Skip to content

ChangeLog.md

Latest commit

 

History

History
119 lines (69 loc) · 4.59 KB

ChangeLog.md

File metadata and controls

119 lines (69 loc) · 4.59 KB
Raw

Ver. 1.1.3 (Aug. 19, 2026)

  • The average over degenerate modes is now taken for the scattering rate (i.e. the self-energy Gamma), not for the lifetime, in agreement with ANPHON. ANPHON averages the ph-ph damping damping3 before it is combined with the isotope/boundary contributions and converted to a lifetime (Conductivity::compute_kappa()), whereas Scattering averaged the lifetime after all scattering processes had been summed. Since tau is proportional to 1/Gamma, the arithmetic mean of the lifetimes differs from the inverse of the mean rate whenever the rates of the degenerate modes are not exactly identical. The averaging now happens in set_scattering_rate_phph(). (auto_kappa/io/scattering.py, tests/test_analyzer_degeneracy.py)

  • Lifetime averaging over degenerate modes no longer overwrites the averaged values. get_average_at_degenerate_point() had two defects. (1) The band loop ran over range(nb-1), so the highest band was never used as the starting index of a degenerate group and kept its zero-initialized value whenever it was not degenerate with the band below it, zeroing that mode's lifetime and its contribution to kappa. (2) A group was re-detected starting from every band it contains, and each pass overwrote the value written before, so the last mode of every group was left unaveraged: for omega = [0, 0, 0, 120, 120, 300] and tau = [3, 5, 10, 2, 8, 1] the result was [6, 7.5, 10, 5, 8, 1] instead of [6, 6, 6, 5, 5, 1], with the error growing with the degeneracy (e.g. triply degenerate optical modes at the Gamma point of cubic crystals). Each degenerate group is now detected once and the already-averaged modes are skipped, matching ANPHON's Conductivity::average_self_energy_at_degenerate_point(). (auto_kappa/io/analyzer.py, tests/test_analyzer_degeneracy.py)

Ver. 1.1.2 (Jul. 3, 2026)

  • Fixed a bug in BORNINFO generation where StructureMatcher.get_transformation() returned non-identity atom mappings for highly symmetric crystals (e.g. CdAs2, I4₁/amd), causing Born effective charge tensors to be assigned to the wrong atoms and reversing the sign of off-diagonal components for symmetry-related atom pairs. Replaced with direct nearest-neighbour matching in fractional coordinates.

Ver. 1.1.1 (2026)

  • Arrange the workflow related to --analyze_with_larger_sc, --scph, and --four options.

Ver. 1.1.0 (2026)

  • Excessive displacements in the random displacement method were limited to the smaller of 10% of the nearest-neighbor distance and 1.0 nm.

  • The bug in determining KMESH_INTERPOLATE has been fixed.

  • The default value of SCPH iterative was modified to 1,000, default value of ALAMODE.

Ver. 1.0.1 (Dec. 20, 2025)

New

  • Add --config and --optimize_klength.

  • Delete --vasp_parameters.

  • VASP parameters can be modified with a user configuration file read by --config option.

  • K-length can be automatically optimized using --optimize_klength option.

Ver. 1.0.0 (Nov. 22, 2025)

Fix

  • Run dfc2 command properly when the file name contains brackets

  • Modify the determination method of KMES_INTERPOLATE and KMESH_SCPH

Ver. 1.0.0 (Sept. 28, 2025)

New

  • Release the package

  • Add 'command_vasp_gam' for Gamma point calculation.

  • Modify the structure relaxation with EOS. (Set the maximum number of iteration and strain range.)

  • Read the previously-obtained relaxed structure for the volume relaxation (strict optimization).

  • Add 'calculate_forces' option. When this option is set to be "0", forces are not calculate but structure with displacement patterns are prepared.

  • Add 'four', 'command_anphon_ver2', and 'frac_kdensity_4ph' options.

Ver. 0.4.0 (April 24, 2023)

New

  • Set the primitive cell tolerance for ALAMODE calculations to match that of Spglib.

Fix

  • Change the minimum number of displacement patterns for LASSO to 10.

Ver. 0.3 (September 8, 2023)

New

  • New options --analyze_with_larger_supercell, --delta_max_natoms, and --max_loop_for_largesc were added. These options enable the calculation of harmonic properties with larger supercells when negative frequencies appear with the default size of supercell.

  • A log file ak.log is created in the working directory. Reading this log file, one can easily know the status of the calculation.

Ver. 0.2 (April 3, 2023)

New

  • A new option --file_structure was added. This option enables the automation calculation with an arbitrary structure.

Fix

  • Fix a few bugs

  • k-mesh densities are properly selected.

Modifications

  • Variable name: from dir_phdb to base_dir in ak_script.py