Dataset request: feature_comparison_NJP2023

[jvita]

Contribute content

Contributor/requester

Contact information about the person contributing/requesting the data. Used for communication purposes. ​

Name: Josh Vita Email: vita1@llnl.gov

Dataset

Any information necessary to help the ColabFit find and access the data, and to correctly cite relevant material. The "name" and "description" will be used when publishing to the ColabFit exchange, and should be human-readable. Author list should include full first names, unless the author is normally attributed by initials. Links should include relevant publications and online location of dataset, if available.

Name: feature_comparison_NJP2023

Authors: Ting Han, Jie Li, Liping Liu, Fengyu Li, Lin-Wang Wang

Links:

• Publication link: https://iopscience.iop.org/article/10.1088/1367-2630/acf2bb
• Dataset link: contact authors

Description:

Calculations

Details regarding how the data was computed in order to improve reproducibility. Provide as much information as possible. Input files are highly encouraged. Additional details might include functional, basis set, energy cutoff, k-point grid, reference energy, etc. ​

Method: DFT Software: Unknown Additional details: Unknown Files: Unknown

Included properties

See the current list of ColabFit property definitions. If you believe your data does not match one of the existing definitions, then you must submit a new property definition following the template provided in the examples folder.

Name	Units	Notes

Configurations

Basic information explaining the types of configurations in the dataset, and how they are organized.
Elements should be listed by chemical symbol

Elements: C, S Number of configurations: 4000? Storage format: Unknown

Naming convention

If your configurations have names, please describe where their names can be found (e.g., as a field in an ASE.Atoms.info dictionary). ​

​

Configuration sets

Configuration sets are used to define a conceptual grouping over a collection of atomic configurations. Configuration sets are constructed via regex filtering on specified keys. ​

Paper contents suggest that configuration sets have been defined. However, this can't be verified without access to the data.

Key	Regex	Description

Configuration labels

Configuration labels can be attached to your data to improve interpretability. This is done via regex matching on specified keys. ​

Paper contents suggest that configuration labels may exist. However, this can't be verified without access to the data.

Key	Regex	Label

Distribution License

The license under which the content will be distributed (e.g. Creative Commons Zero)

Unknown

[jvita]

More notes that I took:

• Does Github have an option for "Open an issue using this MD file as a template"?
• We should be able to add tags like "missing information" or "contact author"
• We should track which institutions are contributing data. Maybe easier once we have something like "profiles" on ColabFit

[gpwolfe]

contacted corresponding author

[gpwolfe]

data link appears to be repaired. base link: http://www.pwmat.com/mlff DFT dataset(s) link: https://github.com/LonxunQuantum/PWMLFF_library Parser may be adapted from https://github.com/LonxunQuantum/PWMLFF_library/blob/main/utils/MOVEMENT2xyz.py

A number of the splits of the dataset (split by element type and by number of images in the trajectory) are only available via download through baidu netdisk client. Others are prepared for ingest