openmm

.github/workflows/CI.yml

@@ -165,7 +165,7 @@ jobs:
       #if: false
       run: |
         conda remove qcelemental --force
-        python -m pip install 'git+https://github.com/loriab/QCElemental.git@csse_layout_536a' --no-deps
+        python -m pip install 'git+https://github.com/loriab/QCElemental.git@csse_layout_536b' --no-deps
 
       # note: conda remove --force, not mamba remove --force b/c https://github.com/mamba-org/mamba/issues/412
       #       alt. is micromamba but not yet ready for setup-miniconda https://github.com/conda-incubator/setup-miniconda/issues/75

docs/source/changelog.rst

@@ -70,11 +70,13 @@ Misc.
 
 MUST (Unmerged)
 +++++++++++++++
-- When qcengine.compute() fails and forms a fop = FailedOperation (raise_error=T), with v2, `fop.input_data` will be an <>Input model (when possible; if the error was in forming the model, it'll still be a dict), not always a dict like v1.
-- When <executor>.compute() fails and collects the input for processing, with v2 it now uses the <>Input model passed to the executor, not the model-or-dict passed into compute().
-- The net result of the two above is that whereas fop.input_data in v1 was reliably a dict and its contents would reflect whether a model or dict was passed to qcengine.compute(), now in v2, fop.input_data is a model whenever possible (to mirror <>Result.input_data) regardless of model or dict passed to qcengine.compute(); the only case where it's a dict is if the error was in forming the model.
-- DFTD3 & DFTD4 (new intf) - intercept ``v1.AtomicResult`` with ``success=False`` and ``error`` fields set from QCSchema interfaces and return ``FailedOperation``s. Someday when upstream switches to v2, request packages return FaileOp directly and use ``input_error`` rather than ``input error``.
-- ``qcengine run`` learned new argument ``--return-version`` analogous to ``qcengine.compute(..., return_version=1|2)`` so CLI matches API capabilities. Note *not* ported to phasing-out ``qcengine run-procedure``.
+If you're missing something from AtomicResult.extras, check AtomicResult.input_data.extras in case it was passed in on input
+- (:pr:`459`) OpenMM gained AtomicResult.properties.return_gradient
+- (:pr:`458`) When qcengine.compute() fails and forms a fop = FailedOperation (raise_error=T), with v2, `fop.input_data` will be an <>Input model (when possible; if the error was in forming the model, it'll still be a dict), not always a dict like v1.
+- (:pr:`458`) When <executor>.compute() fails and collects the input for processing, with v2 it now uses the <>Input model passed to the executor, not the model-or-dict passed into compute().
+- (:pr:`458`) The net result of the two above is that whereas fop.input_data in v1 was reliably a dict and its contents would reflect whether a model or dict was passed to qcengine.compute(), now in v2, fop.input_data is a model whenever possible (to mirror <>Result.input_data) regardless of model or dict passed to qcengine.compute(); the only case where it's a dict is if the error was in forming the model.
+- (:pr:`458`) DFTD3 & DFTD4 (new intf) - intercept ``v1.AtomicResult`` with ``success=False`` and ``error`` fields set from QCSchema interfaces and return ``FailedOperation``s. Someday when upstream switches to v2, request packages return FaileOp directly and use ``input_error`` rather than ``input error``.
+- (:pr:`458`) ``qcengine run`` learned new argument ``--return-version`` analogous to ``qcengine.compute(..., return_version=1|2)`` so CLI matches API capabilities. Note *not* ported to phasing-out ``qcengine run-procedure``.
 
 WIP (Unmerged)
 ++++++++++++++

qcengine/programs/openmm.py

@@ -324,13 +324,16 @@ def compute(self, input_model: "AtomicInput", config: "TaskConfig") -> "AtomicRe
 
             # Force to gradient
             ret_data["return_result"] = -1 * q
+            ret_data["properties"]["return_gradient"] = -1 * q
+            ret_data["properties"]["calcinfo_natom"] = len(input_model.molecule.symbols)
         else:
             raise InputError(f"Driver {input_model.driver} not implemented for OpenMM.")
 
         ret_data["success"] = True
-        ret_data["extras"] = input_model.extras
+        ret_data["input_data"] = input_model
+        ret_data["molecule"] = input_model.molecule  # should connectivity be added from off_mol?
 
         # Move several pieces up a level
         ret_data["provenance"] = Provenance(creator="openmm", version=openmm.version.short_version, nthreads=nthreads)
 
-        return AtomicResult(**{**input_model.dict(), **ret_data})
+        return AtomicResult(**ret_data)