Changed swarm output directory to be the current swarm id

[wjs20]

The way that lightdock-rust currently reads and writes files is a little confusing. Because swarms writes to the current directory by default, you either have to copy the structure files into every swarm directory, which seems inefficient, or you have to cd into every directory and refer to the structure and settings files using dotted relative paths. The change means that instead of doing this:

cd swarm_0
cp ../lightdock_1czy_protein.pdb ../lightdock_1czy_peptide.pdb .
cp ../rec_nm.npy ../lig_nm.npy .
cp -R ../../../data .
time ../../../target/release/lightdock-rust ../setup.json ../init/initial_positions_0.dat 100 dfire

You can just do this:

/target/release/lightdock-rust setup.json init/initial_positions_0.dat 100 dfire

You can also run this in parallel, because each process will be writing to a different directory and the file names will not clobber each other. It does this by extracting the swarm id from the initial_positions_number.dat filename, and using this as the directory name to write the gso_number.out files.

Let me know what you think.

[wjs20]

My initial implementation didn't work with nested paths (e.g. init/initial_positions_0.dat) I have fixed this in edb6c37

[brianjimenez]

Thank you @wjs20 for your contribution, did you have a look at 0.3.0 branch? That is the most updated code (still not merged into main though) and something similar was implemented (no need to copy structures or data anymore). Please, could you rebase your contribution into 0.3.0? Thanks in advance for your time and contributed code!