0.5.0

• Particle mesh Ewald Coulomb solver
• Regression fix in empirical bond-order potentials (dimers were handle incorrectly) - regression introduced in 0.4.4

0.4.6

• More robust EAM (low and high density configurations)
• More robust version extraction for LAMMPD and standalone code (uses versioneer)

0.4.5

• Maintenance: Fixed problem with passing c_null_ptr through c_f_pointer when
  using XLF

0.4.4

• Bug fix: Occasional NaNs in bond-order potentials
• Maintenance: Fixed segfault on BlueGene

0.4.3

Minor fixes and preparation for PyPI release

v0.4.2

Support for new ASE calculator interface (thanks to James Kermode)

v0.4.1

• Regression fix: Charge extrapolation did not work because charges were
  overriden with initial charges at every step. This was introduced when
  changing to the new ASE Calculator class.

v0.4.0

• Python 3 compatibility.
• Python interface based on new ASE Calculator base class.
• Tight binding: Removed necessity for an 'elements.dat' file.
  Code auto-detects elements from files found in the database directory.
• Tight binding: Added support for d, sd and pd electronic configurations.
• LAMMPS interface automatically checks git fingerprint to ensure
  compatibility between pair_style and Atomistica library.
• Fixed proper stopping of timer when errors are raised or passed in some
  parts of the code.

v0.3.2

• Added general mechanism for object introspection from Python
• Exposed NOTB internals to Python (Hamiltonian, overlap and density matrices,
  eigenvalues and eigenvectors if LAPACK solver is used)
• Made NOTB per-bond analysis tools available from Python (added
  get_per_bond_property method)
• Some bug fixes to parts of the standalone code

v0.3.1

• Implemented charge extrapolation for SCC NOTB
• Fixed a couple of OpenMP related regression that broke compilation.
• Fixed a regression that lead to wrong unit conversion for Hubbard-Us.
• Fixed a regression that lead to OutputEnergy not being called.
• Coulomb solvers now have an energy_and_forces rather than a
  potential_and_field function.