Read Motion Capture CSV Data

Generic CSV reader that handles common motion capture CSV exports from various systems including Qualisys, Vicon, and generic marker position formats. Returns a PhysioExperiment with position_x, position_y, and position_z assays.

Usage

readMoCapCSV(
  path,
  format = c("auto", "xyz", "wide", "long", "qualisys", "vicon"),
  sampling_rate = NULL,
  header_rows = 1L,
  skip = 0L,
  sep = ",",
  marker_names = NULL,
  coord_columns = NULL
)

Arguments

path

Path to the CSV file.

format

Format hint: "auto" (detect), "xyz" (columns like marker1_x, marker1_y, marker1_z), "wide" (columns like Time, M1X, M1Y, M1Z), "long" (columns: frame, marker, x, y, z), "qualisys" (Qualisys TSV export), or "vicon" (Vicon CSV export). Default "auto".

sampling_rate

Sampling rate in Hz. Required if not detectable from the file (e.g., from a Time column). If NULL and not detectable, an error is raised.

header_rows

Number of header rows. Default 1.

skip

Number of lines to skip before reading. Default 0.

sep

Column separator. Default ",".

marker_names

Explicit marker names (overrides auto-detection from column names). Must match the number of markers detected from columns.

coord_columns

Mapping of coordinate types as a named list of regex patterns, e.g., list(x = "_X$", y = "_Y$", z = "_Z$"). Used only when format is "wide" or "auto".

Value

A PhysioExperiment with assays:

position_x

X coordinates matrix (frames x markers)

position_y

Y coordinates matrix (frames x markers)

position_z

Z coordinates matrix (frames x markers)

The colData contains columns label (marker names) and type ("marker"). The metadata list contains format, source_file, and optionally time (if a Time column was found).

Details

Auto-detection logic

When format = "auto", the function inspects column names:

1. If columns match *_x, *_y, *_z pattern (case-insensitive) with consistent marker prefixes, the "xyz" format is used.

2. If columns match *X, *Y, *Z suffix pattern, the "wide" format is used.

3. If the header contains "Qualisys" or "QTM", the "qualisys" format is used.

4. If columns include frame, marker, x, y, z (case-insensitive), the "long" format is used.

If a Time or time column is present and contains numeric values, the sampling rate is computed from the median time step. An explicit sampling_rate argument always takes precedence.

References

Wickham H (2014). "Tidy Data." Journal of Statistical Software, 59(10), 1-23.

See also

readASF() and readAMC() for Acclaim skeleton and motion files, readMoCapAuto() for automatic format detection.

Examples

path <- tempfile(fileext = ".csv")
df <- data.frame(Marker1_X = rnorm(10), Marker1_Y = rnorm(10),
                 Marker1_Z = rnorm(10))
write.csv(df, path, row.names = FALSE)
pe <- readMoCapCSV(path, format = "wide", sampling_rate = 100)
pe
#> class: PhysioExperiment
#> dim: 10 x 1 
#> assays(3): position_x, position_y, position_z
#> samplingRate: 100 Hz
#> channels(1): Marker1_
#> colData names(2): label, type