rapt offline mode

[OH-AU]

In ticket #10 there is a reference to running rapt in offline mode. We also have a need to do this.
I've managed to modify the python script so it uses a local rapt.sif image via singularity. lsof confirms it's using the local image, but the last statement when running the --test says:
[2023-04-21 16:44:04] Running functional tests...
With no further activity as it can't access the internet.
Using an actual fastq file for input, the program appears to run as expected, but after completing with DONE, it seems to hang at "Error sending metrics to NCBI ..." without ever returning a prompt.
When using a real it tries to download more data from amazon.

Summary: Could the instructions on running locally be made publicly available. The compute nodes on the cluster have no/limited internet access (same as I suspect the user in #10) so an offline option is the only way we can get usage of the tool.
Thanks.

[thibaudnis]

Hi, thanks for your question. Can you please running run_rapt.py with --no-usage-reporting? I think it will address the issue you see.

[OH-AU]

1 . This does address the hang at the end of processing - Thank you. Is there an environment variable that could be used instead of a command line option as I don't expect end users to remember to use that switch every single time?
2 . It still required downloading a number of datasets to run the the analysis - so would still like to have details on how to run the pipeline offline.

[thibaudnis]

There is currently no environment variable that can be used as an alternative, but yes, this is something we should think about... As for downloads, I don't see a way to avoid them. The annotation does require a largish database, and of course the pipeline itself has to be uploaded on the machine. Once it is there though, access to the internet is not needed.