metacountregressor 0.1.33__py3-none-any.whl → 0.1.35__py3-none-any.whl

metacountregressor/main.py

@@ -18,9 +18,9 @@ import pandas as pd
 import helperprocess
 print('loaded helper')
 
-from metaheuristics import (differential_evolution_non_mp,
-                            harmony_search_non_mp, 
-                            simulated_annealing_non_mp)
+from metaheuristics import (differential_evolution,
+                            harmony_search, 
+                            simulated_annealing)
 print('loaded algorithnms')
 from solution import ObjectiveFunction
 print('loaded solution')
@@ -299,7 +299,7 @@ def main(args):
         'transformations': ['no', 'no', 'no', 'no'],
         'dispersion': 0
         }
-        manual_fit_spec = None
+        #manual_fit_spec = None
         
         
     elif dataset == 8:
@@ -418,7 +418,7 @@ def main(args):
           
                # results = simulated_annealing(obj_fun, 1, 1, None, arguments_hyperparamaters)
             
-            results = simulated_annealing_non_mp(obj_fun, None, **arguments_hyperparamaters)
+            results = simulated_annealing(obj_fun, None, **arguments_hyperparamaters)
             helperprocess.results_printer(results, arguments['algorithm'], int(arguments['is_multi']))
             if dual_complexities:
                 arguments['complexity_level'] = secondary_complexity
@@ -426,7 +426,7 @@ def main(args):
                
                     #results = simulated_annealing(obj_fun, 1, 1, None, arguments_hyperparamaters)
                 
-                results = simulated_annealing_non_mp(obj_fun, None, **arguments_hyperparamaters)
+                results = simulated_annealing(obj_fun, None, **arguments_hyperparamaters)
                 helperprocess.results_printer(results, arguments['algorithm'], int(arguments['is_multi']))
              
         elif arguments['algorithm'] == 'hs':
@@ -435,7 +435,7 @@ def main(args):
             arguments_hyperparamaters = {
                                          'Manual_Fit': arguments['Manual_Fit']}
             
-            results = harmony_search_non_mp(obj_fun, None, **arguments_hyperparamaters)
+            results = harmony_search(obj_fun, None, **arguments_hyperparamaters)
             helperprocess.results_printer(results, arguments['algorithm'], int(arguments['is_multi']))    
             
             if dual_complexities:
@@ -444,12 +444,12 @@ def main(args):
                 #if multi_threaded:
                 #    results = harmony_search(obj_fun, 1, 1, None)
                 #else:
-                results = harmony_search_non_mp(obj_fun, None, **arguments_hyperparamaters)
+                results = harmony_search(obj_fun, None, **arguments_hyperparamaters)
                 helperprocess.results_printer(results, arguments['algorithm'], int(arguments['is_multi']))  
 
                       
         elif arguments['algorithm'] =='de':
-            arguments_hyperparamaters = {'_AI': 1, 
+            arguments_hyperparamaters = {'_AI': 2,
                                         '_crossover_perc': float(arguments['crossover']), 
                                          '_max_iter': int(arguments['_max_imp'])
                                          ,'_pop_size': int(arguments['_hms']), 'instance_number': int(args['line'])
@@ -461,7 +461,7 @@ def main(args):
                 
               #  results = differential_evolution(obj_fun, 1, 1, None, **arguments_hyperparamaters)
            # else:
-            results = differential_evolution_non_mp(obj_fun, None, **arguments_hyperparamaters) 
+            results = differential_evolution(obj_fun, None, **arguments_hyperparamaters) 
             
             helperprocess.results_printer(results, arguments['algorithm'], int(arguments['is_multi'])) 
             if dual_complexities:
@@ -472,7 +472,7 @@ def main(args):
                # results = differential_evolution(obj_fun, 1, 1, None, **arguments_hyperparamaters)
                 
               
-                results = differential_evolution_non_mp(obj_fun, None, **arguments_hyperparamaters) 
+                results = differential_evolution(obj_fun, None, **arguments_hyperparamaters) 
                 
                 helperprocess.results_printer(results, arguments['algorithm'], int(arguments['is_multi'])) 
                   

metacountregressor/metaheuristics.py

@@ -216,7 +216,7 @@ def different_evolution(objective_function, initital_slns=None, **kwargs):
     _plot(x, y, z, 'Iterations', 'Incumbent', 'Best')
     return results
 
-def differential_evolution_non_mp(objective_function, initial_slns=None, **kwargs):
+def differential_evolution(objective_function, initial_slns=None, **kwargs):
     if not isinstance(objective_function, ObjectiveFunction):
         raise Exception
     start = datetime.now()
@@ -321,7 +321,7 @@ def differential_evolution(objective_function, num_processes, num_iterations, in
 
