Output track file duplication when TempExt fails

[qntmCharles]

I have been running TempestExtremes via the one-shot wrapper provided by TCTrack on one-year chunks of data on JASMIN. Tracks were successfully produced for the year 1950, but all subsequent years failed during the StitchNodes step because the data pre-processing resulted in mis-matched time steps. However, TCTrack continued to produce output files. Upon further investigation I found that these files were all identical, containing the tracks for 1950. I am running with the option to produce intermediate files 'nodes_out.dat' and 'tracks_out.txt', which are stored in the same directory that the computation script (Jupyter notebook) is running.

It seems like TCTrack is using the intermediatea files to produce an output track file even when the TempExt step fails.

[jatkinson1000]

@qntmCharles can you provide us a script you used for thios so we can try and reproduce on CSD3 to further investigate?

Is it that we need to name the output files "nodes_out_1.dat", "nodes_out_2.dat" etc.?

[jatkinson1000]

This perhaps relates to #110 and as I finalise TSTORMS I realise we may want to allow users to provide a prefix/suffix to prepend/append to outputs if they are being placed in a common output location.