Update version

.github/workflows/docs.yml

@@ -103,12 +103,9 @@ jobs:
           cd docs
           make html
 
-# event type from https://stackoverflow.com/questions/61886993/in-github-actions-how-to-get-the-type-of-a-trigger-event-as-a-variable
-# conditional control ("if:") from https://docs.github.com/en/actions/using-jobs/using-conditions-to-control-job-execution
-# dependency from https://docs.github.com/en/actions/using-jobs/using-jobs-in-a-workflow
-
   deploy-docs:
-    if: github.event.action == 'push'
+    name: Publish documentation
+    if: startsWith(github.ref, "refs/tags/") 
     needs: build-docs
     runs-on: ubuntu-latest
 

README.md

@@ -10,38 +10,23 @@ For examples and more information see [documentation](https://libatoms.github.io
 
 `wfl` and its dependencies may be installed via `pip install wfl`. 
 
-NOTE: (as of 14 June 2024) `wfl` is only tested against the latest ASE pip release, currently v3.23.0. 
-For the time being, v3.22.1 is the minimum version listed as a (pip) prerequisite of `wfl`, because
-it at least mostly works, but it may not be fully compatible and is not actively tested.
 
 # Recent changes
 
-v0.2.3:
+v0.3.1:
 
-- Add wfl.generate.neb, with required improved support for passing ConfigSet.groups() to 
-  autoaparallelized functions
+- additional updates to file-based calculators for ASE v3.23.
+- fixes to parity plots
 
-- Improved handling of old and new style ase.calculators.espresso.Espresso initialization
+v0.3.0:
 
-v0.2.2:
+- Update the file-based calculators (Orca, FHI-Aims, Vasp, Quantum Espresso, Castep) to work 
+  with with ASE v3.23. This update breaks backwards-compatibility. For compatibility with with 
+  the ASE v3.22 see use wfl v0.2.7 or earlier. 
 
-- Improve checking of DFT calculator convergence
+v0.2.7:
 
-v0.2.1:
+- Latest version compatible with ASE v3.22.x. To install, use `pip install wfl==0.2.7`. 
 
-- Fix group iterator
+For older changes see [documentation](https://libatoms.github.io/workflow).
 
-v0.2.0:
-
-- Change all wfl operations to use explicit random number generator [pull 285](https://github.com/libAtoms/workflow/pull/285), to improve reproducibility of scripts and reduce the chances that on script rerun, cached jobs will not be recognized due to uncontrolled change in random seed (as in [issue 283](https://github.com/libAtoms/workflow/issues/283) and [issue 284](https://github.com/libAtoms/workflow/issues/284)).  Note that this change breaks backward compatibility because many functions now _require_ an `rng` argument, for example
-  ```python
-  rng = np.random.default_rng(1)
-  md_configs = md.md(..., rng=rng, ...)
-  ```
-
-v0.1.0:
-
-- make it possible to fire off several remote autoparallelized ops without waiting for their jobs to finish
-- multi-pass calculation in `Vasp`, to allow for things like GGA followed by HSE
-- MACE fitting, including remote jobs
-- various bug fixes

docs/source/conf.py

@@ -12,6 +12,8 @@
 #
 import os
 import sys
+from wfl import __version__
+
 sys.path.insert(0, os.path.abspath('../../wfl'))
 
 
@@ -21,6 +23,8 @@
 #copyright = '2021, N. Bernstein, T. K. Stenczel, E. Gelzinyte'
 author = 'N. Bernstein, T. K. Stenczel, E. Gelzinyte'
 
+release = __version__
+
 
 # -- General configuration ---------------------------------------------------
 

docs/source/index.rst

@@ -23,14 +23,6 @@ Quick start that installs all of the mandatory dependencies:
 
 	python3 -m pip install wfl
 
-.. note::
-
-    (as of 14 June 2024) `wfl` is only tested against the latest
-    ASE pip release, currently v3.23.0.  For the time being, v3.22.1 is
-    the mininum version listed as a (pip) prerequisite of `wfl`, because
-    it at least mostly works, but it may not be fully compatible and is
-    not actively tested.
-
 
 ***************************************
 Repository
@@ -39,6 +31,67 @@ Repository
 Please find the code, raise issues and cotribute at https://github.com/libAtoms/workflow.  
 
