Move number of assemblies out of global space

[john-science]

There have been several bugs recently due to strange race conditions and edge cases to do with the fact that the total number of assemblies in a reactor is stored in a global variable:

armi/armi/reactor/assemblies.py

Lines 36 to 44 in 907f2b4

	# to count the blocks that we create and generate a block number
	_assemNum = 0
	def incrementAssemNum():
	global _assemNum # tracked on a module level
	val = _assemNum # return value before incrementing.
	_assemNum += 1
	return val

It appears all of these bugs have been present in the code for years. And in the past few weeks I have had to work on several of these issues.

Since the rest of the codebase is Object Oriented, I see now reason why this one Reactor value shouldn't be part of the Reactor object. And that is my suggestion for fixing it.

[john-science]

Okay, I found what I think is the only reason that we have a global _assemNum:

armi/armi/reactor/assemblies.py

Lines 99 to 100 in 907f2b4

	if assemNum is None:
	assemNum = incrementAssemNum()

When we create a new Assembly object, we want to give it a new, unique assembly number that is only one higher than the next largest assembly number in the Reactor.

[john-science]

Perhaps we could just give each assembly that needs a unique ID one from the linux timestamp:

from time import time

assemNum = int(time() * 1e8)

Then each assembly should have a unique number. And when we need create a reactor, we can go through each assembly and re-number them quickly, starting with zero (and preserving the order).

[john-science]

Also, I feel like the biggest troubles come from giving assemblies number when building a Reactor from a Case Settings file.

It seems like, in that scenario, we can just provide sequential assembly numbers while reading through the CS file.

[john-science]

This was closed with: #1383