CCpy: A coupled-cluster package written in Python
CCpy is a research-level Python package for performing non-relativistic electronic structure
calculations for molecular systems using methods based on the ground-state coupled-cluster
(CC) theory and its equation-of-motion (EOM) extension to electronic excited, attached,
and ionized states. As a design philosophy, CCpy favors simplicity over efficiency, and this
is reflected in the usage of computational routines that are transparent enough so that they
can be easily used, modified, and extended, while still maintaining reasonable efficiency.
To this end, CCpy employs a hybrid Python-Fortran programming approach made possible with the
f2py package, which allows one to compile Fortran code into shared object libraries
containing subroutines that are callable from Python and interoperable with Numpy arrays.
Contents:
- Installation
- The methods API: a calculation guide
- Molecules, references and composing calculations
- Ground states and energy corrections
- Excitation, attachment, ionization, ADC and STEOM
- Initial guesses, symmetry and choosing roots
- FNOs, active spaces and selected determinants
- Three different reductions
- Enable DF-MP2 FNOs at the start of a chain
- Automatic composite energy completion
- State-specific ADC-FNOs
- Multiple targets and root identities
- Protected orbitals and reference inspection
- Active and selected spaces after FNO preparation
- Custom FNO providers
- External determinant-space readers
- Cluster-in-molecule local correlation
- Left states, densities, response and gradients
- Multipole moments and harmonic IR intensities
- Analytic gradients and finite differences
- Geometry optimization, transition states and IRC validation
- Installation and units
- Minimize an electronic method directly
- Use a callback for custom calculations
- Optimization controls and result fields
- Hessian evaluation and cost
- Find a transition state
- Projected vibrational analysis
- Run both IRC branches with run_irc
- Optimize and validate the endpoints
- Electronic excited states and custom surfaces
- Runnable end-to-end examples
- Validation of the shipped example
- External documentation and version checks
- Computing vibrational frequencies
- Solver controls, diagnostics and custom calculations
- Complete methods API catalog
- Moller-Plesset perturbation theory
- Ground-state coupled cluster
- Equation-of-motion methods
- Algebraic-diagrammatic construction
- STEOM-CCSD
- Extended STEOM
- Noniterative and adaptive corrections
- Left ground and excited states
- Linear response
- State-universal multireference coupled cluster
- Similarity-transformed Hamiltonians
- References and materialization warnings
- Factorized Hamiltonian
- FNO policies, evaluations and preparation
- Experimental cluster-in-molecule local correlation
- Shared method protocol
- Preparation and registry functions
- A complete first calculation
- What is shared
- Native DF versus materialized equations
- Runnable recipes and documentation maintenance
- Frozen natural orbitals
- Density-fitted ADC(2)
- Spin-integrated STEOM-CCSD
- Legacy Driver interface
- Getting Started
- Computational Options
- CCD
- CCSD
- CCSD(T)
- CC3
- CCSDt
- CCSDT
- CC4
- CCSDTQ
- EOMCCSD
- EOM-CC3
- EOMCCSDT(a)*
- EOMCCSDt
- EOMCCSDT
- IP-EOMCCSD(2h-1p)
- IP-EOMCCSD(3h-2p)
- IP-EOMCCSDT(a)*
- EA-EOMCCSD(2p-1h)
- EA-EOMCCSD(3p-2h)
- EA-EOMCCSDT(a)*
- DEA-EOMCCSD(2p)
- DEA-EOMCCSD(3p-1h)
- DEA-EOMCCSD(4p-2h)
- SF-EOMCCSD
- CR-CC(2,3)
- CR-CC(2,4)
- CC(t;3)
- CIPSI-driven CC(P ;Q) aimed at converging CCSDT
- Adaptive CC(P ;Q) aimed at converging CCSDT
- CR-EOMCC(2,3) and \(\delta\)-CR-EOMCC(2,3)
- ec-CC-II
- ec-CC-II3
- ec-CC-II3,4
- Specific Walkthroughs
- Coupled-Cluster Options
- CC(P ;Q) Methodologies
- Externally Corrected Coupled-Cluster
- References
- CCpy Development Team