From 3c40e230e2b6c1c09b9158dda5cc3890aa5d1407 Mon Sep 17 00:00:00 2001 From: Jammy2211 Date: Fri, 11 Sep 2026 11:23:06 -0400 Subject: [PATCH] Declare Pixelization.__solved_parameters__ for model figures (PyAutoLens#736) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The source reconstruction is solved by the inversion, not sampled. Declaring it through PyAutoFit's new `__solved_parameters__` protocol lets the model figure draw a dashed `reconstruction ยท solved` pill on the Pixelization card instead of an empty component. One class attribute, no behaviour change. Co-Authored-By: Claude Fable 5.1 Claude-Session: https://claude.ai/code/session_0178yLU9v19GtACcRPFggjGa --- autoarray/inversion/pixelization.py | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/autoarray/inversion/pixelization.py b/autoarray/inversion/pixelization.py index 477bf059e..7d667456c 100644 --- a/autoarray/inversion/pixelization.py +++ b/autoarray/inversion/pixelization.py @@ -8,6 +8,13 @@ class Pixelization: + # The source reconstruction is solved for by the inversion, not sampled by + # the non-linear search, so it has no prior and never appears in + # `model.info`. PyAutoFit's `graph_spec` reads this class attribute (the + # `__solved_parameters__` protocol) to draw `reconstruction` as a `solved` + # row in model figures. + __solved_parameters__ = ("reconstruction",) + def __init__( self, mesh: AbstractMesh,