Skip to content

Bugfixes and some new features#505

Open
RagnarB83 wants to merge 40 commits into
masterfrom
NEW
Open

Bugfixes and some new features#505
RagnarB83 wants to merge 40 commits into
masterfrom
NEW

Conversation

@RagnarB83

@RagnarB83 RagnarB83 commented Jun 24, 2026

Copy link
Copy Markdown
Owner

Bugfixes and some small new features

  • Photoelectron: CCSD(T) option, OODFT moread and outputfiles fix
  • DLFIND interface: bugfix for TSOpt
  • create_ML_training_data: more data files written, some restructuring (Madhur)
  • PDBx/mmCIF file support for Fragment and OpenMMTHeory
  • flexible_align_pdb: fix
  • fixes for elements dictionaries: lanthanides and actinides, also dummy atom support for certain FFs
  • pySCF: skala use fix
  • kspies: minor fixes
  • crest interface: support for socket
  • CFOUR: bugfix for NEB runs (Madhur)

RagnarB83 and others added 30 commits May 11, 2026 09:41
…see issue 493

- also fix for print_if_level bug
- also fix for jobtype None bug
…FT unnecessarily) and using UHF instead (works for both HF and DFT)
Singlepoint_fragments: reaction energy unit
…re present. Adding M as element, will use mass and charge of 0. This should allow the use of e.g. water models with virtual sites.

- OpenMMTheory: avoids defining mm_elements if virtual site encountered.
For more than one theory objects that energies, gradients, and train_mace_data.xyz files would be created separately for each theory and also the delta one as before.
a small feature to the create_ml code
sorry, my previous feature added had a bug. For single theory (pure model training), the atomic energy and mace formatted file was not calculated/created.
RagnarB83 and others added 10 commits June 24, 2026 17:51
- element_dict_atname: dummy element
- pyscftheory: CC with GHF enabled
- pyscftheory x2c: now enabling full soc-x2c if scf_type is GHF or GKS
- pyscftheory: ROHF/ROKS, can avoid pop-analysis if requested
- QMMMTheory: preparing for polembed_drude option. Not ready yet.
- added exception error output for mdtraj fails
CFOUR theory works well for relaxed surface scans, but fails for NEB, because important files were not copied to image_* folders.
- NumFreq:
  - IR Boolean option (default True), allows avoiding IR intensities if needed
   - linearity check made consistent for both projection function and detect_linear function. caused problems.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants