Command line tools for runs.

[kazewong]

Currently the main way to run an inference run is to go through the examples, write a script using the Jim library, and run something along the line of python my_script.py.

This is good for exploration, but in production, it would be nice to have some more infrastructure that support programmatically generating config files, then run the inference.

[ThibeauWouters]

More generally, adding more support for new developers working with jim would be beneficial. For the Jim BNS PE paper, I coded up an injection script that made use of argparse to easily allow users to handle hyperparameters etc through the command line. The Python script is at [1] and an example bash script is at [2]. This could be a good first step towards this issue, although there might be significant overlap with the RunManager. @kazewong can perhaps comment on that

[1] https://github.com/ThibeauWouters/TurboPE-BNS/blob/main/injections/injection_recovery.py
[2] https://github.com/ThibeauWouters/TurboPE-BNS/blob/main/injections/run_TaylorF2.sh