DPclust input dataset.Rdata doesn't load existing object

[Lesky47]

Hello and thanks for the great tool! I've been running it with the latest version but run into some issues:

dpclust/R/DirichletProcessClustering.R

Line 134 in d47e349

if (file.exists(paste(outdir, "/", rdata_file_name, sep=""))) {

The "rdata_file_name" is set to contain the exact time in its name, so every time I run runDP(), the "rdata_file_name" will be something like "Seed-123_Date-2024-08-22_14:18:22.734335_dataset.RData" and file.exists(paste(outdir, "/", rdata_file_name, sep="")) will always return FALSE even if there's already a dataset generated previously (Or generated elsewhere using a different name format). Also the current naming is not allowed in my system (':' cannot be used in file name).

[Avramis]

Hi @Lesky47,

Thanks for your ticket.

The fix was initially implemented because we had the opposite issue—when an analysis was rerun with new parameters in the same output folder, the code was loading old data "dataset.RData". We’re currently working on an update so that users can specify whether they want the tool to use an existing dataset.RData file or generate a new one.