hpc-pyscf
PySCF Python quantum chemistry library. Supports HF, DFT, CC, CI, and multi-reference methods. Use for developing custom quantum chemistry workflows and research applications.
SKILL.mdskills/hpc-pyscf/SKILL.md
GitHubHPC-PySCF Skill
PySCF is an open-source Python quantum chemistry software package supporting DFT, HF, and post-HF methods (MP2, CCSD, CASSCF, CASPT2, FCI, etc.).
Quick Start
| Step | Task | Reference |
|---|---|---|
| 1 | Molecule definition (atoms, basis, symmetry) | 01-molecule-definition |
| 2 | Select calculation method (HF/DFT/MP2/CCSD) | 02-calculation-methods |
| 3 | Excited states (TD-DFT, CASSCF, EOM-CC) | 03-excited-states |
| 4 | Advanced features (solvation, periodic, relativistic) | 04-advanced-features |
| - | Diagnose and fix runtime errors | error-recovery |
Skill Map
User Requirements
├─ Molecule Definition → 01-molecule-definition
├─ Energy Calculations
│ ├─ HF Energy → RHF/UHF/ROHF
│ ├─ DFT Energy → RKS/UKS
│ └─ Post-HF → MP2/CCSD/CCSD(T)/CI
├─ Structure Optimization
│ ├─ Geometry Optimization → PySCF + PyBerny
│ └─ Transition State → PySCF + ASE
├─ Electronic Structure
│ ├─ Single Reference → HF/DFT/CC
│ ├─ Multi-Reference → CASSCF/CASPT2/FCI
│ └─ Open-Shell → UHF/UKS/ROHF
├─ Spectroscopy
│ ├─ Vibrational Spectrum → Hessian
│ ├─ Electronic Spectrum → TD-DFT/EOM-CCSD
│ └─ NMR → GIAO
└─ Advanced Applications
├─ Solvation → PCM/COSMO
├─ Periodic Systems → k-point calculations
├─ Relativistic → X2C/DKH
├─ FCI Extensions → SHCI/DMRG
└─ Custom Methods → Flexible Python API
Key Decision Points
| Question | Guide | Summary |
|---|---|---|
| Calculation method? | 02-calculation-methods |
Based on precision requirements |
| Basis set? | 01-molecule-definition |
STO-3G/6-31G/cc-pVDZ/cc-pVTZ — balance precision and efficiency |
| Spin treatment? | 02-calculation-methods |
RHF/UHF/ROHF for closed-shell/open-shell/half-open-shell |
| Correlation method? | 02-calculation-methods |
MP2/CCSD/CCSD(T) for dynamic correlation |
| Multi-reference? | 03-excited-states |
CASSCF/CASPT2 for static correlation/excited states |
| Solvent model? | 04-advanced-features |
PCM/COSMO/SMD for solution environment |
Guardrails
Must Check
- Molecular coordinates are correct (units: Angstrom)
- Basis set and element are matched
- Spin multiplicity is correct
- Charge settings are correct
- Convergence criteria are reasonable
Common Errors
- SCF not converged: Use DIIS, level shift
- Insufficient memory: Reduce basis set or use density fitting
- Symmetry error: Disable symmetry or check structure
- Integration error: Increase integration precision
Outputs
Always report:
- Calculation method and basis set
- SCF convergence status
- Total energy and geometry (if optimized)
- Key output and data structures
- Computation time and memory usage
Assets
When to include: When the skill needs files that will be used in the final output.
Use cases: Templates, boilerplate code, batch scripts that get copied or modified.
| File | Use Case |
|---|---|
assets/templates/sp_hf.py |
HF single point energy calculation |
assets/templates/sp_dft.py |
DFT single point energy calculation |
assets/templates/mp2_ccsd.py |
Post-HF correlation calculations |
assets/templates/casscf.py |
CASSCF multi-reference calculation |
assets/templates/tddft.py |
TD-DFT excited state calculation |
assets/templates/pyscf_slurm.sh |
SLURM submission script |
Reference Summary
All references are used in this skill:
| Document | Topic |
|---|---|
| 01-molecule-definition | Atoms, basis set, symmetry, charge/spin |
| 02-calculation-methods | HF, DFT, MP2, CCSD, CCSD(T), CI methods |
| 03-excited-states | TD-DFT, CASSCF, CASPT2, EOM-CCSD |
| 04-advanced-features | Solvation, periodic, relativistic, FCI extensions |
| error-recovery | SCF convergence, memory, symmetry, integration errors |
Error Recovery
Consult references/error-recovery.md for structured diagnosis of:
- SCF convergence failures
- Memory allocation errors
- Symmetry and integration errors
- Input and API errors
Minimal Example
from pyscf import gto, scf
mol = gto.M(atom='H 0 0 0; H 0 0 0.74', basis='cc-pvdz')
mf = scf.RHF(mol)
energy = mf.kernel()
print(f"Energy: {energy} Ha")