htpolynet 2.13.0__tar.gz → 2.14.0__tar.gz

{htpolynet-2.13.0 → htpolynet-2.14.0}/CHANGELOG.md

@@ -7,6 +7,43 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [2.14.0] - 2026-09-30
+
+### Changed
+
+- **No packaged mdp uses a Berendsen thermostat or barostat any more, and every one of
+  them now passes `grompp` with zero warnings** --- which none of them did before.
+
+- **Every temperature-coupled stage now uses the V-rescale thermostat.**  All seven
+  packaged mdps used Berendsen, which Gromacs deprecates in the same breath as the
+  barostat: *"does not generate the correct kinetic energy distribution."*  That is every
+  stage htpolynet runs, including the ones that produce published densities and
+  transition temperatures.  Same cause as the barostat, and the same remedy; `tau_t` is
+  unchanged at 0.1 ps.
+
+- **Constant-pressure stages no longer use the Berendsen barostat.**  `npt.mdp`,
+  `drag-npt.mdp` and `relax-npt.mdp` now use **C-rescale**, which samples a correct NPT
+  ensemble and is robust far from equilibrium; Berendsen is neither, and Gromacs has said
+  so at every `grompp` since 2025 --- unheard, because those warnings were being
+  discarded (fixed in 2.13.0).  This covers densification, precure and postcure
+  equilibration, the CURE relax and drag ladders, and the ring cure's relax and
+  equilibrate.
+
+  The deformation stages are the exception: they force `pcoupltype = anisotropic`, which
+  C-rescale does not support at all --- `grompp` errors outright --- so `deform` and
+  clamped `shear` now set **Parrinello-Rahman**, which does support it and also samples
+  the right ensemble.  Isochoric shear has no barostat and is unaffected.
+
+  **This changes the sampled ensemble.**  Densities, glass-transition temperatures and
+  moduli from earlier versions are not strictly comparable with ones produced now.  That
+  cost is deliberate and was taken on Cameron's decision, in the same spirit as the 2.7.0
+  constraint change.  `tau_p` is unchanged at 0.5 ps, so that if anything proves unstable
+  there is only one variable to look at; it is the first knob to try.
+
+  A test now fails if any packaged mdp reintroduces a deprecated barostat, or pairs
+  C-rescale with the anisotropic coupling it cannot do.
+
+
 ## [2.13.0] - 2026-09-29
 
 ### Changed

{htpolynet-2.13.0 → htpolynet-2.14.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: htpolynet
-Version: 2.13.0
+Version: 2.14.0
 Summary: Automated MD System Builder for Amorphous Network Polymers
 Project-URL: Source, https://github.com/cameronabrams/htpolynet
 Project-URL: Documentation, https://htpolynet.readthedocs.io/

{htpolynet-2.13.0 → htpolynet-2.14.0}/ROADMAP.md

@@ -601,7 +601,9 @@ Coverage as of the last measurement: **38.8%** overall.
   question is whether a short one converges at all or whether the barostat needs
   something like CURE's 100 ps --- which is the other thing the A/B will show.
 
-- **Properties measured before the threading filter may carry threaded rings.**  The
+- **DECLINED 2026-09-30 by Cameron: "no, I am not interested in answering that
+  question."**  Left here so it is not raised again as though it were open.  What was
+  asked: properties measured before the threading filter may carry threaded rings.  The
   filter shipped in 2.11.4; every build made before it could contain a triazine closed
   around a monomer, which happened in 2 of 4 unfiltered builds --- 4, 2, 0 and 0
   piercings, a mean of 1.5 each.
@@ -1863,31 +1865,6 @@ Coverage as of the last measurement: **38.8%** overall.
 