 """
 
-def simulated_annealing_non_mp(objective_function, initial_slns =None, **kwargs):
+def simulated_annealing(objective_function, initial_slns =None, **kwargs):
     #if hyperparameters is None:
     #    TEMP_ALPHA = 0.99; MAX_STEPS = 10; INTL_ACCEPT = 0.5; STEPS = 20; SWAP_PERC = 0.2; NUM_INTL_SLNS = 25; IS_MULTI = 0
     #else:    
@@ -388,7 +388,7 @@ def simulated_annealing(objective_function, num_processes, num_iterations, initi
   
 """
 
-def harmony_search_non_mp(objective_function, initial_harmonies = None, hyperparameters = None, **kwargs):
+def harmony_search(objective_function, initial_harmonies = None, hyperparameters = None, **kwargs):
     """ Worker for hamony search, 1 cpu process
 
     Args:
@@ -524,6 +524,12 @@ class Metaheuristic(object):
         self._obj_fun = objective_function
         self._num_intl_slns = 10
         self._sa_memory = list()
+        
+        self._pop_size = kwargs.get('_pop_size', 20)
+        self.F = kwargs['_AI']  # mustation scale
+        self.iter = kwargs.get('_max_iter', 10000)
+        self.cr = kwargs.get('_crossover_perc') or kwargs.get('_cr', 0.2)
+        self.instance_number = str(kwargs.get('instance_number',1))
         if objective_function.is_multi:
             
             
@@ -625,11 +631,11 @@ class DifferentialEvolution(object):
         if self._obj_fun._obj_1 is None:
             raise Exception
         
-        self._pop_size = kwargs['_pop_size']
-        self.F = kwargs['_AI']  # mustation scale
-        self.iter = kwargs['_max_iter']
-        self.cr = kwargs['_crossover_perc']
-        self.instance_number = str(kwargs['instance_number'])
+        self._pop_size = kwargs.get('_pop_size', 20)
+        self.F = kwargs.get('_AI', 2)  # mustation scale
+        self.iter = kwargs.get('_max_iter', 10000)
+        self.cr = kwargs.get('_crossover_perc') or kwargs.get('_cr', 0.2)
+        self.instance_number = str(kwargs.get('instance_number',1))
         self.get_direcotory()
         
         self._population = list()
@@ -719,7 +725,7 @@ class DifferentialEvolution(object):
                     current_index = self._obj_fun.get_index(i, struct_a[i])
                     new_index = current_index + self.F*(self._obj_fun.get_index(i, struct_b[i]) - self._obj_fun.get_index(i, struct_c[i]))
                 except Exception as e:
-                    print(e, 'Exception here')
+                    print(e, 'Exception here, with the mutation why??')
                     print(struct_a)
                     print(struct_b)
                     print(struct_c)
@@ -757,6 +763,7 @@ class DifferentialEvolution(object):
         repeats = 0
         # generate trial vector by binomial crossover
         unique = True
+        trial = target
         cr = cr_r
         while unique:
             p = np.random.rand(dims)
@@ -846,7 +853,8 @@ class DifferentialEvolution(object):
             
                 i += 1
                 if self.pf.get_objective_is_multi():
-                    self._pareto_population = self.pf.check_if_dominance(self._pareto_population, obj_trial)
+                    is_updated, self._pareto_population = self.pf.check_if_dominance(self._pareto_population, obj_trial, 1)
+                    
                     if len(self._pareto_population) ==1:
                         print('the size of the population is only 1')
                     if self.pf.check_dominance([obj_trial[self.pf.obj_key_1], obj_trial[self.pf.obj_key_2]], [self._population[j][self.pf.obj_key_1], self._population[j][self.pf.obj_key_2]]):#if solution dominates existing #FIXME some error here true but not entering
@@ -1023,17 +1031,17 @@ class SimulatedAnnealing(object):
         self.current_struct = None
         self.temp_max = 1000000
         self.temp_min = 0.05
-        self.MAX_ITERATIONS = int(kwargs['MAX_ITERATIONS'])
-        self.STEPS_PER_TEMP = int(kwargs['STEPS_PER_TEMP'])
-        self.INTL_ACPT = float(kwargs['INTL_ACPT'])
-        self._num_intl_slns = int(kwargs['_num_intl_slns']) 
-        self.alpha = float(kwargs['alpha'])
+        self.MAX_ITERATIONS = int(kwargs.get('MAX_ITERATIONS', 10000)) or int(kwargs.get('_max_iter', 10000))
+        self.STEPS_PER_TEMP = int(kwargs.get('STEPS_PER_TEMP', 2)) or int(kwargs.get('_ts', 2))
+        self.INTL_ACPT = float(kwargs.get('INTL_ACPT', 0.5))
+        self._num_intl_slns = int(kwargs.get('_num_intl_slns', 20)) 
+        self.alpha = float(kwargs.get('alpha', 0.95))
         self.instance_number = str(objective_function.instance_number)
         self.accept = 0
         self.profiler = []
         self.update_t = self.cooling_linear_m
         self.get_direcotory()
-        self._crossover_perc = float(kwargs['_crossover_perc'])
+        self._crossover_perc = float(kwargs.get('_crossover_perc', 0.2)) or float(kwargs.get('_cr', 0.2))
         if objective_function.is_multi: #TODO Define more specific objectives in the intialiser
             self.obj_1 = objective_function._obj_1 
             self.obj_2 = objective_function._obj_2 
@@ -1300,7 +1308,7 @@ class SimulatedAnnealing(object):
                 
