aggregate_reactions hangs

[KRiedmiller]

For homolysis with a very big system, aggregate_reactions in the recipe_collection freezes KIMMDY for unknown duration.

Workaround is to remove the aggregate_reactions call in runmanager._decide_recipe and to not plot the rates.
For homolysis (which uses rf kmc), this should be no issue, but for HAT and extrande it might. Needs more investigation.

[jmbuhr]

How big is very big?

[jmbuhr]

Looks like this part

kimmdy/src/kimmdy/recipe.py

Lines 373 to 383 in 09f364d

	unique_recipes = []
	unique_recipes_ixs = []
	for i, recipe in enumerate(self.recipes):
	if recipe.recipe_steps not in unique_recipes:
	unique_recipes.append(recipe.recipe_steps)
	unique_recipes_ixs.append([i])
	else:
	for j, ur in enumerate(unique_recipes):
	if recipe.recipe_steps == ur:
	unique_recipes_ixs[j].append(i)

could be more efficient. We are iterating over the unique recipes multiple times. If we make recipe steps hashable we can check for uniqueness easier.

[KRiedmiller]

In fact, I had experimented with making them hashable, but decided that it looked too hacky :D Could be the better solution after all.
It fails for our big riken collagen system.

[jmbuhr]

Ah, I see! That is very big indeed.