2.4.7

• Update doping/carrier concentration functions to be more accurate and robust (following logic discussed
  here: materialsproject/pymatgen#3879).
• Improve reverse-supercell-matrix determination for generate_supercell=False
• Refactor bulk_band_gap_path to bulk_band_gap_vr in DefectsParser/DefectParser,
• Update docstrings to reiterate that bulk supercell VBM is used as VBM reference point for the Fermi level
  by default, unless alternative bulk_band_gap_vr provided.

2.4.6

• Update Defect, DefectEntry and DefectThermodynamics properties/methods to be even more
  efficient with calculations of formation energies and concentrations. Gives ~10x speedup in Fermi
  solving and concentration calculations (e.g. from 2 hours to 12 minutes for 2D chempot vs temp CdTe grid
  in thermodynamics tutorial).
• Avoid unnecessary DeprecationWarnings from latest spglib release.

2.4.5

• Enforce shakenbreak>=2.3.4 requirement.

2.4.4

• Make oxidation state guessing more efficient, semi-significant speed up in generation/parsing for tough cases.
• Add bulk_site_concentration property to DefectEntry, giving the concentration of the corresponding lattice site of that defect in the pristine bulk.
• Minor updates to ensure compatibility with recent pymatgen and ASE releases.

2.4.3

• Remove spglib<=2.0.2 dependency (set to avoid unnecessary warnings), and update installation instructions accordingly.

2.4.2

• Allow cases where the calculated host material is unstable wrt competing phases (above the hull), by downshifting to the hull and warning the user about this.
• General updates to chemical potentials code; more robust (better error catches and messages, API key handling), more informative, cleaner outputs.
• Updates to match recent changes in pymatgen object types (e.g. #68)
• Minor efficiency & robustness updates

2.4.1

• Speed up eigenvalue parsing by using the faster doped site-matching functions rather than MakeDefectStructureInfo from pydefect
• Minor efficiency & robustness updates.
• Minor docs & tutorials updates
• Minor tests updates

JOSS Release

Release of doped to be archived on Zenodo, for the JOSS paper.

Comprises doped version 2.4.0, along with minor updates to documentation and warnings as suggested in the JOSS review.

2.4.0

• Electronic structure analysis by @adair-nicolson & @kavanase:
  • Adds DefectEntry.get_eigenvalue_analysis() method to plot single-particle eigenvalues and
    analyse orbital character and localisation; usage and examples shown on the
    docs Tips <https://doped.readthedocs.io/en/latest/Tips.html#eigenvalue-electronic-structure-analysis>__
    page and the advanced analysis tutorial <https://doped.readthedocs.io/en/latest/advanced_analysis_tutorial.html#eigenvalue-electronic-structure-analysis>__.
  • Projected eigenvalues can be parsed from vasprun.xml(.gz) files (preferred, as more accurate
    with 4 decimal places; c.f. 3 in PROCAR(.gz); more convenient and only ~5% slower) or PROCAR(.gz)
    files in calculation directories (both with significantly expedited parsing compared to pymatgen methods).
    Compatible with spin-polarised, unpolarised and SOC calculations. Comes with update by @kavanase to easyunfold
PROCAR.gz parsing.
• More efficient defect calculation parsing
• Add get_magnetization_from_vasprun and get_nelect_from_vasprun functions to doped.utils.parsing,
  as these attributes are not available from pymatgen.io.vasp.outputs.Vasprun.
• Improve testing efficiency

2.3.3

• General robustness updates:
  • Updated file parsing to avoid hidden files.
  • Sanity check in DefectsGenerator if input symmetry is P1.
  • Add NKRED to INCAR mismatch tests.
  • Re-parse config & spin degeneracies in concentration/symmetry functions if data not already present
    (if user is porting DefectEntrys from older doped versions or manually).
  • Avoid unnecessary DeprecationWarnings
• Updated docs and linting