I’m having some difficulties with CreateBatchFiles and LoadData, specifically with file paths when switching from the local computer I create the pipeline with to the computing cluster.
What I’m doing so far:
- Create a .csv file with the correct file paths for the local computer, along with metadata and group indexes etc.
- At the end of the pipeline use CreateBatchFiles, setting the local root path and the cluster root path. Run this on the local computer to generate a Batch_data.h5
- Then on the cluster I run
cellprofiler --get-batch-commands Batch_data.h5to generate a number of commands.
When I then try to run any of these I get the error file:
INFO:PipelineStatistics:Pipeline saved with CellProfiler version 20140723174500 INFO:PipelineStatistics:Times reported are CPU times for each module, not wall-clock time ERROR:cellprofiler.pipeline:Error detected during run of module LoadData ... ... OSError: [Errno 2] No such file or directory: '/exports/eddie/scratch/s1027820/test_2/Z:/scott/CP_cluster_test/test_2/load_data_eddie_paths.csv' INFO:PipelineStatistics:Thu Jul 7 19:06:04 2016: Image # 1, module LoadData # 1: 0.01 sec
Z:/scott/CP_cluster_test/test_2/ is my local path and
/exports/eddie/scratch/s1027820/test_2/ is the cluster root path. So it looks like it’s pasting together my local and root file paths before the LoadData .csv file.
Any idea where I’m going wrong? The only clear instructions for I can find for LoadData and CreateBatch files are for CellProfiler1.0.