Add a tutorial on rotating equivariants

docs/src/software/rascaline.sec

@@ -15,3 +15,4 @@ version from the `github repository <https://github.com/Luthaf/rascaline>`_
 - examples/roy-gch/roy-gch
 - examples/sample-selection/sample-selection
 - examples/periodic-hamiltonian/periodic-hamiltonian
+- examples/rotate-equivariants/rotate-equivariants

examples/heat-capacity/heat-capacity.py

@@ -14,6 +14,8 @@
 model <http://doi.org/10.1063/1.3167790>`_.
 """
 
+# %%
+
 import os
 import subprocess
 import time

examples/rotate-equivariants/README.rst

@@ -0,0 +1,8 @@
+Rotating equivariant properties
+===============================
+
+This example goes through the rotation of spherical tensor properties.
+This basic, but not obvious, technique is useful in the context of
+training non-equivariant neural networks with rotational augmentation
+or writing tests for equivariant operations within rotationally equivariant
+models.

examples/rotate-equivariants/environment.yml

@@ -0,0 +1,9 @@
+channels:
+  - conda-forge
+dependencies:
+  - python=3.11
+  - pip 
+  - pip:
+    - ase==3.22.1
+    - spherical
+    - featomic