changing DL_Poly units #4983

lexi-x · 2025-03-20T01:45:29Z

Changes made in this Pull Request:

Added unit conversion to DL_Poly
~~Removed forces from TRZ~~
Updated test values for DL_Poly

As mentioned in #2393, this likely requires more work with test cases and code modifications regarding where the DL_Poly conversions are made, so any feedback is appreciated!

PR Checklist

Issue raised/referenced?
Tests updated/added?
Documentation updated/added?
package/CHANGELOG file updated?
Is your name in package/AUTHORS? (If it is not, add it!)

Developers Certificate of Origin

I certify that I can submit this code contribution as described in the Developer Certificate of Origin, under the MDAnalysis LICENSE.

📚 Documentation preview 📚: https://mdanalysis--4983.org.readthedocs.build/en/4983/

github-actions

Hello there first time contributor! Welcome to the MDAnalysis community! We ask that all contributors abide by our Code of Conduct and that first time contributors introduce themselves on GitHub Discussions so we can get to know you. You can learn more about participating here. Please also add yourself to package/AUTHORS as part of this PR.

codecov · 2025-03-21T03:50:15Z

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.41%. Comparing base (dcaa087) to head (a267752).
Report is 13 commits behind head on develop.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #4983      +/-   ##
===========================================
- Coverage    93.42%   93.41%   -0.02%     
===========================================
  Files          177      189      +12     
  Lines        21859    22925    +1066     
  Branches      3078     3078              
===========================================
+ Hits         20422    21415     +993     
- Misses         986     1059      +73     
  Partials       451      451

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

lexi-x · 2025-03-21T04:50:16Z

Based on feedback about TRZ being dropped, I removed the parts regarding TRZ in the latest commit.

orbeckst · 2025-03-30T00:41:38Z

@cbouy and @talagayev could you please review and shepherd this PR? This is a GSOC applicant for one of your projects.

orbeckst

Good start! See comments. Although dividing by 100 gives the right answer, it is a bit more complicated because we want all our readers to be have in a unified manner. Have a look at other readers such as TRRReader where you define the force units

mdanalysis/package/MDAnalysis/coordinates/TRR.py

Lines 150 to 151 in 66e7e5c

    
           units = {'time': 'ps', 'length': 'nm', 'velocity': 'nm/ps', 
        
                    'force': 'kJ/(mol*nm)'}

For DL_POLY, this entry should be "force": "Da*Angstrom/ps" in the _DLPOLY_UNITS dict.

You then need to add at least one entry for "Da*Angstrom/ps" to units.forceUnit_factor

mdanalysis/package/MDAnalysis/units.py

Lines 341 to 350 in 66e7e5c

    
           forceUnit_factor = { 
        
               "kJ/(mol*Angstrom)": 1.0, 
        
               "kJ/(mol*A)": 1.0, 
        
               "kJ/(mol*\u212b)": 1.0, 
        
               "kJ/(mol*nm)": 10.0, 
        
               "Newton": 1e13 / constants["N_Avogadro"], 
        
               "N": 1e13 / constants["N_Avogadro"], 
        
               "J/m": 1e13 / constants["N_Avogadro"], 
        
               "kcal/(mol*Angstrom)": 1 / constants["calorie"], 
        
           }

with the correct conversion factor. Also update relevant docs and tests.

Then in your code, check if unit conversion is requested (similar to what the TRRReader does in

mdanalysis/package/MDAnalysis/coordinates/TRR.py

Lines 213 to 214 in 66e7e5c

    
           if self.convert_units: 
        
               self.convert_forces_from_native(ts.forces)

)

Additionally:

The history reader should also convert forces
add a CHANGELOG entry under Fixes
add a versionchanged 2.10.0 under ConfigReader (and also HistoryReader)

orbeckst · 2025-03-30T00:44:54Z

package/MDAnalysis/coordinates/TRZ.py

@@ -605,4 +605,4 @@ def close(self):
        if self.trzfile is None:
            return
        self.trzfile.close()
-        self.trzfile = None
+        self.trzfile = None


orbeckst · 2025-03-30T00:45:45Z

testsuite/MDAnalysisTests/coordinates/test_dlpoly.py

@@ -70,8 +70,9 @@ def test_velocities(self, ts):
        ref = np.array([1.056610291, -1.218664448, 3.345828610])
        assert_allclose(ts._velocities[0], ref)

+    # modified to original numbers / 100


remove comment – we can always look in the git blame to find out when something was changed

orbeckst · 2025-03-30T00:45:53Z

testsuite/MDAnalysisTests/coordinates/test_dlpoly.py

@@ -118,8 +119,9 @@ def test_vel(self, u):
        ref = np.array([2.637614561, 0.5778767520e-01, -1.704765568])
        assert_allclose(u.atoms[2].velocity, ref)

+    # modified to original numbers / 100 


remove comment

@lexi-x removing this would also fix the linters / black error here https://github.com/MDAnalysis/mdanalysis/actions/runs/13984105125/job/39154722804?pr=4983#step:5:151

orbeckst · 2025-03-30T00:54:34Z

package/MDAnalysis/coordinates/DLPoly.py

@@ -99,7 +99,7 @@ def _read_first_frame(self):
        if has_vels:
            velocities = np.array(velocities, dtype=np.float32)
        if has_forces:
-            forces = np.array(forces, dtype=np.float32)
+            forces = np.array(forces, dtype=np.float32) / 100


add a comment above this line stating that forces in DL_POLY are 10 J/(mol.Å) and mention issue #2393 — this is weird enough to warrant a hint for future coders

Actually, we should not just divide by 100 here. Instead we should use the normal conversion machinery where you store the native force units in _DLPOLY_UNITS and store the conversion factor in units, then use convert_from_native.

See also https://userguide.mdanalysis.org/stable/units.html .

cbouy

I second the changes requested by @orbeckst (I wasn't familiar with the unit conversion API to be honest, today I learned!)

cbouy · 2025-04-02T20:44:40Z

testsuite/MDAnalysisTests/coordinates/test_dlpoly.py

@@ -118,8 +119,9 @@ def test_vel(self, u):
        ref = np.array([2.637614561, 0.5778767520e-01, -1.704765568])
        assert_allclose(u.atoms[2].velocity, ref)

+    # modified to original numbers / 100 


@lexi-x removing this would also fix the linters / black error here https://github.com/MDAnalysis/mdanalysis/actions/runs/13984105125/job/39154722804?pr=4983#step:5:151

lexi-x · 2025-04-03T20:08:53Z

@orbeckst @cbouy Thank you for your recommendations! Sorry for the delayed response, I was caught up in making the proposal and another pull request. I made the changes you recommended and I wanted to make sure that my thought process was valid in implementing them before making commits.

After changing the DL_POLY force units to Da*Angstrom/ps, I added the conversion factor to units.py based on my understanding of the unit conversions:

forceUnit_factor = {
    "kJ/(mol*Angstrom)": 1.0,
    "kJ/(mol*A)": 1.0,
    "kJ/(mol*\u212b)": 1.0,
    "kJ/(mol*nm)": 10.0,
    "Newton": 1e13 / constants["N_Avogadro"],
    "N": 1e13 / constants["N_Avogadro"],
    "J/m": 1e13 / constants["N_Avogadro"],
    "kcal/(mol*Angstrom)": 1 / constants["calorie"],
    "Da*Angstrom/ps": 1000.0 / constants["N_Avogadro"] / 1e10 * 1e12
}

Regarding making the same unit conversions for History Reader, when I originally added the conversion to both functions, the test cases returned results that were scaled by 10000 instead of 100, so I suspect it was making double conversions. Would this still be an issue when using native force units instead of direct conversion? Thank you!

orbeckst · 2025-04-11T06:31:26Z

@lexi-x if all conversions are done in the standard way then there shouldn't be any double conversions. Try it and make the tests pass.

orbeckst · 2025-04-11T06:32:50Z

Explain to us why what you're doing is correct — that's the best way to be sure that you're right and makes it easy for us to review.

Lexi Xu added 3 commits March 19, 2025 20:19

modifying force units

e787218

removing force from TRZ

ad9c571

modifying dlpoly test case

445ec3a

github-actions bot reviewed Mar 20, 2025

View reviewed changes

Lexi Xu added 4 commits March 19, 2025 21:06

changing conversion location

5585acc

typo fix

cfa4e25

fix inconsistencies

f11800b

disregarding TRZ (dropped in 3.0)

a267752

lexi-x marked this pull request as ready for review March 21, 2025 04:48

orbeckst assigned cbouy and talagayev Mar 30, 2025

orbeckst changed the title ~~Dropping TRZ reading forces and changing DL_Poly units~~ changing DL_Poly units Mar 30, 2025

orbeckst requested changes Mar 30, 2025

View reviewed changes

cbouy reviewed Apr 2, 2025

View reviewed changes

	units = {'time': 'ps', 'length': 'nm', 'velocity': 'nm/ps',
	'force': 'kJ/(mol*nm)'}

	forceUnit_factor = {
	"kJ/(mol*Angstrom)": 1.0,
	"kJ/(mol*A)": 1.0,
	"kJ/(mol*\u212b)": 1.0,
	"kJ/(mol*nm)": 10.0,
	"Newton": 1e13 / constants["N_Avogadro"],
	"N": 1e13 / constants["N_Avogadro"],
	"J/m": 1e13 / constants["N_Avogadro"],
	"kcal/(mol*Angstrom)": 1 / constants["calorie"],
	}

	if self.convert_units:
	self.convert_forces_from_native(ts.forces)

changing DL_Poly units #4983

Are you sure you want to change the base?

changing DL_Poly units #4983

Uh oh!

Conversation

lexi-x commented Mar 20, 2025 • edited by orbeckst Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Checklist

Developers Certificate of Origin

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Uh oh!

codecov bot commented Mar 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

lexi-x commented Mar 21, 2025

Uh oh!

orbeckst commented Mar 30, 2025

Uh oh!

orbeckst left a comment

Choose a reason for hiding this comment

Uh oh!

orbeckst Mar 30, 2025

Choose a reason for hiding this comment

Uh oh!

orbeckst Mar 30, 2025

Choose a reason for hiding this comment

Uh oh!

orbeckst Mar 30, 2025

Choose a reason for hiding this comment

Uh oh!

cbouy Apr 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

orbeckst Mar 30, 2025

Choose a reason for hiding this comment

Uh oh!

orbeckst Mar 30, 2025

Choose a reason for hiding this comment

Uh oh!

cbouy left a comment

Choose a reason for hiding this comment

Uh oh!

cbouy Apr 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lexi-x commented Apr 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

orbeckst commented Apr 11, 2025

Uh oh!

orbeckst commented Apr 11, 2025

Uh oh!

Uh oh!

lexi-x commented Mar 20, 2025 •

edited by orbeckst

Loading

codecov bot commented Mar 21, 2025 •

edited

Loading

cbouy Apr 2, 2025 •

edited

Loading

cbouy Apr 2, 2025 •

edited

Loading

lexi-x commented Apr 3, 2025 •

edited

Loading