Changelog

Changelog#

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased#

0.5.0 - 2026-04-09#

Added#

Changed#

  • all asyncmd.utils methods now require a MDEngine (class) to dispatch to the correct engine submodule methods

  • GmxEngine apply_constraints and generate_velocities methods: rename wdir argument to workdir to make it consistent with prepare and prepare_from_files (also add the workdir argument to the MDEngine ABC).

Fixed#

  • a rare condition that could result in the same TrajectoryFunctionWrapper being applied twice to the same Trajectory, which resulted in unnecessary computations and a ValuesAlreadyStoredError when trying to cache the calculated values twice.

0.4.1 - 2025-07-29#

Added#

  • Improved handling of h5py caches: Improved asyncmd.config.register_h5py_cache method and added asyncmd.config.deregister_h5py_cache method. It is now possible to use (multiple) read-only h5py cache sources to retrieve values from, even if no writeable cache source is registered. Also add an option to copy over all values from (read-only or writeable) previously registered h5py cache sources, when registering a new h5py cache source.

  • “py.typed” file to support type checking

0.4.0 - 2025-07-22#

Added#

  • Pylint configuration (in pyproject.toml), automatically run pylint as a github action

  • Module and file level docstrings with a short description for each module or file

  • Improve and expand documentation, include example notebooks into it, and setup deployment on read the docs

  • SlurmGmxEngine (and GmxEngine) now expose the mdrun_time_conversion_factor to enable users to control the expected time it takes to set up the environment inside the slurm job. Both engines also have improved consistency checks for mdp options when performing energy minimization.

  • MDEngine, GmxEngine, and SlurmGmxEngine: removed unused running property

  • Some tests for trajectory.propagate and gromacs.mdengine modules

Changed#

  • Renamed asyncmd.trajectory.construct_TP_from_plus_and_minus_traj_segments to asyncmd.trajectory.construct_tp_from_plus_and_minus_traj_segments. Also minimal change to the call signature, arguments to the TrajectoryConcatenator class are now not explicit arguments anymore but instead all kwargs will be passed to the TrajectoryConcatenator.concatenate method.

  • asyncmd.trajectory.propagate: Changes to the call signatures of TrajectoryPropagator classes, the arguments that were previously only passed trough to the concatenator class are now not explicit anymore and instead all additional kwargs passed to the TrajectoryPropagator.cut_and_concatenate and TrajectoryPropagator.propagate_and_concatenate methods will be directly passed to the concatenate method of the TrajectoryConcatenator.

  • Renamed asyncmd.config.set_default_trajectory_cache_type to asyncmd.config.set_trajectory_cache_type and remove option to set cache type individually for Trajectory objects and instead always use the centrally configured cache. These changes were made while drastically simplifying the trajectory function value caching code (mostly under the hood) to increase maintainability and extensibility.

  • TrajectoryFunctionWrapper and their interaction with Trajectories: One notable semi-under-the-hood change is that TrajectoryFunctionWrapper subclasses now need to implement the _get_values_for_trajectory method instead of the get_values_for_trajectory method. The call signature of the function has not changed and the change should in all cases be as easy as appending the _ to the function name in subclasses. This change was deemed necessary to avoid that users call the get_values_for_trajectory method expecting that the values would be cached (which they are not).

  • TrajectoryConcatenator: remove_double_frames is now an instance attribute instead of an argument to the concatenate and concatenate_async methods

Fixed#

  • refactor to reduce number of pylint messages

0.3.3 - 2025-05-06#

Added#

  • Add CHANGELOG.md file

  • SlurmProcess now supports arbitrary sbatch options via sbatch_options, this is also true for the SlurmGmxEngine and SlurmTrajectoryFunctionWrapper classes via the added keyword argument sbatch_options

  • Example IPython notebook on FrameExtractors

  • Some tests for slurm module

Fixed#

  • ConditionalTrajectoryPropagator, InPartsTrajectoryPropagator, and weighted ensemble example IPython notebooks now use the correct paths to input files

  • No more f-strings in logging statements

  • Add scipy to dependencies

0.3.2 - 2025-01-10#

Added#

  • First release on PyPi