurons in Amira hxlineset format
write.neurons(Cell07PNs, dir="testwn", format='hxlineset')

# write some neuron meshes in Stanford ply format (the default for meshes)
write.neurons(myneurons, dir="testwn")
# specify the format to avoid a warning. Write to a zip file.
write.neurons(myneurons, dir="testmeshes.zip", format='ply')
# Wavefront obj format
write.neurons(myneurons, dir="testwn", format='obj')

# organise new files in directory hierarchy by glomerulus and Scored.By field
write.neurons(Cell07PNs,dir="testwn",
  subdir=file.path(Glomerulus,Scored.By),format='hxlineset')
# ensure that the neurons are named according to neuronlist names
write.neurons(Cell07PNs, dir="testwn", files=names(Cell07PNs),
  subdir=file.path(Glomerulus,Scored.By),format='hxlineset')
# only write a subset
write.neurons(subset(Cell07PNs, Scored.By="ACH"),dir="testwn2",
  subdir=Glomerulus,format='hxlineset')
# The same, but likely faster for big neuronlists
write.neurons(Cell07PNs, dir="testwn3",
  INDICES=subset(Cell07PNs,Scored.By="ACH",rval='names'),
  subdir=Glomerulus,format='hxlineset')
# set file name explicitly using a field in data.frame
write.neurons(subset(Cell07PNs, Scored.By="ACH"),dir="testwn4",
  subdir=Glomerulus, files=paste0(ID,'.am'), format='hxlineset')
}