 ## Simulation defaults
 
-- **Every constant-pressure stage uses the Berendsen barostat, which does not sample the
-  NPT ensemble and is deprecated upstream.**  `npt.mdp`, `drag-npt.mdp` and
-  `relax-npt.mdp` all set `pcoupl = Berendsen`, and nothing in the Python overrides it,
-  so this covers densification, precure and postcure equilibration, the CURE relax
-  ladder, the ring cure's relax and equilibrate, and both deformation stages.  Gromacs
-  2025 says so at every `grompp`: *"The Berendsen barostat does not generate any strictly
-  correct ensemble, and should not be used for new production simulations... We recommend
-  using the C-rescale barostat instead."*
-
-  Raised by htpolynet-study 2026-09-28 in the context of moduli, where it matters most:
-  a barostat that does not sample the right volume fluctuations is a poor foundation for
-  a mechanical measurement.
-
-  **The reason this has not simply been changed is reproducibility, not doubt.**
-  Switching to C-rescale changes the sampled ensemble, so every density, Tg and modulus
-  htpolynet has produced becomes not strictly comparable with anything produced after.
-  There is precedent for taking that cost deliberately --- 2.7.0 changed the constraint
-  scheme for the same kind of reason and said so in the notes --- and the same treatment
-  fits here: change all three together, say plainly in the changelog that prior results
-  are not comparable, and ideally re-measure one known system across the change so the
-  size of the shift is on record rather than assumed.
-
-  Whoever takes it should also check whether `tau_p` wants revisiting: a value tuned for
-  Berendsen's weak coupling is not necessarily right for C-rescale.
-
 - **The density-convergence tolerance may be tight for example 6.**  In the
   v2.10.0 sweep, example 6 failed the gate for the first time in four sweeps:
   "sem 1.070 exceeds tolerance 1.000" after 6 extensions, with the last window

{htpolynet-2.13.0 → htpolynet-2.14.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "htpolynet"
-version = "2.13.0"
+version = "2.14.0"
 authors = [
   { name="Cameron F Abrams", email="cfa22@drexel.edu" },
 ]

{htpolynet-2.13.0 → htpolynet-2.14.0}/src/htpolynet/analysis/postsim.py

@@ -339,6 +339,9 @@ class PostSimDeform(PostSimMD):
             'tau_p':1.0,
             'refcoord_scaling': 'com',
             'pcoupltype': 'anisotropic',
+            # the packaged mdp uses C-rescale, which does not support anisotropic
+            # coupling; Parrinello-Rahman does and samples the right ensemble too
+            'pcoupl': 'Parrinello-Rahman',
             # stress is what this stage exists to measure, so sample it far more often
             # than the packaged npt.mdp does.  At 1 ps the frames are already
             # near-independent, so ten times as many cut the standard error on a fitted
@@ -467,6 +470,8 @@ class PostSimShear(PostSimMD):
             'tau_p':1.0,
             'refcoord_scaling': 'com',
             'pcoupltype': 'anisotropic',
+            # as in PostSimDeform: C-rescale cannot do anisotropic, Parrinello-Rahman can
+            'pcoupl': 'Parrinello-Rahman',
             # the Cartesian component the shear moves is uncoupled; the other two
             # normal directions stay at P.  Gromacs rejects the run otherwise.
             'ref_p':f'{params["P"]} {params["P"]} {params["P"]} 0 0 0',

{htpolynet-2.13.0 → htpolynet-2.14.0}/src/htpolynet/resources/mdp/drag-npt.mdp

@@ -19,11 +19,11 @@ fourierspacing  = 0.12
 pme_order       = 4
 ewald_rtol      = 1e-5
 optimize_fft    = yes
-tcoupl      = Berendsen
+tcoupl              = V-rescale     ; correct canonical ensemble; Berendsen does not give the right kinetic energy distribution (deprecated 2025+)
 tc-grps     = System
 tau_t       = 0.1
 ref_t       = 300
-pcoupl              = Berendsen     ; Pressure coupling on in NPT, also weak coupling
+pcoupl              = C-rescale     ; correct NPT ensemble and robust far from equilibrium; Berendsen is neither (deprecated 2025+)
 pcoupltype          = isotropic     ; uniform scaling of x-y-z box vectors
 tau_p               = 0.5           ; time constant, in ps
 ref_p               = 1.0

{htpolynet-2.13.0 → htpolynet-2.14.0}/src/htpolynet/resources/mdp/drag-nvt.mdp

@@ -19,7 +19,7 @@ fourierspacing  = 0.12
 pme_order       = 4
 ewald_rtol      = 1e-5
 optimize_fft    = yes
-tcoupl      = Berendsen
+tcoupl              = V-rescale     ; correct canonical ensemble; Berendsen does not give the right kinetic energy distribution (deprecated 2025+)
 tc-grps     = System
 tau_t       = 0.1
 ref_t       = 300

