Analyze HDF5 files

[ARTUR551]

Hi,

DLCAnalizer is great, thanks for developing it.
Do you think it's possible to add the feature of analyzing HDF5 files?

Best,

A.

[lukasvonziegler]

Sorry for the late response, but i do not get a notification about issues unless you @ me in the messages. I still respond because even if it is way to late for you it may still be useful for other users

i wrote a importer for SLEAP (https://sleap.ai/) that loads data from hdf5 files (code below). you may need to adjust it based on the structure of the hdf5 file from DLC. Hope this is helpful. Best, Lukas

#' Reads SLEAP Tracking data from .h5 file(s) and returns a TrackingData object
#'
#' @param file path to a SLEAP tracking .h5 file
#' @param fps frames per second of the recording. required to enable time resolved metrics
#' @return An object of type TrackingData
#' @examples
#' ReadSLEAPDataFromh5("SLEAPData/Data.h5", fps = 25)
ReadSLEAPDataFromh5 <- function(files,fps = 1){
library(rhdf5)
library(data.table)
out <- list()
out$data <- list()
for(j in files){
h5f = H5Fopen(j)
for(i in 1:length(h5f$node_names)){
out$data[[paste(h5f$node_names[i])]] <- data.frame(frame = 1:length(h5f$tracking_scores),
x = h5f$tracks[,i,1,],
y = h5f$tracks[,i,2,],
likelihood = h5f$point_scores[,i,])
process <- is.nan(out$data[[paste(h5f$node_names[i])]]$x)
out$data[[paste(h5f$node_names[i])]]$x[process] <- NA
out$data[[paste(h5f$node_names[i])]]$y[process] <- NA
}
}

if(fps == 1){
warning("no fps set. setting fps to 1. keep in mind that time based analyses are resolved in frames / second")
}
out$frames <- 1:length(h5f$tracking_scores)
out$fps <- fps
out$seconds <- out$frames / fps

out$median.data <- NULL
for(i in names(out$data)){
out$median.data <- rbind(out$median.data, data.frame(PointName = i, x = median(out$data[[i]]$x, na.rm = TRUE), y = median(out$data[[i]]$y, na.rm = TRUE)))
}
rownames(out$median.data) <- out$median.data$PointName

out$point.info <- data.frame(PointName = names(out$data), PointType = "NotDefined")
out$distance.units <- "pixel"
out$labels <- list()
out$filename <- last(strsplit(files[1],split = "/")[[1]])
out$object.type = "TrackingData"

return(out)
}