Tested and fixed EULA code.

nnmm/mods/mortality.py

@@ -37,7 +37,8 @@ def init( model ):
  # Find index of first value after eula_age
  split_index = np.searchsorted(model.population.age, eula_age)
  # Get list of indices we would EULA-ify
- dod_filtered = model.population.dod[0:split_index]
+ # If pop is age order, lowest to highest, we keep the "left chunk" and eula-ify the middle chunk
+ dod_filtered = model.population.dod[split_index:model.population.count]
  # Convert these all to "simulation year of death"
  death_years = dod_filtered // 365
  # Now invoke function in Population class to calculate the expected deaths by simulation year
@@ -88,7 +89,11 @@ def do_non_disease_deaths(model, tick):
  year = tick // 365
  if model.population.expected_new_deaths_per_year is not None:
  for nodeid in range(len(model.population.expected_new_deaths_per_year)):
- model.nodes.population[nodeid,tick+1] -= (model.population.expected_new_deaths_per_year[nodeid][year]/365)
+ # converting yearly number to daily basically sucks
+ deaths = (model.population.expected_new_deaths_per_year[nodeid][year]/365)
+ model.nodes.population[nodeid,tick+1] -= deaths 
+ # Reporting only
+ model.nodes.deaths[nodeid,year] += deaths
 
  pq = model.nddq
  while len(pq) > 0 and pq.peekv() <= tick:

src/idmlaser/numpynumba/population.py

@@ -327,7 +327,7 @@ def expected_deaths_over_sim(self, death_years, split_index, sim_years=10):
  death_years[death_years > sim_years] = sim_years # Cap deaths at years-1
 
  # Initialize the expected_new_deaths array
- nodeids_filtered = self.nodeid[0:split_index]
+ nodeids_filtered = self.nodeid[split_index:self.count]
  unique_nodeids = np.unique(nodeids_filtered) # slow way of calculating node count
  nodeid_indices = {nodeid: i for i, nodeid in enumerate(unique_nodeids)}
  self.expected_new_deaths_per_year = np.zeros((len(unique_nodeids), sim_years+1), dtype=int)