metacountregressor 0.1.347__tar.gz → 1.0.2__tar.gz

{metacountregressor-0.1.347/metacountregressor.egg-info → metacountregressor-1.0.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: metacountregressor
-Version: 0.1.347
+Version: 1.0.2
 Summary: Extensive Testing for Estimation of Data Count Models
 Home-page: https://github.com/zahern/CountDataEstimation
 Author: Zeke Ahern

{metacountregressor-0.1.347 → metacountregressor-1.0.2}/metacountregressor/solution.py

@@ -459,14 +459,12 @@ class ObjectiveFunction(object):
 
         if self.G is not None:
             #TODO need to handle this for groups
-            print('group by groups off for now')
-            NAH_GROUP = 0
-            if NAH_GROUP:
-                self._distribution = ["trad| " + item for item in self._distribution
+            
+           
+            self._distribution = ["trad| " + item for item in self._distribution
                                   ] + ["grpd| " + item for item in self._distribution]
-            else:
-                self._distribution = ["trad| " + item for item in self._distribution
-                                  ]
+        
+                                  
 
         # output information
         self.convergence = None
@@ -1793,23 +1791,40 @@ class ObjectiveFunction(object):
 
         counter = 0
         while b < self._min_characteristics and counter < max_loops:
-            counter +=1
+            counter += 1
             weights = [1 if x == 0 else 0 for x in only_ints_vals]
             get_rdm_i = random.choices(only_ints, weights=weights)[0]
 
-            # todo if only 1 this will get stuck
+            # Safeguard: Check if only one valid value exists to handle the "stuck" case
+            if len(only_ints_vals) == 1 and only_ints_vals[0] == 0:
+                print("Breaking out: only one valid value exists and it is 0.")
+                break
+
             if vector[get_rdm_i] == 0:
+                max_inner_loops = 100  # Limit for the inner loop
+                inner_counter = 0
                 while new_j == 0:
-                    # print('this should only ever be 3 ocasionally 2', self.get_num_discrete_values(get_rdm_i))
-                    get_rdm_j = random.randint(
-                        0, self.get_num_discrete_values(get_rdm_i) - 1)
+                    inner_counter += 1
+                    if inner_counter > max_inner_loops:
+                        print("Breaking out of inner loop: exceeded maximum iterations.")
+                        break
+                    
+                    get_rdm_j = random.randint(0, self.get_num_discrete_values(get_rdm_i) - 1)
                     new_j = self.get_value(get_rdm_i, get_rdm_j)
+
+                # If the inner loop exited early, check and continue
+                if new_j == 0:
+                    print("Inner loop exited without finding a valid value.")
+                    break
+
+                # Assign the value after successfully exiting the inner loop
                 vector[get_rdm_i] = new_j
                 prmVect[get_rdm_i] = new_j
 
                 b += new_j
                 new_j = 0
 
+
         if hasattr(self, 'forced_variables'):
             self.force_inclusion(vector)
 
@@ -2477,7 +2492,7 @@ class ObjectiveFunction(object):
             layout = vector.copy()
 
             while max_trial <= max_routine and trial_run == 1:  # and max_trial <= 200:
-
+                
                 obj_1, model_mod = self.makeRegression(model_nature, layout=layout, **a)
 
                 if obj_best is None:
@@ -7212,7 +7227,7 @@ class ObjectiveFunction(object):
             if len(self.none_handler(zi)) > 0:
                 the_name = the_name + '_zi'
             if model_nature is not None:
-                if 'XG' in model_nature:
+                if 'XG' in model_nature and model_nature.get('XG') is not None:
                     if model_nature.get('XG').shape[2] > 0:
                         the_name = the_name + 'grp'
 
@@ -7285,10 +7300,11 @@ class ObjectiveFunction(object):
         if grouped_rpm is None:
             a =len(self.rdm_cor_fit)+len(self.rdm_fit) != Xr.shape[2]
         else:
-            a = len(self.none_handler(self.rdm_cor_fit)) + len(self.none_handler(self.rdm_fit)) + len(self.none_handler(grouped_rpm)) != Xr.shape[2]
+            a = len(self.none_handler(self.rdm_cor_fit)) + len(self.none_handler(self.rdm_fit)) + len(self.none_handler(grouped_rpm))*self._Gnum != Xr.shape[2]
             
         if a:
             print('why')
+
             print('The number of random effects does not match the data')
             print(Xr.shape)
             print(self.rdm_cor_fit)
@@ -7368,16 +7384,19 @@ class ObjectiveFunction(object):
                                  self.group_dummies.shape[2]) if self.grouped_rpm != [] else []
             X_set = df_tf[:, :, indices4]
             XG = np.tile(self.group_dummies, len(self.grouped_rpm)) * X_set if X_set.shape[2] != 0 else None
+            if XG is not None:
+                model_nature['XG'] = XG
+            
         else:
             XG = None
         X = df_tf[:, :, indices]
         XH = df_tf[:, :, indices5]
 
