Interface index numbering changes within a snapshot run

[keckler]

I noticed that during a snapshot run the numbering of interfaces as they are printed in the STDOUT changes between the interactInit() and interactBOC() times. Here are some pictures from my specific example:

Note that the depletion interface is removed from the stack printing at BOC and later. Presumably this is because the depletion interface is not active during snapshot runs. But I would expect the interface indices to remain constant throughout the simulation.

This is of very low importance right now, just pointing it out.

[john-science]

Note that the depletion interface is removed from the stack printing at BOC and later.

This might be a function of code written in your depletion module. It might be moving itself in the stack order, and that is by design and out of the control of ARMI code.

My first thought is I would want to look at that code.

[keckler]

I think what is happening is that the CrossSectionGroupManager is being removed from the stack during a snapshot run, and then this enumerate call is only seeing the real-time stack:

armi/armi/bookkeeping/report/reportingUtils.py

Lines 407 to 435 in 8ab9099

	def getInterfaceStackSummary(o):
	data = []
	for ii, i in enumerate(o.interfaces, start=1):
	data.append(
	(
	"{:02d}".format(ii),
	i.__class__.__name__.replace("Interface", ""),
	i.name,
	i.function,
	"Yes" if i.enabled() else "No",
	"Reversed" if i.reverseAtEOL else "Normal",
	"Yes" if i.bolForce() else "No",
	)
	)
	text = tabulate.tabulate(
	data,
	headers=(
	"Index",
	"Type",
	"Name",
	"Function",
	"Enabled",
	"EOL order",
	"BOL forced",
	),
	tablefmt="armi",
	)
	text = text
	return text

To fix this, the original stack for the run would need to be stored somewhere. Not the most straightforward this to do, but surely it could be done.