0.6.1-1

Modify the Zenodo record to reflect Tsjerk A. Wassenaar as an author on
this project.

v0.6.1

Marieke Westendorp (6):
      Update README to mention the Zenodo record
      Update lockfiles to reflect the 0.6 version
      Add more thanks to the README
      Skip read_large_steps test
      Actually skip the read_large_steps test
      Bump version to 0.6.1

Semen Yesylevskyy (1):
      Adding write_frame_parts() method to efficiently write coordinates from arbitrary iterator

v0.6.0

This release brings a number of awaited changes.

- The project now has a Zenodo record and will have a citable doi.
- A `CONTRIBUTING.md` now describes the contribution guidelines for this
  project.
- Methods for skipping frames were introduced.
- The `glam` dependency has been removed. (This is a subtle breaking
  change for the library interface.)
- The _molly_ command line application now sits behind the `cli` feature
  flag.
- The `clap` dependency is now an optional dependency.
- The `chemfiles` development dependency was dropped. This was used for
  testing the validity of _molly_ output, but since `xdrfile` still
  fulfills this role it is easier to leave this CMake nightmare behind
  us.
- Some XTC writer bugs/papercuts were caught in usage and fixed.

I want to emphasize my appreciation for the contributors to this
release.

@ma3ke (Marieke Westendorp) (12):
      Properly format comments in README code example
      Remove glam dependency
      Make clap dependency an optional feature
      Add LICENSE file in accordance to this project's license
      Remove chemfiles dependency
      Remove trailing whitespace from doc comment
      Explain how seek_prev works and increase its initial step back
      Add @yesint to zenodo record
      Ignore MDAnalysis xtc_offsets.npz files
      Add CONTRIBUTING.md
      Add Aster Kovács to the Zenodo record
      Bump version to 0.6.0

@aster31 (Aster Kovács) (2):
      Set Frame default precision to 1000
      Python binding XTCReader.frame reference semantics

@mlund (Mikael Lund) (2):
      Add Zenodo metadata for DOI archival
      Use split_first strategy to slightly improve Buffer::pop performance

@yesint (Semen Yesylevskyy) (1):
      Add methods for skipping frames forward and backward

Introducing XTC writing support

Thanks to Mikael Lund, molly can now write XTC files as well.

0.5.0

Allow reading trajectories with a long simulation step.

Thanks to Lada for finally fixing this long-standing issue!

0.4.0

Fix Mask and Until AtomSelections.

Before, it was possible to supply selections that would leave some
values in the frame positions array unset.

Changes:
- Mask and Until AtomSelections now operate as described and expected.
- The default value with which uninitialized values in a frame positions
  array are filled during resizing is now NaN. This was done with the
  intention of making it more apparent when the final state of the
  positions array is not as expected. If NaNs are encountered that
  points to a problem in this library.

Breaking:
- The `molly::reader::read_compressed_positions` function now takes an
  additional `header_natoms` argument.

0.3.3

Improvement: An Err is returned when an unknown magic number is
encountered, rather than a panic. This is much better library design.
Thanks, Lada, for pointing that out!

Bug fix: Found and fixed a problem with the FrameSelection::Range
indexing. I added tests to replicate the issue and fixed the problem.

Python bindings: Updated pyo3 and numpy crates to 0.23.

Marieke Westendorp <ma3ke.cyber@gmail.com> (6):
      Emit an error when an invalid magic number is encountered
      Migrate to pyo3 0.21
      Fix range indexing
      Update pyo3 to 0.23
      Update packages
      Bump version to 0.3.3

Fix the output unit

Add support for the 2023 magic number