Add MD progress bar

[k-harris27]

Resolves #405

I think there are a few ways that this could be done - this one (attaching the progress bar as an observer function) does not discriminate between heating and production MD steps. I tried using irun as had been suggested, but since half of the ASE ensembles don't support irun right now, this seemed like the best way.

There's a hack at md.py lines 1060/1061 to get around ASE's NPT ensemble calling observers 1 less time than expected. Once the fix is incorporated (https://gitlab.com/ase/ase/-/merge_requests/3598), those lines will need to be deleted.

There is also a test that should check each ensemble fills the progress bar completely by the end of the run.

[ElliottKasoar]

Resolves #405

I think there are a few ways that this could be done - this one (attaching the progress bar as an observer function) does not discriminate between heating and production MD steps. I tried using irun as had been suggested, but since half of the ASE ensembles don't support irun right now, this seemed like the best way.

There's a hack at md.py lines 1060/1061 to get around ASE's NPT ensemble calling observers 1 less time than expected. Once the fix is incorporated (https://gitlab.com/ase/ase/-/merge_requests/3598), those lines will need to be deleted.

There is also a test that should check each ensemble fills the progress bar completely by the end of the run.

Thanks for this!

When you say does not discriminate, do you mean it has a single progress bar for a combined heating + MD simulation, with the steps for both summed?

[ElliottKasoar]

I'm not sure the heating is currently filling up completely (with or without MD after)?

janus-core % janus md --ensemble nvt --temp-time 50 --temp-start 0 --temp-end 20 --temp-step 10 --traj-every 1 --stats-every 1 --struct tests/data/NaCl.cif
cuequivariance or cuequivariance_torch is not available. Cuequivariance acceleration will be disabled.
Using Materials Project MACE for MACECalculator with /Users/elliottkasoar/.cache/mace/20231210mace128L0_energy_epoch249model
Using float64 for MACECalculator, which is slower but more accurate. Recommended for geometry optimization.
[codecarbon INFO @ 17:30:49] offline tracker init
[codecarbon WARNING @ 17:30:49] Multiple instances of codecarbon are allowed to run at the same time.
Performing MD simulation... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╸━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 100/150 0:00:02
(janus-core) elliottkasoar@dhcp-10-248-26-147 janus-core % 
(janus-core) elliottkasoar@dhcp-10-248-26-147 janus-core % 
(janus-core) elliottkasoar@dhcp-10-248-26-147 janus-core % janus md --ensemble nvt --temp-time 50 --temp-start 0 --temp-end 20 --temp-step 10 --traj-every 1 --stats-every 1 --struct tests/data/NaCl.cif --steps 100
cuequivariance or cuequivariance_torch is not available. Cuequivariance acceleration will be disabled.
Using Materials Project MACE for MACECalculator with /Users/elliottkasoar/.cache/mace/20231210mace128L0_energy_epoch249model
Using float64 for MACECalculator, which is slower but more accurate. Recommended for geometry optimization.
[codecarbon INFO @ 17:31:13] offline tracker init
[codecarbon WARNING @ 17:31:13] Multiple instances of codecarbon are allowed to run at the same time.
Performing MD simulation... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╺━━━━━━━━━━━━━━━━━━━━━━━ 200/250 

[k-harris27]

I'm not sure the heating is currently filling up completely (with or without MD after)?

janus-core % janus md --ensemble nvt --temp-time 50 --temp-start 0 --temp-end 20 --temp-step 10 --traj-every 1 --stats-every 1 --struct tests/data/NaCl.cif
cuequivariance or cuequivariance_torch is not available. Cuequivariance acceleration will be disabled.
Using Materials Project MACE for MACECalculator with /Users/elliottkasoar/.cache/mace/20231210mace128L0_energy_epoch249model
Using float64 for MACECalculator, which is slower but more accurate. Recommended for geometry optimization.
[codecarbon INFO @ 17:30:49] offline tracker init
[codecarbon WARNING @ 17:30:49] Multiple instances of codecarbon are allowed to run at the same time.
Performing MD simulation... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╸━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 100/150 0:00:02
(janus-core) elliottkasoar@dhcp-10-248-26-147 janus-core % 
(janus-core) elliottkasoar@dhcp-10-248-26-147 janus-core % 
(janus-core) elliottkasoar@dhcp-10-248-26-147 janus-core % janus md --ensemble nvt --temp-time 50 --temp-start 0 --temp-end 20 --temp-step 10 --traj-every 1 --stats-every 1 --struct tests/data/NaCl.cif --steps 100
cuequivariance or cuequivariance_torch is not available. Cuequivariance acceleration will be disabled.
Using Materials Project MACE for MACECalculator with /Users/elliottkasoar/.cache/mace/20231210mace128L0_energy_epoch249model
Using float64 for MACECalculator, which is slower but more accurate. Recommended for geometry optimization.
[codecarbon INFO @ 17:31:13] offline tracker init
[codecarbon WARNING @ 17:31:13] Multiple instances of codecarbon are allowed to run at the same time.
Performing MD simulation... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╺━━━━━━━━━━━━━━━━━━━━━━━ 200/250 

I noticed this too when the CI tests ran. I can fix it, but I'm trying to see if there's a better way to approach this since this method seems really fragile at the moment.

[k-harris27]

Superseded by #444