A few small updates to the drivers

What's Changed

• Remove net force for non-conservative forces in metatomic driver by @frostedoyster in #473
• Metatomic variants by @ceriottm in #474
• Feat/mace by @EliaStocco in #465
• Added Example of UQ tracking by @ceriottm in #475
• Update version number for 3.1.7 release by @ceriottm in #476

Full Changelog: v3.1.6...v3.1.7

A few new useful features, including the possibility of evaluating multiple replicas at once with a batched evaluator, and of having PES files into a custom location

What's Changed

• Update elphmod example by @janberges in #461
• feat/morse by @EliaStocco in #464
• Psiflow PES by @armaet in #466
• the trimsim tool now works with ase formatted output files too by @venkatkapil24 in #468
• Monitor energy conservation in constrained barostats by @ceriottm in #455
• Batch evaluation of replicas using FFDirect by @ceriottm in #471
• Revamped PES organization by @EliaStocco in #472

Full Changelog: v3.1.5.1...v3.1.6

Very minor bugfix release, with a new pes.

What's Changed

• Fixing examples in client/aims by @mahrossi in #460
• Feat/spherical lj by @EliaStocco in #459
• Fix a deprecated numpy dtype that breaks on some versions of numpy by @ceriottm in #462

Full Changelog: v3.1.5...v3.1.5.1

Bugfix release, with a couple of new features, scripting and driver improvements

What's Changed

• add missing files to lammps example by @litman90 in #450
• Feat/mace by @EliaStocco in #432
• Add an interface to metatomic ML models by @Luthaf in #448
• Fix "specialty" thermostats by @ceriottm in #449
• Added (and cleaned) debug messages around FF evaluation and depend objects by @ceriottm in #451
• Fixed volume barostat by @armaet in #441
• Scripting improvements by @ceriottm in #453
• FFCommittee of MACE models with FFDirect by @EliaStocco in #456
• Fixed imports for drivers by @ceriottm in #458

New Contributors

• @armaet made their first contribution in #441

Full Changelog: v3.1.4...v3.1.5

Bugfix release for metatensor driver, including also an experimental scripting helper that allows defining a custom stepper for a motion class.

What's Changed

• Fix virial in metatensor driver by @frostedoyster in #445
• Fix vesin device by @frostedoyster in #444
• Generic stepper by @ceriottm in #446

Full Changelog: v3.1.3...v3.1.4

What's Changed

• Bump pytorch-lightning from 2.0.8 to 2.4.0 in /demos/ensemble-deltamu by @dependabot in #433
• Allow non-conservative forces and stresses in metatensor driver by @frostedoyster in #435
• Deprecate PET and librascal drivers by @Luthaf in #440
• Remove the dependency on ase calculator for the metatensor driver by @Luthaf in #438

Full Changelog: v3.1.2...v3.1.3

A quick-fix release for a regression bug

What's Changed

• Bugfix on return shape of energy by @ceriottm in #431

Full Changelog: v3.1.1...v3.1.2

This is primarily a bugfix release, addressing some recent and some long-standing bugs, as well as improving a few corners of the documentation

What's Changed

• Adding information to documentation by @mahrossi in #418
• Added git information to the wheel by @EliaStocco in #408
• Update black formatting version by @EliaStocco in #428
• Fixed ASE cell order in the scripting helpers by @ceriottm in #430
• Bug in FFDirect by @EliaStocco in #429
• Removing drift from conserved quantity. by @mahrossi in #423
• Implement a native metatensor driver by @frostedoyster in #424
• Bugfix/get temp by @venkatkapil24 in #417
• Fixing bugs for EDA by @EliaStocco in #404

Full Changelog: v3.1...v3.1.1

This is a point release that introduces several new features and a few breaking changes to the input files.
Main highlights are a ffdirect forcefield class that sidesteps the need for an external driver for serial evaluation of Pythonic PES, consolidation of the python PES and a first minimalistic API to call i-PI from a python script.

What's Changed

• Clean up the kinetic temperature calculator by @ceriottm in #368
• A utility function to install the driver by @ceriottm in #369
• Bump actions/download-artifact from 3 to 4.1.7 in /.github/workflows by @dependabot in #370
• A "t" too much in a lbfgs if clause by @mahrossi in #373
• Just cleaning up a bit the docstrings and the adaptive FD logic for the heat capacity estimators by @ceriottm in #375
• update documentation by @V-Alizade in #377
• Better dev interface to force.copy (now renamed force.clone) by @ceriottm in #380
• Use "atomic units" for time and pressure in docs by @ceriottm in #381
• Fixed smotion mode setting for metad by @ceriottm in #378
• Added a utils library for ACF and wrapped the postprocessing tool around it by @ceriottm in #382
• bugfix instools by @litman90 in #383
• add xtb driver to python drivers by @V-Alizade in #376
• Moved GLE spectrum deconvolution code to tools by @ceriottm in #384
• Add SO3LR calculator by @sirmarcel in #388
• Prevent new xtb module from crashing simulations by @frostedoyster in #389
• Changing the default behaviour of the seed of the random number generator. by @mahrossi in #393
• Making the dummy driver pass a json-formatted string for extras field by @mahrossi in #395
• Implement fixatoms_dof tag and variables by @litman90 in #394
• A "direct" FF using the same PES interfaces as the python driver by @ceriottm in #390
• Consolidate tutorials in the docs and restyle by @BarakHirshberg in #392
• Fix multiple random number outputs by @mahrossi in #405
• Fixing elphmod by @EliaStocco in #403
• Elphmod by @EliaStocco in #406
• A rudimentary scripting interface for i-PI by @ceriottm in #397
• update xtb client by @V-Alizade in #407
• Fix traj parsing by @ceriottm in #409
• Functionalities to incorporate rotational averaging in i-PI by @ceriottm in #372
• Improve the PLUMED interface by @ceriottm in #411
• Update docs by @litman90 in #402
• Release/v3.1 by @ceriottm in #413
• update python version in setup.cfg by @litman90 in #415
• Fix broken YML by @ceriottm in #416

