read.mrtrix.stream: Read the data payload of an MRtrix TCK/TSF file.

In freesurferformats: Read and Write 'FreeSurfer' Neuroimaging File Formats

Read the data payload of an MRtrix TCK/TSF file.

Description

Reads the payload sequentially in chunks and splits it into groups of values_per_point finite values, which are the points of the individual streamlines.

The end of the data is normally marked by a group of Inf values, but that marker is not always present: the TSF test data shipped with this package has none, and the MRtrix reference reader treats the plain end of the file as a normal end of the streamline data as well. Its absence is therefore not an error here, it is only reported as a warning when the header states a larger number of streamlines than were found, i.e., when data really is missing.

Usage

read.mrtrix.stream(
  con,
  offset,
  dsize,
  endian,
  values_per_point,
  max_groups = Inf,
  skip_groups = 0L,
  bbox = NULL,
  chunk_values = 4e+06,
  filepath = "",
  gzipped = FALSE,
  expected_groups = NA_real_
)

Arguments

con	a connection opened in binary read mode.
offset	numeric, byte offset of the payload.
dsize	integer, bytes per value (4 for Float32, 8 for Float64).
endian	character string, 'little' or 'big'.
values_per_point	integer, 3 for TCK (xyz triples) and 1 for TSF.
max_groups	numeric, stop after this many groups have been selected. Use Inf to read everything.
skip_groups	integer, the number of streamlines to skip before selecting any. Skipped streamlines are never held in memory.
bbox	numeric vector of length 6 or NULL. If given, only streamlines that have at least one point inside the box are selected. The box is given as c(xmin, xmax, ymin, ymax, zmin, zmax) in the coordinate system of the stored data.
chunk_values	integer, number of values to read per chunk.
filepath	character string, the path of the file, used in messages only.
gzipped	logical, whether the connection reads a gzip-compressed file. R cannot seek on gzip connections, so the header bytes are then skipped by reading and discarding them instead.
expected_groups	numeric, the number of groups (streamlines) stated in the header, or NA if the header does not contain a usable count. Only used to warn about a possibly truncated file, never to limit reading.

Value

named list with entries data (matrix with values_per_point columns holding the concatenated groups), lengths (integer vector, number of points per group) and terminator_seen (logical).

freesurferformats documentation built on Sept. 25, 2026, 1:07 a.m.