{htpolynet-2.13.0 → htpolynet-2.14.0}/src/htpolynet/resources/mdp/npt.mdp

@@ -18,11 +18,11 @@ fourierspacing  = 0.12
 pme_order       = 4
 ewald_rtol      = 1e-5
 optimize_fft    = yes
-tcoupl      = Berendsen
+tcoupl              = V-rescale     ; correct canonical ensemble; Berendsen does not give the right kinetic energy distribution (deprecated 2025+)
 tc-grps     = System
 tau_t       = 0.1
 ref_t       = 300
-pcoupl              = Berendsen     ; Pressure coupling on in NPT, also weak coupling
+pcoupl              = C-rescale     ; correct NPT ensemble and robust far from equilibrium; Berendsen is neither (deprecated 2025+)
 pcoupltype          = isotropic     ; uniform scaling of x-y-z box vectors
 tau_p               = 0.5           ; time constant, in ps
 ref_p               = 10.0

{htpolynet-2.13.0 → htpolynet-2.14.0}/src/htpolynet/resources/mdp/nvt.mdp

@@ -17,7 +17,7 @@ fourierspacing  = 0.12
 pme_order       = 4
 ewald_rtol      = 1e-5
 optimize_fft    = yes
-tcoupl      = Berendsen
+tcoupl              = V-rescale     ; correct canonical ensemble; Berendsen does not give the right kinetic energy distribution (deprecated 2025+)
 tc-grps     = System
 tau_t       = 0.1
 ref_t       = 300

{htpolynet-2.13.0 → htpolynet-2.14.0}/src/htpolynet/resources/mdp/relax-npt.mdp

@@ -19,11 +19,11 @@ fourierspacing  = 0.12
 pme_order       = 4
 ewald_rtol      = 1e-5
 optimize_fft    = yes
-tcoupl      = Berendsen
+tcoupl              = V-rescale     ; correct canonical ensemble; Berendsen does not give the right kinetic energy distribution (deprecated 2025+)
 tc-grps     = System
 tau_t       = 0.1
 ref_t       = 300
-pcoupl              = Berendsen     ; Pressure coupling on in NPT, also weak coupling
+pcoupl              = C-rescale     ; correct NPT ensemble and robust far from equilibrium; Berendsen is neither (deprecated 2025+)
 pcoupltype          = isotropic     ; uniform scaling of x-y-z box vectors
 tau_p               = 0.5           ; time constant, in ps
 ref_p               = 1.0

{htpolynet-2.13.0 → htpolynet-2.14.0}/src/htpolynet/resources/mdp/relax-nvt.mdp

@@ -19,7 +19,7 @@ fourierspacing  = 0.12
 pme_order       = 4
 ewald_rtol      = 1e-5
 optimize_fft    = yes
-tcoupl      = Berendsen
+tcoupl              = V-rescale     ; correct canonical ensemble; Berendsen does not give the right kinetic energy distribution (deprecated 2025+)
 tc-grps     = System
 tau_t       = 0.1
 ref_t       = 300

{htpolynet-2.13.0 → htpolynet-2.14.0}/src/htpolynet/resources/mdp/single-molecule-nvt.mdp

@@ -14,7 +14,7 @@ rcoulomb    = 1.0
 rvdw        = 1.0
 coulombtype = cut-off
 vdw_type	= cut-off
-tcoupl      = Berendsen
+tcoupl              = V-rescale     ; correct canonical ensemble; Berendsen does not give the right kinetic energy distribution (deprecated 2025+)
 tc-grps     = System
 tau_t       = 0.1
 ref_t       = 300

