Building dataframe from geoseries instead of list and dtype fixed

scenariosAleatoires/MCMC.py

@@ -99,6 +99,7 @@ class MCMC:
         :param scenarios: (ScenariosStack) list of the scenarios used as base for sampling the new scenarios.
         '''
         # Loop of sampling and scoring
+        start_time = time.time()
         data={'scenarios_patches':[], 'scenarios_cult':[], 'previous_indexes':[]}
         for index, scenario in tqdm(scenarios.sample(nb_particles, self.rng).iterrows(), total=nb_particles, disable=disable_progress):
             scenario = scenarios.reallocate(index)
@@ -106,12 +107,11 @@ class MCMC:
                 scenarios.setFullReallocated(index, True)
             else:
                 data['scenarios_patches'].append(scenario)
-                data['scenarios_cult'].append(scenario['cultgeopat'].to_list())
+                data['scenarios_cult'].append(scenario['cultgeopat'])
                 data['previous_indexes'].append(index)
-        start_time = time.time()
         scenarios.reconstitute(iter_nb, **data)
         elapsed_time = time.time() - start_time
-        print('scenarios evaluated and stored in {} - '.format(time.strftime("%M:%S", time.gmtime(elapsed_time))), end="", flush=True)
+        print('Iteration duration: {} - {:.2f} scen/s '.format(time.strftime("%M:%S", time.gmtime(elapsed_time)), nb_particles/elapsed_time), end="", flush=True)
         # Computing pareto front
         start_time = time.time()
         pareto_mask = MCMC.is_pareto_efficient(scenarios.scores[['Resilience','Proximity','Productivity','Biodiversity','Social','TargetDelta']].values)

scenariosAleatoires/scenarios.py

@@ -72,7 +72,7 @@ class ScenariosStack:
 
     def reconstitute(self, iter_nb, scenarios_patches, scenarios_cult, previous_indexes, disable_progress=False):
         indexes = [next(self._counter) for id in range(len(scenarios_cult))]
-        self.cultgeopat = pd.DataFrame(scenarios_cult, index=indexes, columns=self.cultgeopat.columns)
+        self.cultgeopat = pd.DataFrame(scenarios_cult, index=indexes, columns=self.cultgeopat.columns, dtype='int8')
         list_scores = []
         for i, previous_index in tqdm(enumerate(previous_indexes), total=len(previous_indexes), disable=disable_progress):
             scenario_patch = scenarios_patches[i]