Example: a custom fourfit fringe-plot plugin that *wraps* the stock renderer.
fourfit's default fringe plot is produced by
``hops_visualization.fourfit_plot``. To customize it you do **not** need to copy
that ~1000-line module -- import the stock renderer and override only the pieces
you want to change. This keeps your plugin tiny and, more importantly, means it
automatically inherits every future fix and layout change to the default plot
instead of silently drifting from it.
How fourfit calls this
----------------------
Put this file where the embedded interpreter can import it -- the directory in
``$HOPS_USER_PLUGINS_DIR`` (or the default plugin dir), which HOPS appends to
``sys.path`` at startup -- then name it in the control file::
plot_backend matplotlib
python_custom_plot custom_fourfit_plot make_custom_fourfit_plot_wrapper
The ``python_custom_plot <module_path> <function_name>`` statement tells
MHO_DefaultPythonPlotVisitor to import ``module_path`` (dot-separated) and call
``function_name(fringe_data_interface)`` instead of the default
``hops_visualization.fourfit_plot:make_fourfit_plot_wrapper``. It only takes
effect with ``plot_backend matplotlib``, and only in a build with embedded
Python (HOPS_ENABLE_EMBEDDED_PYTHON=ON); the libpython-free subprocess backend
cannot run custom plots and renders the default plot instead.
The customization below is deliberately trivial (a small footer annotation) so
the *pattern* is obvious: keep a handle to the stock helper, wrap it, and
reinstall it on the imported module so ``make_fourfit_plot()`` picks it up.