htpolynet-2.14.0/tests/unit/test_mdp_ensembles.py

@@ -0,0 +1,109 @@
+"""
+
+.. module:: test_mdp_ensembles
+   :synopsis: the packaged mdps ask for couplings that sample the right ensemble
+
+.. moduleauthor: Cameron F. Abrams, <cfa22@drexel.edu>
+
+"""
+import importlib.resources as ir
+import unittest
+
+DEPRECATED_BAROSTATS = {'berendsen'}
+"""Gromacs 2025+: "The Berendsen barostat does not generate any strictly correct
+ensemble, and should not be used for new production simulations." """
+
+DEPRECATED_THERMOSTATS = {'berendsen'}
+"""Gromacs 2025+: "The Berendsen thermostat does not generate the correct kinetic
+energy distribution... We would recommend the V-rescale thermostat." """
+
+
+def mdps():
+    d = ir.files('htpolynet').joinpath('resources/mdp')
+    return {f.name: f.read_text() for f in d.iterdir() if f.name.endswith('.mdp')}
+
+
+def setting(text, key):
+    for line in text.split('\n'):
+        if '=' in line and line.split('=')[0].strip() == key:
+            return line.split('=')[1].split(';')[0].strip()
+    return None
+
+
+class TestNoPackagedMdpUsesADeprecatedBarostat(unittest.TestCase):
+    """This went unnoticed for as long as Gromacs has warned about it, because
+    htpolynet runs grompp with -maxwarn 4 and discarded its output on success.  A test
+    is the only thing that would have caught it, so here is the test."""
+
+    def test_every_pressure_coupled_mdp_uses_a_correct_ensemble(self):
+        bad = {n: setting(t, 'pcoupl') for n, t in mdps().items()
+               if (setting(t, 'pcoupl') or 'no').lower() in DEPRECATED_BAROSTATS}
+        self.assertEqual(bad, {})
+
+    def test_the_constant_pressure_mdps_are_actually_coupled(self):
+        # guard against "fixing" the warning by switching pressure coupling off
+        for name in ('npt.mdp', 'drag-npt.mdp', 'relax-npt.mdp'):
+            with self.subTest(name):
+                self.assertEqual((setting(mdps()[name], 'pcoupl') or 'no').lower(),
+                                 'c-rescale')
+
+    def test_c_rescale_is_only_used_where_it_is_supported(self):
+        # C-rescale does not support anisotropic coupling; grompp errors outright.
+        # The deformation stages override pcoupltype at runtime and set their own
+        # barostat, so no packaged mdp may pair the two.
+        for name, text in mdps().items():
+            with self.subTest(name):
+                if (setting(text, 'pcoupl') or '').lower() == 'c-rescale':
+                    self.assertNotEqual((setting(text, 'pcoupltype') or '').lower(),
+                                        'anisotropic')
+
+
+class TestNoPackagedMdpUsesADeprecatedThermostat(unittest.TestCase):
+    """Every mdp htpolynet ships used the Berendsen thermostat, on every stage including
+    the ones that produce published numbers.  Same cause as the barostat: grompp said so
+    at every run, into output that was thrown away."""
+
+    def test_no_mdp_uses_one(self):
+        bad = {n: setting(t, 'tcoupl') for n, t in mdps().items()
+               if (setting(t, 'tcoupl') or 'no').lower() in DEPRECATED_THERMOSTATS}
+        self.assertEqual(bad, {})
+
+    def test_the_temperature_coupled_mdps_are_actually_coupled(self):
+        # guard against "fixing" it by switching temperature coupling off
+        for name, text in mdps().items():
+            with self.subTest(name):
+                if setting(text, 'tcoupl') is not None:
+                    self.assertEqual(setting(text, 'tcoupl').lower(), 'v-rescale')
+
+
+class TestTheDeformationStagesCarryTheirOwnBarostat(unittest.TestCase):
+    """They force anisotropic coupling, which C-rescale cannot do."""
+
+    def built(self, cls, params):
+        import os
+        import shutil
+        import tempfile
+
+        import numpy as np
+        d = tempfile.mkdtemp()
+        f = os.path.join(d, 'm.mdp')
+        shutil.copy(str(ir.files('htpolynet').joinpath('resources/mdp/npt.mdp')), f)
+        cls(params).build_mdp(f, box=np.diag([5.6, 5.6, 5.6]))
+        return open(f).read()
+
+    def test_deform_uses_parrinello_rahman(self):
+        from htpolynet.analysis.postsim import PostSimDeform
+        text = self.built(PostSimDeform, {'direction': 'x', 'edot': 1e-4, 'ps': 300})
+        self.assertEqual(setting(text, 'pcoupl'), 'Parrinello-Rahman')
+        self.assertEqual(setting(text, 'pcoupltype'), 'anisotropic')
+
+    def test_clamped_shear_uses_parrinello_rahman(self):
+        from htpolynet.analysis.postsim import PostSimShear
+        text = self.built(PostSimShear, {'direction': 'xy', 'edot': 1e-4, 'ps': 300})
+        self.assertEqual(setting(text, 'pcoupl'), 'Parrinello-Rahman')
+
+    def test_isochoric_shear_has_no_barostat_at_all(self):
+        from htpolynet.analysis.postsim import PostSimShear
+        text = self.built(PostSimShear, {'direction': 'xy', 'edot': 1e-4, 'ps': 300,
+                                         'coupling': 'isochoric'})
+        self.assertEqual(setting(text, 'pcoupl'), 'no')