 
+***************************************
+Development
+***************************************
+
+To install all dependencies needed for running unit tests:
+
+.. code-block:: sh
+
+	python3 -m pip install /path/to/workflow[test] 
+
+
+Some of the `wfl` functions rely on the `quippy-ase` package, which currently (2 September 2024) supports no higher Python versions than v3.9.
+
+The file-based calculator tests need ASE calculator configiguration file to be present; `see the ASE documentation <https://wiki.fysik.dtu.dk/ase/ase/calculators/calculators.html#calculator-configuration>`. 
+
+
+***************************************
+Recent Changes
+***************************************
+
+v0.3.0:
+
+- Update the file-based calculators (Orca, FHI-Aims, Vasp, Quantum Espresso, Castep) to work 
+  with with ASE v3.23. This update breaks backwards-compatibility. For compatibility with with 
+  the ASE v3.22 see use wfl v0.2.7 or earlier. 
+
+v0.2.7:
+
+- Latest version compatible with ASE v3.22.x. To install, use `pip install wfl==0.2.7`. 
+
+v0.2.3:
+
+- Add wfl.generate.neb, with required improved support for passing ConfigSet.groups() to 
+  autoaparallelized functions
+
+- Improved handling of old and new style ase.calculators.espresso.Espresso initialization
+
+v0.2.2:
+
+- Improve checking of DFT calculator convergence
+
+v0.2.1:
+
+- Fix group iterator
+
+v0.2.0:
+
+- Change all wfl operations to use explicit random number generator [pull 285](https://github.com/libAtoms/workflow/pull/285), to improve reproducibility of scripts and reduce the chances that on script rerun, cached jobs will not be recognized due to uncontrolled change in random seed (as in [issue 283](https://github.com/libAtoms/workflow/issues/283) and [issue 284](https://github.com/libAtoms/workflow/issues/284)).  Note that this change breaks backward compatibility because many functions now _require_ an `rng` argument, for example
+  ```python
+  rng = np.random.default_rng(1)
+  md_configs = md.md(..., rng=rng, ...)
+  ```
+
+v0.1.0:
+
+- make it possible to fire off several remote autoparallelized ops without waiting for their jobs to finish
+- multi-pass calculation in `Vasp`, to allow for things like GGA followed by HSE
+- MACE fitting, including remote jobs
+- various bug fixes
+
+
 ***************************************
 Contents
 ***************************************

pyproject.toml

@@ -1,12 +1,12 @@
 [project]
 
 name = "wfl"
-version = "0.3.0"
 requires-python = ">=3.9"
 dependencies = [ "click>=7.0", "numpy<2", "ase>=3.22.1", "pyyaml", "spglib",
-                 "docstring_parser", "expyre-wfl", "universalSOAP" ]
+                 "docstring_parser", "expyre-wfl", "universalSOAP", "pandas" ]
 readme = "README.md"
 license = { file = "LICENSE" }
+dynamic = ["version"]
 
 [project.scripts]
     wfl = "wfl.cli.cli:cli"
@@ -16,3 +16,24 @@ license = { file = "LICENSE" }
 
 [tool.setuptools.packages.find]
     exclude = [ "test*" ]
+
+[tool.setuptools.dynamic]
+version = {attr  = "wfl.__version__"}
+
+[project.optional-dependencies]
+test = [
+    "pytest",
+    "mace-torch",
+    "quippy-ase",
+    "requests",
+    "rdkit==2024.3.3",
+]
+doc = [
+    "sphinx",
+    "sphinx_click",
+    "nbsphinx",
+    "myst_parser",
+    "sphinx_book_theme",
+]
+all = ["wfl[test,doc]"]
+

