Skip to content

Code for the paper "Automatic Discovery of Chemical Reactions Using Imposed Activation"

License

Notifications You must be signed in to change notification settings

gomesgroup/iacta

 
 

Repository files navigation

Automatic Discovery of Chemical Reactions Using Imposed Activation

This repository contains code for the paper: Automatic Discovery of Chemical Reactions Using Imposed Activation.

Original authors: Cyrille Lavigne, Gabe Gomes, Robert Pollice, Alán Aspuru-Guzik

Prerequsites:

iacta needs python >= 3.7. Before being able to use the repository, you will need to install a few additional programs and packages. First, common packages that are required are numpy and pandas.

Second, you will need to install xtb version >= 6.3.0. Importantly, by default the xtb and crest binaries are assumed to be in $PATH. If they are not, the path to xtb has to be defined when running iacta with the appropriate option.

Third, you will need to install openbabel version >= 3.1.1 and the associated python bindings. The easiest is to do this is via conda:

conda install -c conda-forge openbabel

Finally, you also need to install yaml and pyaml, again easiest to do via conda:

conda install -c conda-forge pyyaml

Afterwards, simply clone the GitHub repository and installation is complete.

It is also recommended to define the environment variable LOCALSCRATCH before running. This variable should point to a directory to be used for temporary folders and files generated by iacta. Since iacta generates a lot of them, it is recommended to use a location with fast I/O.

How to run:

There are two ways to run iacta. The easiest is via yaml input files. There are a few example input files for the case studies shown in the paper in the subdirectories of the test-set and other-examples directories. These input files are used with rsearch-restart.py for iacta simulations as follows:

python3 /path/to/iacta/rsearch-restart.py /path/to/user.yaml -o path/to/output/

The second possibility is to use rsearch.py. There is one XYZ file in the example input files for the case studies shown in the paper. It can be found in the subdirectory test-set/Sn2MeI. The iacta search can be initiated as follows:

python3 /path/to/iacta/rsearch.py /path/to/initial.xyz 1 2 6.0 -o path/to/output/

For more detailed control over the input, please consult the help options -h for the rsearch.py and rsearch-restart.py scripts.

Analyze output

We also implemented convenient tools to analyze the results generated by iacta. They are used as follows:

python3 /path/to/iacta/read_reactions.py /path/to/output/

This script provides some summary of the reactions discovered and also generates the files parsed_reactions.csv and parsed_species.csv which provide the most important reactions and species discovered in a comma-separated table together with the path to the directories that contain the corresponding reaction trajectories.

Questions or Problems?

Please make a GitHub issue and be as clear and descriptive as possible. Feel free to reach out to: gabegomes[AT]cmu[DOT]edu and r[DOT]pollice[AT]rug[DOT]nl.

License

The contents of the iacta repository are provided under the MIT license.

DOI

About

Code for the paper "Automatic Discovery of Chemical Reactions Using Imposed Activation"

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 97.4%
  • Shell 2.6%