makeReactionRateTable is inefficient and poorly named

[keckler]

armi/armi/physics/neutronics/isotopicDepletion/crossSectionTable.py

Lines 164 to 188 in 4d45672

	def makeReactionRateTable(obj, nuclides: List = None):
	"""
	Generate a reaction rate table for given nuclides.
	Often useful in support of depletion.
	.. impl:: Generate cross section table.
	:id: I_ARMI_DEPL_TABLES1
	:implements: R_ARMI_DEPL_TABLES
	Parameters
	----------
	nuclides : list, optional
	list of nuclide names for which to generate the cross-section table.
	If absent, use all nuclides obtained by self.getNuclides().
	Notes
	-----
	This also used to do some caching on the block level but that has been removed
	and the calls to this may therefore need to be re-optimized.
	See Also
	--------
	armi.physics.neutronics.isotopicDepletion.isotopicDepletionInterface.CrossSectionTable
	"""

This function has a couple of issues.

The first is its name. The logic in this function is pretty confusing, IMO, but to my understanding it does NOT make a table of reaction rates. It makes a table of 1-group macroscopic cross sections. So the name seems confusing and/or wrong. Please, somebody correct me if I'm wrong here, because tracing this method up the tree was pretty confusing.

The second issue is the inefficiency. For n nuclides in c children of the composite (likely a block) that is passed in, this method will find the core via getAncestorWithFlags n*c times, get the XSID of the via calls to getAncestor().getMicroSuffix() block n*c times, and get the integrated MG flux via calls to getIntegratedMgFlux n*c + 1 times. Each of those operations only need to be called once for a given block. They are being done many extra times because makeReactionRateTable uses getReactionRates to do the work, but getReactionRates uses a helper method which makes all of the aforementioned duplicate calls.

This code was touched during #911 , but it appears that the mentioned inefficiencies were already in place before that, so this has been around for a while.

Kinda-not-really related to #813

[john-science]

I am hoping the isotopicDepletion module will be moved out of ARMI "soon", but if not, this function could certainly use some TLC.