New Contributors

• @dependabot made their first contribution in #370
• @V-Alizade made their first contribution in #377
• @frostedoyster made their first contribution in #389

Full Changelog: v3.0...v3.1

This is the 3.0 stable release of i-PI.
The new reference paper for the code is Litman et al., J. Chem. Phys. 161, 062504 (2024)

The main highlights of this new release are:

• a deep re-design of the depend mechanism and of the socket communication, reducing the overhead of i-PI for several
  common workflows by a factor of 2 or more
• simplified adding python-side PES to the i-PI driver
• several new simulation features, as documented in the accompanying paper
• improved documentation and examples, to make features easier to find and to use
• a million papercut bugfixes

What's Changed

• Faster ipi fix by @litman90 in #294
• Updated README to reflect that the pypi package has been renamed by @ceriottm in #293
• Fix for 2-temperature thermostat needed by @mahrossi in #287
• Automatic import of PES drivers by @ceriottm in #282
• Add ams example by @mhellstr in #297
• Demo for 2D-IR-Raman spectrum calculation using equilibrium-nonequilibrium RPMD by @tbegusic in #279
• Fix bug and update sGDML force field function by @sauceda in #197
• lint fix consequence of stalled PR 197 by @litman90 in #298
• Better documentation of the handling of the lattice parameters by @ceriottm in #301
• Added regtests for open-path simulations by @ceriottm in #303
• Bosonic PIMD in quadratic time by @yotamfe in #258
• Examples cleanup part 4 by @ceriottm in #283
• Constrained centroid MD by @ceriottm in #304
• Update to latest black formatting by @ceriottm in #306
• Updating bug in comment line of phonons.py by @mahrossi in #307
• Metatensor driver by @ceriottm in #305
• Cleaning up docs by @ceriottm in #302
• Fixed PET and rascal drivers by @ceriottm in #310
• Add (and update) lost input.xml from an old example by @ceriottm in #312
• Full depend object implementation of ExchangePotential by @ceriottm in #311
• Allow passing committee data in the extras as a JSON string by @ceriottm in #314
• Instanton3 (Position dependent RPI) by @litman90 in #272
• Weighted baseline example with JSON committee by @ceriottm in #318
• Adding utilities to parse i-PI output files into more standard python / ase objects by @ceriottm in #313
• Another round of performance improvements based on profiling by @ceriottm in #317
• Update version numbers (oops) by @ceriottm in #327
• Doc, list of features: update reference to bosonic pimd by @yotamfe in #329
• Fix nasty queueing bug by @ceriottm in #328
• Add elphmod example by @janberges in #323
• Better PLUMED interface by @ceriottm in #330
• Fixed some sneaky bugs in the new PRNG module by @ceriottm in #333
• time-dependent NVE integrator by @EliaStocco in #309
• Adding python driver capable of running example of phonon steering by @mahrossi in #300
• Update PET driver to work with current code by @sirmarcel in #334
• Fix several minor bugs for the instanton calculations by @litman90 in #324
• Added an option to override the default prefix for Unix sockets by @ceriottm in #322
• Update the example model for metatensor by @Luthaf in #315
• Ensure that sockets_prefix is passed to run_driver by @sirmarcel in #338
• Added links and contact email in demos/2D-IR-Raman. by @tbegusic in #339
• Add rudimentary parsing of literals to PET driver by @sirmarcel in #341
• apply string replace from longest to shortest label by @svandenhaute in #344
• Enh/picg by @venkatkapil24 in #346
• Gh page update by @litman90 in #349
• trajectory class cleanup by @venkatkapil24 in #335
• Gh page update2 by @litman90 in #351
• add LICENSE.md file before v3.0 release by @litman90 in #347
• A demo of committee models and uncertainty propagation by @ceriottm in #319
• Fix aspects of interface between i-Pi and PLUMED that are not working by @gtribello in #340
• Master is main by @ceriottm in #352
• Fixed buggy env.sh by @ceriottm in #353
• Te PIGS demo by @venkatkapil24 in #354
• ffsocket pbc option is now false as default! by @mahrossi in #355
• Carry on and display more meaningful errors for py driver by @ceriottm in #356
• Fix replay with ASE inputs by @ceriottm in #357
• Handles fixcom = false automatically if some atomic position was fixed by @mahrossi in #360
• Fixed typo and mention that colvar fetching requires plumed 2.10 by @ceriottm in #365

New Contributors

• @mhellstr made their first contribution in #297
• @tbegusic made their first contribution in #279
• @janberges made their first contribution in #323
• @EliaStocco made their first contribution in #309
• @svandenhaute made their first contribution in #344
• @gtribello made their first contribution in #340

Full Changelog: v2.6.1...v3.0