-        if XG is not None:
+        if XG is not None and self.is_multi:
             indices4_test = np.repeat(self.get_named_indices(self.grouped_rpm),
                                       self.group_dummies_test.shape[2]) if self.grouped_rpm != [] else []
             XGtest = np.tile(self.group_dummies_test, len(self.grouped_rpm)) * df_test[:, :, indices4_test]
-            model_nature['XG'] = XG
+            #model_nature['XG'] = XG
             model_nature['XGtest'] = XGtest
 
         model_nature['X'] = X
@@ -7396,7 +7415,7 @@ class ObjectiveFunction(object):
         if np.isin(X, [np.inf, -np.inf, None, np.nan]).any():  # type ignore
             raise Exception('there is some kind of error in X')
 
-        # numpy data setup fpr estimation
+        # numpy data setup for estimation
         indices2 = self.get_named_indices(self.rdm_fit)
         Xr = df_tf[:, :, indices2]
         if self.rdm_cor_fit is not None:

metacountregressor-1.0.2/metacountregressor/test_code.py

@@ -0,0 +1,43 @@
+from solution import ObjectiveFunction
+from metaheuristics import (harmony_search,differential_evolution,simulated_annealing)         
+#from . import helperprocess
+import pandas as pd
+
+df = pd.read_csv("https://raw.githubusercontent.com/zahern/data/refs/heads/main/rural_int.csv")
+y = df['crashes']  # Frequency of crashes
+
+df.drop(columns=[ 'year', 'orig_ID',
+                                    'jurisdiction', 'town', 'maint_region', 'weather_station', 'dummy_winter_2', 'month', 'inj.fat', 'PDO', 'zonal_ID', 'ln_AADT', 'ln_seg'], inplace=True)  # was dropped postcode
+
+           
+
+arguments_hs = {'_par': 0.3, '_hms': 20}
+arguments = {'test_percentage': 0.2, 'complexity_level': 5, 'reg_penalty':0} #Objective args
+# Step 2: Process Data
+model_terms = {
+    'Y': 'crashes',         # Dependent variable
+    'group': 'county',       # Grouping column (if any)
+    'panels': 'element_ID',      # Panel column (if any)
+    'Offset': None       # Offset column (if any)
+}
+
+
+X = df.drop(columns=['crashes']) # setup X based on data
+X.columns
+print(X.columns)
+
+manual_fit_spec = {
+'fixed_terms': ['const', 'DP10'],
+'rdm_terms':  [ 'DX32:normal'],
+'rdm_cor_terms': [],
+'group_rdm': ['DPO1:triangular'],
+'hetro_in_means': [],
+'transformations': ['no', 'no', 'no', 'no', 'no', 'no'],
+'dispersion': 0
+}
+arguments = {'test_percentage': 0.2, 'complexity_level': 6, 'reg_penalty':0, 'group':'county', 'panels':'element_ID'} #Objective args
+arguments['Manual_Fit'] = manual_fit_spec
+#initial_solution = None
+obj_fun = ObjectiveFunction(X, y, **arguments)
+initial_solution = None
+results_hs = harmony_search(obj_fun, initial_solution, **arguments_hs)

{metacountregressor-0.1.347 → metacountregressor-1.0.2/metacountregressor.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: metacountregressor
-Version: 0.1.347
+Version: 1.0.2
 Summary: Extensive Testing for Estimation of Data Count Models
 Home-page: https://github.com/zahern/CountDataEstimation
 Author: Zeke Ahern

{metacountregressor-0.1.347 → metacountregressor-1.0.2}/metacountregressor.egg-info/SOURCES.txt

@@ -19,6 +19,7 @@ metacountregressor/pareto_logger__plot.py
 metacountregressor/setup.py
 metacountregressor/single_objective_finder.py
 metacountregressor/solution.py
+metacountregressor/test_code.py
 metacountregressor/test_generated_paper2.py
 metacountregressor.egg-info/PKG-INFO
 metacountregressor.egg-info/SOURCES.txt

metacountregressor-1.0.2/version.txt

@@ -0,0 +1 @@
+1.0.2

metacountregressor-0.1.347/version.txt

@@ -1 +0,0 @@
-0.1.347