Implement #HQ directives

[spirali]

Implement equivalent to #PBS directives in shell scripts

[giovannipizzi]

This would be very useful! Since our software AiiDA expects/requires the directives to be put within the file, and submit simply with hq submit script.sh (as a way to store this information for later debugging, if needed) at the moment we are planning to write a simple wrapper script hq-wrapper-submit script.sh that will parse lines like

#HQ --cpus=32 
#HQ --cwd=run2bis

...

and actually submit hq submit --cpus=32 --cwd=run2bis script.sh.

However, having this directly supported by hq would be great and avoids us maintaining a wrapper, and having to ask users to install also this together with hq.

I think the logic might be simple:

1. start from the first line in the file, and check line by line
2. skip all lines that are empty or only contains white spaces after removing comments (signalled by #); this also includes the first shebang line (#!/bin/bash or similar) if present
3. if the line starts with #HQ , get the rest and pass it as a command line (we would do it "blindly", but since you know the parameters you could already check if they are valid to avoid injection of commands, e.g. if one writes #HQ --cpus=32; rm *, or at least escape each of them with quotes? :-)
4. as soon as you find a non-empty line, stop processing the #HQ commands (i.e., in a file like this:

   #HQ --cpus=32 
   echo 123
   #HQ --cwd=run2bis
   

   only the first directive would be passed to hq, but the second --cwd would be ignored because there is a non-comment line echo 123 before).

[Kobzol]

Something like this indeed sounds reasonable, we could take inspiration from e.g. Slurm commands.
It's currently a bit lower priority though, as there are a lot of other issues (a lot of them suggested by you :D ) that we want to resolve first.