Create script for accessing cohort query matches on BIC

[surchs]

As a query user at the BIC, when I find a cohort of subjects that exist at the BIC with the query tool, I want to then be able to follow a few easy steps to start working with these data locally, so that I can do analysis on this cohort.

Steps that are needed here:

• Decide on: how should a local BIC user (assuming access to the data) proceed after finding a cohort of interest, in order to start accessing / working with the data we show them
• A simple CLI or script (maybe installed already on the BIC) that can take a query tool output and generate symlinks to a target directory of the desired subjects
• a way to download easily the input to the script, in a way that is robust to me including remote (i.e. inaccessible) datasets in my cohort

Assumptions:

• User has access to the root dataset folder, e.g. /data/pd
  • usually, this means they will need to be added to a specific user group
• Different datasets are stored on the same server (e.g., Calgary dataset on CC)
• Dataset results are disaggregated

Limitations:

• Cannot provide specific access to pheno data for now since we don't store the path of the TSV (and it doesn't have a guaranteed location within the dataset)
• Cannot provide specific access to derivatives for now

Desired outcome:

• A simple bash script, without argparse capabilities, that takes at least 2 positional args:
  • participants results TSV from query tool
  • a target location for the symlinks
  • dataset root? or name? (how to know where the directory tree begins)?
• Output should be a skeleton BIDS-like dataset directory going down to the level of the subject or session (whichever is the lowest level in the provided paths), which then is a symlink to the actual directory in the data storage location
• Script should check if the symlinked location exists, otherwise error + exclude that subject-session (?)
• Script should also create a simple README.md that includes a disclaimer that the created directory tree is not an actual valid BIDS dataset (and that BIDS validation should be skipped if using BIDS app pipelines)
• Script should live under /data/pd for now