lazear/sage

Input handling guardrails and glob expansion

Opened this issue · 2 comments

acesnik commented

Hi there, this is an exciting tool; thanks for making it!

Here are a couple of minor requests on input handling:

  1. It could be nice to check for file existence prior to generating the fragment index. One has to wait for that process to finish to find out the mzML doesn't exist in the current directory, for example.
  1. Performing glob::glob expansion of asterisks in the mzML list could make a little improvement in CLI experience
lazear commented
  • This is actually desired behavior on my end. Crashing or exiting because files don't exist or are corrupted is annoying if you are running a big search (e.g. failing on file 900 out of 1000)
  • This would be a nice feature (for windows users, since *nix shells will handle globbing automatically if specified via CLI) - would also be great to enable glob usage for S3 (via list-objects-v2 API). Are you interested in contributing a PR for this feature?

I would note that I also strongly recommend running sage via WSL if you're on windows... performance is definitely worse on Windows

acesnik commented

Ah, sounds good!

  • How about creating a warning for file existence early in the run instead of erroring out?
  • I drafted up a glob feature PR (untested)
  • Interesting about the performance on Windows! I'll definitely keep that in mind.