POMFinder 1.0.0

POMFinder is a tree-based supervised learning algorithm which can predict the polyoxometalate cluster from a Pair Distribution Function.

ChemRxiv | [Paper]

POMFinder

Welcome to POMFinder! This is a simple machine learning tool for structure characterisation of polyoxometalate clusters using total scattering Pair Distribution Function (PDF) analysis.
Simply provide a PDF and the model will output best best structural models from it structure catalog which contains 443 polyoxometalate clusters.

1. App
2. Install
3. Usage
4. Authors
5. Cite
6. License

App

You can use POMFinder on a single PDF on the following HuggingFace App: https://huggingface.co/spaces/AndySAnker/POMFinder

Install

To install POMFindeer you will need to have Python or Anaconda installed. We recommend running POMFinder on Python version 3.7. If you have installed Anaconda you can create a new environment and activate it.

conda create --name POMFinder_env python=3.7
conda activate POMFinder_env

Now you are ready to install what you actually come for! Currently POMFinder is not available through PyPI or conda so the package needs to be downloaded manually Run the following command to install the POMFindeer package.

pip install .
or
python setup.py install

To verify that POMFinder has been installed properly, try calling the help argument.

POMFinder --help

>>> usage: POMFinder [-h] -d DATA [-n NYQUIST] [-i QMIN] [-a QMAX] [-m QDAMP] [-f FILE_NAME]       
>>> 
>>> This is a package which takes a directory of PDF files 
>>> or a specific PDF file. It then determines the best structural 
>>> candidates based on a polyoxometalate catalog. Results can
>>> be fitted to the PDF. 

This should output a list of possible arguments for running POMFinder and indicate that it could find the package!

Usage

Now that POMFinder is installed and ready to use, let's discuss the possible arguments.

Arg	Description	Default
	Required argument
-h or --help	Prints help message.
-n or --nyquist	Is the data nyquist sampled. bool	-n True
-i or --Qmin	Qmin value of the experimental PDF. float	-i 0.7
-a or --Qmax	Qmax value of the experimental PDF. float	-a 30
-m or --Qdamp	Qdamp value of the experimental PDF. float	-m 0.04
-f or --file_name	Name of the output file. str	-o ''
-d or --data	A directory of PDFs or a specific PDF file. str	-d 5

For example

POMFinder --data "Experimental_Data/DanMAX_AlphaKeggin.gr" --nyquist "no" --Qmin 0.7 --Qmax 20 --Qdamp 0.02

>>> The 1st guess from the model is:  icsd_427457_1_0.9rscale.xyz
>>> The 2nd guess from the model is:  icsd_427379_0_0.9rscale.xyz
>>> The 3rd guess from the model is:  icsd_281447_0_1.0rscale.xyz
>>> The 4th guess from the model is:  icsd_423775_0_0.9rscale.xyz
>>> The 5th guess from the model is:  icsd_172542_0_1.1rscale.xyz

OBS: As demonstrated in section E of the supplementary information, the Qmin, Qmax, and Qdamp values have only a minor impact on the prediction and can thus be omitted.

Authors

Andy S. Anker¹
Emil T. S. Kjær¹
Kirsten M. Ø. Jensen¹

¹ Department of Chemistry and Nano-Science Center, University of Copenhagen, 2100 Copenhagen Ø, Denmark.

Should there be any question, desired improvement or bugs please contact us on GitHub or through email: [e-mail].

Cite

If you use our code or our results, please consider citing our paper. Thanks in advance!

https://chemrxiv.org/engage/chemrxiv/article-details/64e5fef7dd1a73847f5951b9

License

This project is licensed under the Apache License Version 2.0, January 2004 - see the LICENSE file for details.