Releases: rowansci/rowan-python
Releases · rowansci/rowan-python
Release list
v3.1.12
New
- adds protein preparation workflow
- adds grouped solubility submission
Changes
- retrieve_workflows() now automatically splits large retrievals into API-safe batches of 100
- adds the public ProteinUUID semantic type alias and updated protein workflow examples to pass stored
proteins by UUID
Full Changelog: v3.1.11...v3.1.12
v3.1.11
New
- API keys have new attributes: budget and credits_used. When credits_used exceeds budget, workflows associated with that API key will be stopped. API keys now have a refresh() method. create_api_key() now has a budget param, and budgets can be updated with new update_api_key_budget() call.
Changes
- MKDocs now builds multiple pages of documentation instead of just one page.
Full Changelog: v3.1.10...v3.1.11
v3.1.10
New
- LogP workflow predicting the octanol/water partition coefficient from a neural network, Wildman-Crippen, or COSMO-RS
- Binder specification in protein MD, adding per-frame MM/GBSA scores and binder RMSD
- Cα backbone restraints and binder-proximal solvent saving in protein MD
Changes
- Protein MD forcefield validation no longer flags binder small molecules, which are parameterized from
SMILES rather than the protein forcefield
Bug fixes
- Fixed retrieval and download of proteins reachable only through a workflow shared with you
- Fixed workflow result types so editors and type checkers see each workflow's own result class instead of the base class
- Corrected examples and skill references: PDB code argument order, missing ligand_smiles, and download_pdb_file filename handling
Breaking
- validate_protein_forcefield() renames exclude_residue_names to exclude_residues, which now accepts residue names or 0-based non-polymer indices
Full Changelog: v3.1.9...v3.1.10
v3.1.9
New
- rowan.NMR_SUPPORTED_SOLVENTS exported at the top level, listing the solvents NMR
prediction is parameterized for
Changes
- NMR prediction is out of private beta
- NMR conformer ranking now reweights the conformer ensemble with a solvent-aware AIMNet2
single point (CPCM-X where supported, otherwise ALPB) instead of ranking gas-phase,
improving predicted-shift accuracy in solvent - submit_nmr_workflow now validates solvent against the supported set and raises at
submission time instead of failing during compute
Full Changelog: v3.1.8...v3.1.9
v3.1.8
New
- analogue_names on submit_analogue_docking_workflow for naming analogues; each docked pose's Molecule.name is set so best_poses can be re-keyed by name instead of SMILES
- Molecule.name getter/setter exposing the underlying stjames name field
- settings property on RBFE perturbation results, plus TMDRBFESettings exported at the top level, for resubmitting on a prior run's exact settings
- write_ligand_results_csv() on RBFE perturbation results, exporting per-ligand ΔG, ΔG error, and SMILES
- get_project() for getting or creating a project by exact name, with creation opt-in
Full Changelog: v3.1.7...v3.1.8
v3.1.7
New
- Context-local api_credentials() manager for safely using per-request API keys and project
UUIDs in concurrent applications - Claude Code and Codex plugin installation instructions
Changes
- Agent skill reorganized into lightweight shared guidance with dedicated MCP and Python SDK references
Bug fixes
- upload_protein() now surfaces errors from both PDB conversion and protein creation instead of attempting to parse unsuccessful responses
Full Changelog: v3.1.6...v3.1.7
v3.1.6
New
- Covalent inhibitor scan workflow with scan-point geometries and energies
- gnina support for noncovalent and covalent docking
- GSM and NEB support for double-ended transition state search workflows
- Optional endpoint optimization and endpoint result access for IRC workflows
- Actual forward and backward step sizes exposed in IRC results
- MM/GBSA scores added to docking, analogue docking, and protein cofolding results
Changes
- Docking now accepts VinaSettings or GninaSettings through docking_settings
- Legacy docking arguments such as executable, scoring_function, exhaustiveness, and
max_poses are deprecated - Double-ended transition state searches now use unified string-method settings with
geodesic interpolation by default - Minimum stjames version updated from 0.0.217 to 0.0.228
Breaking
- Double-ended transition state search settings changed — FSMInterpolation and FSMOptimizationCoordinates were removed; use Interpolation with StringMethodSettings or NEBSettings
- Legacy FSMSettings tuning fields such as node counts, optimizer iterations, and
displacement are no longer available
Full Changelog: v3.1.5...v3.1.6
v3.1.5
- binding_affinity workflow for predicting protein-ligand binding affinity
- generate_intermediate_ligands setting on the RBFE graph workflow — splits oversized
perturbation edges by generating synthetic intermediate ligands that recombine R-groups
already present in the two parent ligands - New periodic/crystalline properties on Molecule: symmetry, xrd_peaks, band_structure,
band_gap, density_of_states, elastic_tensor - Hydration site results (pamd) on the pose analysis MD workflow
Full Changelog: v3.1.4...v3.1.5
v3.1.4
New
- perturb() and resubmit_with_perturbations() methods on molecules for geometry
perturbation and workflow resubmission - Protein MD result fields (pamd and related)
- Timeout errors now surfaced from the API
- MSA workflow types re-exported from rowan — no more import stjames needed for cofolding
workflows
Changes
- HTTP error messages from the server are now passed through directly
- Folder and workflow list calls scoped to projects
- Improved folder interaction API
- Unique ligand name validation
Bug fixes
- Fixed docking workflow submissions failing against updated server
- Fixed blocked workflow error handling — now raises a clear error instead of hanging
- Fixed basic calculation tasks field not being populated correctly
- Fixed create_protein_from_pdb_id argument order in docking examples
Breaking
- create_protein_from_pdb_id argument order changed — code is now first, name is now
optional (previously name, code both required)
Full Changelog: v3.1.3...v3.1.4