Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/muammar/centerfinder

Look for localized molecular orbitals (LMO) near to atoms based on the center of charges to use in the method of increments
https://github.com/muammar/centerfinder

Last synced: 20 days ago
JSON representation

Look for localized molecular orbitals (LMO) near to atoms based on the center of charges to use in the method of increments

Awesome Lists containing this project

README 

        
centerfinder

============



This python script intends to look for localized molecular orbitals (LMO) near

to atoms based on the center of charges and prepare a MOLPRO input file in

order to perform later incremental calculations as stated in _H. Stoll, Chem.

Phys. Lett., 1992, 19_. This is normally a very tedious work to be done by

hand, and it's there when centerfinder comes to help.



It reads your MOLPRO output file, and then it extracts coordinates and center

of charges to then print which LMO are in between which atoms in the molecule.

In your MOLPRO output file, a localization calculation must have been done

beforehand. Take into account that your calculation has to be done using the

`symmetry,nosym;` card in MOLPRO. On the other hand, you have to save your LMO

to the record `2103.2`. Below, an example of the `locali block` in MOLPRO to be

used to run this program:



```

{locali;

save,2103.2;

core,10;

group,28.1,31.1,-34.1;

group,11.1,-27.1,29.1,30.1;

}

```



*Note:* The script prepares two input files with the rotations needed to do

one- and two- body interactions, as well as the CASSCF and CCSD input. I will

generalize more this script in the coming time.



## Requirements



In order to run this script, you need to install in your system:



- Python 2.7 (I am porting this to work with Python >3)

- NumPy.

- SciPy.



You can use `pip` to install them or the package manager of your favorite Linux

distribution or Mac OS X.



## How to use it



### Installation and execution



Clone this git repository:



```bash

$ git clone https://github.com/muammar/centerfinder.git

```



Then you excute the program as follows:



```bash

$ python $PATH/centerfinder/centers.py $input

```

Where `$input` is the name of your MOLPRO output file. The program will ask you:



1. Number of CORE MOs used in your calculation for doing the localization.

- Number of Localized Molecular orbitals.



Then, for the CASSCF calculation section it will ask you:



3. Number of CLOSED MOs.

- Number of FROZEN MOs.

- Number of OCCUPIED MOs.

- The wavefunction in the format: `wf,#electrons,Sym,Multiplicity`.

- Atoms to be discarded by symmetry reasons.



*Note*: The default answers are `NO` so you can just press enter and the

program will continue.



Finally, a `molpro.in` and `molpro2.in` files will be created at the end of the

execution. Please *note* that in MOLPRO you can only run maximum of about `100`

calculations per input file.



### The input file created is too long, how can I split it?



If you are using linux, there is a very nice command called `split` (`man

split` for more information). So, if you have one input file created by

`centerfinder` that has more than 500 lines and you want to split it in

2 parts, you can do:



```bash

$ split -l 250 molpro2.in

```

and you will split `molpro2.in` in two files containing 250 lines each one.



Please note that you have to check that every MOLPRO block when splitting is

correctly done. I am thinking in adding another program that let you do this

instead of using `split`.



### How do I check that this input file was correctly generated?



I suggest you to make the one body calculations given in `molpro.in` and make

an analysis of the energies and check if the LMO are well assigned (using any

of the available molecular visualizers). If after checking your one body

results you find they are fine, so you are good to go. Actually, the two body

interactions are built from the one body ones. The script proceeds as follows:

a) it calculates the possible combinations; b) it asks you if you want to

discard atoms to not be taken into account in the two body interaction part; c)

and finally it writes `molpro2.in`.



### The onebody calculations are wrong, the LMO doesn't match when I plot them using a molecular visualizer. What can I do?.



There are 4 parameters you can change in the following lines:



* `atnearat=np.argwhere((distances > 2.3) & (distances < 2.7))`

* `lmonat=np.argwhere((dcoordcoc > 0.5) & (dcoordcoc < 2.0))`



If you do something wrong, you can revert the changes by doing:



```bash

$ git checkout $PATH/centerfinder/centers.py

```



### How to update it?



To update this program, just do a `git pull` in your `$PATH/centerfinder/`

folder.



## Feedback?



This is my first program, if it can be called like that, written in python. So,

I'm sure it is not optimal at all. However, you can send to me

[suggestions, and improvements](https://github.com/muammar/centerfinder/issues).

They are very welcome.