                 
             if num_of_changeablePARMs==0:
-                print('fuck line 631')     
+                print('check this line in particular 631')     
         return neighbour
     def _initialize(self, initial_slns=None, model_nature = None):
         """

metacountregressor/pareto_file.py

@@ -15,6 +15,7 @@ class Solution(dict):
         self['aic'] = 10000000.0
         self['bic'] = 10000000.0
         self['MAE'] = 10000000.0
+        self['MSE'] = 10000000.0
         self['layout'] = None
         self['sol_num'] = Solution.sol_counter
         self['num_parm'] = 1000
@@ -37,8 +38,10 @@ class Solution(dict):
         if MAE is not None:
             if MAE < 10000000:
                 self['MAE'] = MAE
+                self['MSE'] = MAE
             else:
                 self['MAE'] = 10000000
+                self['MSE'] = 10000000
         if loglik is not None:
             self['loglik'] = loglik
         if aic is not None:
@@ -148,7 +151,7 @@ class Pareto(object):
                     self.run()
                     return
 
-    def check_if_dominance(self, Structs, New_Struct):
+    def check_if_dominance(self, Structs, New_Struct, return_struct = 0):
         """
         Funtion for non-dominant for newly proposed sln
         ni - the number of solutions which dominate the solution i
@@ -162,10 +165,13 @@ class Pareto(object):
 
         val_key_1 = self.obj_key_1
         val_key_2 = self.obj_key_2
-
+        dominance = False
         if len(Structs) == 0:
             Structs.append(New_Struct)
-            return Structs
+            if return_struct:
+                return True, Structs
+            else:
+                return True
 
         if isinstance(Structs, list):
 
@@ -177,7 +183,7 @@ class Pareto(object):
                     if dominance:
 
                         Structs.append(New_Struct)
-                        Structs = self.pareto_run(Structs)
+                        Structs = self.pareto_run(Structs, True)
                         print('new solution, pareto optimal')
                         break
                 elif self.minimise_1 and not self.minimise_2:  # minimise 1 and maximise 2
@@ -186,7 +192,7 @@ class Pareto(object):
                     if dominance:
 
                         Structs.append(New_Struct)
-                        Structs = self.pareto_run(Structs)
+                        Structs = self.pareto_run(Structs, True)
                         print('new solution, pareto optimal')
                         break
                 elif not self.minimise_1 and self.minimise_2:  # maximise 1 and minise 2
@@ -195,7 +201,7 @@ class Pareto(object):
                     if dominance:
 
                         Structs.append(New_Struct)
-                        Structs = self.pareto_run(Structs)
+                        Structs = self.pareto_run(Structs, True)
                         print('new solution, pareto optimal')
                         break
 
@@ -205,11 +211,14 @@ class Pareto(object):
                     if dominance:
 
                         Structs.append(New_Struct)
-                        Structs = self.pareto_run(Structs)
+                        Structs = self.pareto_run(Structs, True)
                         print('new solution, pareto optimal')
                         break
-
-            return Structs
+            if return_struct:
+                return dominance, Structs
+            else:
+                return dominance        
+          
         else:
             print('not a list')
             if self.minimise_2:
@@ -244,7 +253,10 @@ class Pareto(object):
                         Structs.append(New_Struct)
                         Structs = self.pareto_run(Structs)
                         print('new solution, pareto optimal')
-            return Structs
+            if return_struct:
+                return dominance, Structs
+            else:
+                return dominance      
 
     def get_fronts(self, Struct) -> dict:  # TODO get rid of HM
         """