tests/calculators/test_ase_fileio_caching.py

@@ -8,8 +8,8 @@
 ########################
 # test Vasp calculator
 
-from tests.calculators.test_vasp import test_vasp_mark
-@pytest.mark.skipif(test_vasp_mark, reason='Vasp testing env vars missing')
+from tests.calculators.test_vasp import pytestmark as vasp_pytestmark
+@vasp_pytestmark
 def test_vasp_cache_timing(tmp_path, monkeypatch):
     from ase.calculators.vasp import Vasp as Vasp_ase
     from wfl.calculators.vasp import Vasp as Vasp_wrap
@@ -27,7 +27,7 @@ def test_vasp_cache_timing(tmp_path, monkeypatch):
 ########################
 # test quantum espresso calculator
 from tests.calculators.test_qe import espresso_avail, qe_pseudo
-@pytest.mark.skipif(not espresso_avail, reason='qe testing env vars missing')
+@espresso_avail
 def test_qe_cache_timing(tmp_path, monkeypatch, qe_pseudo):
     from ase.calculators.espresso import Espresso as Espresso_ASE
     from wfl.calculators.espresso import Espresso as Espresso_wrap

tests/calculators/test_calc_generic.py

@@ -15,6 +15,8 @@
 from wfl.calculators.espresso import Espresso
 from wfl.autoparallelize import AutoparaInfo
 
+from tests.calculators.test_qe import espresso_avail
+
 ref_lj_energy = -4.52573996914352
 ref_morse_energy = -3.4187397762024867
 
@@ -139,6 +141,7 @@ def test_generic_autopara_defaults():
     sys.stderr = sys.__stderr__
     assert "num_inputs_per_python_subprocess=3" in l_stderr.getvalue()
 
+@espresso_avail
 def test_generic_DFT_autopara_defaults(tmp_path, monkeypatch):
     ats = [Atoms('Al2', positions=[[0,0,0], [1,1,1]], cell=[10]*3, pbc=[True]*3) for _ in range(50)]
 

tests/calculators/test_wrapped_calculator.py

@@ -6,7 +6,7 @@
 ########################
 # test a RuntimeWarning is raised when using the Espresso Calculator directly from ase
 from tests.calculators.test_qe import espresso_avail, qe_pseudo
-@pytest.mark.skipif(not espresso_avail, reason='qe testing env vars missing')
+@espresso_avail
 def test_wrapped_qe(tmp_path, qe_pseudo):
     from ase.calculators.espresso import Espresso as Espresso_ASE
     from wfl.calculators.espresso import Espresso as Espresso_wrap
@@ -27,4 +27,4 @@ def test_wrapped_qe(tmp_path, qe_pseudo):
 
     direct_calc = (Espresso_ASE, [],  kwargs)
     kwargs_generic = dict(inputs=ConfigSet(config), outputs=OutputSpec(), calculator=direct_calc)
-    pytest.warns(RuntimeWarning, generic.calculate, **kwargs_generic)
+    pytest.warns(RuntimeWarning, generic.calculate, **kwargs_generic)

wfl/__init__.py

@@ -0,0 +1 @@
+from .__version__ import __version__

wfl/__version__.py

@@ -0,0 +1 @@
+__version__ = "0.3.1"

wfl/calculators/utils.py

@@ -22,31 +22,4 @@ def clean_rundir(rundir, keep_files, default_keep_files, calculation_succeeded):
     else:
         clean_dir(rundir, keep_files, force=False)
 
-def parse_genericfileio_profile_argv(argv):
-    """Parse a command provided as a conventional argv into the separate
-    structures that generic file-io calculators use to construct their Profile
 
-    Parameters
-    ----------
-    argv: list(str)
-        command to execute, split into separate arguments (e.g. using shlex.split?)
-
-    Returns
-    -------
-    binary: str binary to execute
-    parallel_info: dict with parallel info, in particular "binary" for mpirun/mpiexec/srun etc,
-                   and additional fields to reconstruct rest of command line (all fake, depending
-                   on details of how ASE constructs the final command line
-    """
-    binary = argv[-1]
-    parallel_info = None
-    if len(argv) > 1:
-        # assume earlier arguments are parallel execution dependent, in particular
-        # mpirun/mpiexec/srun [other mpi argument] pw_executable
-        parallel_info = {"binary": argv[0]}
-        for arg in argv[1:-1]:
-            # add additional arguments, faked into a dict that ASE will convert into
-            # a proper command line
-            parallel_info[arg] = True
-
-    return binary, parallel_info