pyAutoMR

The method used by this program is quite similar to MOKIT. However, we try to do everything with PySCF and without Gaussian.

This program aims to do:

• HF guess strategy
• automatic guess for CASSCF/GVB/SUHF
• interface for post-MR

Installation

Pre-requisites

• MOKIT (no need to fully compile, only autopair is needed)
• PySCF
• mrh (optional, for MC-PDFT)
• ExSCF (optional, for SUHF)
• pyNOF (optional, for GVB)

Install

• git clone and add /path/to/pyAutoMR to your PYTHONPATH

Features

• UHF -> UNO (-> PM LMO -> assoc rot) (-> GVB) -> CASSCF
• UHF -> SUHF -> CASSCF
• RHF (-> vir MO projection -> PM LMO -> pairing) (-> GVB ) -> CASSCF
• CASSCF -> NEVPT2
• CASSCF -> MC-PDFT
• CASSCF(dry run) -> SA-CASSCF

UHF, RHF can be auto-detected.

Utilities

• guess for UHF/UKS
  • mix
  • fragment
  • from_fch
  • flipspin (by lmo or by site)
• internal stability for RHF/RKS, UHF/UKS, ROHF/ROKS
  • optimize wavefunction until stable
• dump CASCI coefficients
• dump (active) orbital compositions

Quick Start

from automr import guess, autocas

xyz = 'N 0.0 0.0 0.0; N  0.0 0.0 1.9' 
bas = 'cc-pvdz'

mf = guess.from_frag(xyz, bas, [[0],[1]], [0,0], [3,-3], cycle=50)
mf = guess.check_stab(mf)

mf2 = autocas.cas(mf)

Tutorials

• Tutorial: Symmetry-broken wavefunction
• Auto CASSCF: UHF case

TODO

• TDDFT NO -> CASSCF

Citation

Please cite pyAutoMR as

Shirong Wang, pyAutoMR, https://github.com/hebrewsnabla/pyAutoMR (accessed month day, year)

and cite every program called by pyAutoMR, such as PySCF, MOKIT, mrh, etc.

If you peform calculations involving GVB, please also cite the following two papers

DOI: 10.1021/acs.jctc.8b00854; DOI: 10.1021/acs.jpca.0c05216.