[WIP] Better input set generator and sets for io.lammps #4368
Conversation
…LammpsInputFile objects as input for the template to make use of all the other io.lammps functions in pmg
update - 12/3
…rd lammps) files with the input set that are prepared independent of the input file
Merge updates - 1/6/2025
Added a new (and probs better?) implementation for generating lammps input sets, with a more expressive MD template. Everything is now based on the template to ensure the atomate2 workflows being written work without the user having to specify too much. Inputs and states have to be enumed.
…LammpsInputFile objects as input for the template to make use of all the other io.lammps functions in pmg
…rd lammps) files with the input set that are prepared independent of the input file
Added a new (and probs better?) implementation for generating lammps input sets, with a more expressive MD template. Everything is now based on the template to ensure the atomate2 workflows being written work without the user having to specify too much. Inputs and states have to be enumed.
…class for storing and validating user defined settings
Changes to be committed: Reworked the way the input set generator generates the input file to be less cumbersome and use fewer if-else clauses, this new approach goes back to the original generator in order to be a lot more flexible (esp for custom lammps jobs). A LammpsSettings object is now present in order to validate the inputs specified by the user prior to actually passing them to lammps. This approach works out for the makers implemented currently (NVT/NPT/NVE/Minimization).
Redid how input sets are validated, in particular how "general" settings can also be stored as attrs of LammpsSettings, and only validate keys from _BASE_LAMMPS_SETTINGS. This way, allows for very flexible inputs from user to also be present in the final taskdoc.inputs. Also used PathLike instead of Path objects.
vir-k01
changed the title
src/pymatgen/core/periodic_table.py
Outdated
| @@ -776,6 +777,19 @@ def is_rare_earth(self) -> bool: | |||
| """ | |||
| return self.is_lanthanoid or self.is_actinoid or self.symbol in {"Sc", "Y"} | |||
|
|
|||
| @property | |||
| @deprecated( | |||
There was a problem hiding this comment.
I guess reverting changes to core.periodic_table and io.exciting is not intended?
If so you could revert those changes with (warning, this would drop all changes to that file, make sure this is what you need) git checkout upstream/master -- file_to_revert
There was a problem hiding this comment.
Just behind the master branch, synced up now
There was a problem hiding this comment.
Hi Vir,
Thanks for opening this PR, it looks very good and much better than what we had already done. I have left a few comments - those are only suggestions, I don't have a strong opinion on any of them.
I think indeed including support for molecules would be beneficial. I have added a few comments/suggestions going in that direction, let me know what you think.
The force field description is definitely a challenge with so many different ways to specify it. I'm not sure checking/validating what the user specifies is absolutely necessary at this stage - after all the code is also not checking the validity of other user-specified settings. I would leave it to the user to know how to specify the force field they want to use. But if others have ideas on how to implement that efficiently, I'm definitely not against it either :-)
Looking forward to having this integrated in pymatgen!
Best,
Guillaume
src/pymatgen/io/lammps/generators.py
Outdated
|
|
||
| MODULE_DIR = os.path.dirname(os.path.abspath(__file__)) | ||
| TEMPLATE_DIR = f"{MODULE_DIR}/templates" | ||
| _BASE_LAMMPS_SETTINGS = { |
There was a problem hiding this comment.
To address the different default settings for solids or molecules, maybe this could be a nested dictionary like
_COMMON_SETTINGS = {} # Add any default that would be common to any type of data.
_BASE_LAMMPS_SETTINGS = {
"structure": _COMMON_SETTINGS | {"atom_style": "atomic", "boundary": ("p", "p", "p")},
"molecule": _COMMON_SETTINGS | {"atom_style": "full", "boundary": ("f", "f", "f")},
"general": _COMMON_SETTINGS,
}
that would then be used according to the type of the object passed to get_input_set.
There was a problem hiding this comment.
Just as a small comment, I wanted to point out that in principle the Lattice has a pbc attribute:
pymatgen/src/pymatgen/core/lattice.py
Line 48 in 39308ec
so in principle even mixed use cases could be handled directly with a
Structure object.Maybe it would be easier to have the
boundary and atom_style settings be part of the data as produced by LammpsData.from_structure(from_molecule) instead? Would it make sense?
There was a problem hiding this comment.
Thank you so much for the comments! I've made some of the changes, and let me try what @gpetretto mentioned for the molecules
| input_file = LammpsInputFile.from_str(filtered_input_str, keep_stages=self.keep_stages) | ||
|
|
||
| return LammpsInputSet( | ||
| inputfile=input_file, data=data, calc_type=self.calc_type, additional_data=write_data, **kwargs |
There was a problem hiding this comment.
It would be nice to pass the argument template_file=... so that a trace of the initial template is kept somehow. This could probably be achieved in different ways though, not sure what would be best.
src/pymatgen/io/lammps/generators.py
Outdated
|
|
||
| def get_input_set( | ||
| self, | ||
| data: Structure | LammpsData | CombinedData, |
There was a problem hiding this comment.
This would have to accept Molecule as well.
src/pymatgen/io/lammps/generators.py
Outdated
| species = "" | ||
| if isinstance(data, Structure): | ||
| species = " ".join({s.symbol for s in data.species}) | ||
| data = LammpsData.from_structure(data, atom_style=atom_style) |
There was a problem hiding this comment.
For Molecules, I think the LammpsData.from_molecule classmethod should be implemented.
Moved the FF warning to inside the get_input_set function, better position cause if the FF hasn't been specified by then it's most likely an error. Also used .get() methods for the conditionals in the get_input_set function for when the user is using the custom maker
…ized json where atoms and velocities are not resolved into dataframes before checking condition. Unsure why that is the case, commenting out for now
…data from system.data
|
@vir-k01 any news about this PR ? |
|
@davidwaroquiers apologies for slacking off on this. After a bit of experimentation, I've added a |
Summary
This PR is part of an effort to include workflows for running LAMMPS via atomate2. A concurrent PR will be opened in atomate2 that has the flows, while this PR aims to update the input set generators to match the inputs required there.
These changes are based on initial work here: https://github.com/Matgenix/atomate2-lammps done by @ml-evs and @gbrunin. Also tagging in @esoteric-ephemera who helped structure the code in this PR, and @davidwaroquiers.
Major changes:
Marked this as a WIP as there might be further updates here dependent on the updates made in the concurrent atomate2 PR.
TODO: