robmaz / distmap

Sequence alignment on Hadoop
0 stars 1 forks source link

Add a dry-run option #58

Open magicDGS opened 6 years ago

magicDGS commented 6 years ago

Currently, the --verbose option prints the parsed inputs in the terminal and then exit; nevertheless, there is no real check for the arguments or requirements (such as ReadTools version or existance of files).

It will be nice to debug the current behaviour to have a --dry-run option that will construct the pipeline internally and do the checks, printing on the command line what the pipeline will be (this can be a good addition also if not dry-run is specified, at the beginning of the output). For example, it could say:

PIPELINE (job = ${job_description})
===================================
* Indexing (${reference})
* Upload (${inputs})
* Trimming (${trimming_script} with arguments: "${trim_args}")
* Map (${mapper} with arguments  "${mapper_args}")

WARNING: trimming is perform on upload and it is OBSOLETE

This will be useful for debugging, but also for the final user to check how the procedure will be.

robmaz commented 6 years ago

Good idea in principle, but difficult to implement with the current convoluted structure.

magicDGS commented 6 years ago

So for the long term!