Update comment

Author: RemiLehe

Examples/Tests/embedded_boundary_em_particle_absorption/analysis.py

@@ -1,13 +1,17 @@
 #!/usr/bin/env python
 
-# TODO: update description: explain two particles separate, absorbed by the boundary
-# with EM solver, there can be a static spurious divE (spurious charge build-up) that remains at the position where particle was absorbed
-# this checks this is not the case. Since divE also has a (physical) component that propagate along the EB (due to EM waves reflecting),
-# we average in time to remove this component and only check the static part.
-# TODO: Adjust tolerances
 """
-This analysis script checks for any spurious charge build-up at the embedded boundary, when particles are removed in 3D.
-It averages the divergence of the electric field (divE) over the last 20 time steps and compares the results with a specified tolerance.
+This analysis script checks that there is no spurious charge build-up when a particle is absorbed by an embedded boundary.
+
+More specifically, this test simulates two particles of oppposite charge that are initialized at
+the same position and then move in opposite directions. The particles are surrounded by a spherical
+embedded boundary, and are absorbed when their trajectory intersects this boundary. With an
+electromagnetic solver, this can lead to spurious charge build-up (i.e., div(E)!= rho/epsion_0)
+that remains at the position where particle was absorbed.
+
+Note that, in this test, there will also be a (non-spurious) component of div(E) that propagates
+along the embedded boundary, due to electromagnetic waves reflecting on this boundary.
+When checking for static, spurious charge build-up, we average div(E) in time to remove this component.
 """
 
 from openpmd_viewer import OpenPMDTimeSeries
@@ -21,6 +25,9 @@
 end_avg_iter = 100
 divE_avg = divE_stacked[start_avg_iter:end_avg_iter].mean(axis=0)
 
+# Adjust the tolerance so that the remaining error due to the propagating
+# div(E) (after averaging) is below this tolerance, but so that any typical
+# spurious charge build-up is above this tolerance. This is dimension-dependent.
 dim = ts.fields_metadata["divE"]["geometry"]
 if dim == "3dcartesian":
     tolerance = 5e-11