Feature/optimiser config

autofit/optimize/grid_search.py

@@ -114,8 +114,10 @@ def __init__(
         self.paths = paths
 
         self.parallel = parallel
-        self.number_of_cores = conf.instance.non_linear.get(
-            "GridSearch", "number_of_cores", int
+        self.number_of_cores = conf.instance.non_linear.config_for(
+            "GridSearch"
+        ).get(
+            "general", "number_of_cores", int
         )
         self.phase_tag_input = paths.phase_tag
 

autofit/optimize/non_linear/multi_nest.py

@@ -51,11 +51,14 @@ def __init__(self, paths=None, sigma=3, run=pymultinest.run):
         self.log_zero = self.config("log_zero", float)
         self.max_iter = self.config("max_iter", int)
         self.init_MPI = self.config("init_MPI", bool)
-        self.terminate_at_acceptance_ratio = conf.instance.non_linear.get(
-            "MultiNest", "terminate_at_acceptance_ratio", bool
+        multinest_config = conf.instance.non_linear.config_for(
+            "MultiNest"
         )
-        self.acceptance_ratio_threshold = conf.instance.non_linear.get(
-            "MultiNest", "acceptance_ratio_threshold", float
+        self.terminate_at_acceptance_ratio = multinest_config.get(
+            "general", "terminate_at_acceptance_ratio", bool
+        )
+        self.acceptance_ratio_threshold = multinest_config.get(
+            "general", "acceptance_ratio_threshold", float
         )
 
         self.run = run
@@ -147,11 +150,14 @@ def prior(cube, ndim, nparams):
 
             return cube
 
-        stagger_resampling_likelihood = conf.instance.non_linear.get(
-            "MultiNest", "stagger_resampling_likelihood", bool
+        multinest = conf.instance.non_linear.config_for(
+            "MultiNest"
+        )
+        stagger_resampling_likelihood = multinest.get(
+            "general", "stagger_resampling_likelihood", bool
         )
-        stagger_resampling_value = conf.instance.non_linear.get(
-            "MultiNest", "stagger_resampling_value", float
+        stagger_resampling_value = multinest.get(
+            "general", "stagger_resampling_value", float
         )
 
         class Fitness:
@@ -162,8 +168,8 @@ def __init__(
                 Fitness function that only handles resampling
                 """
                 self.stagger_accepted_samples = 0
-                self.resampling_likelihood = conf.instance.non_linear.get(
-                    "MultiNest", "null_log_evidence", float
+                self.resampling_likelihood = multinest.get(
+                    "general", "null_log_evidence", float
                 )
 
             def __call__(self, cube, ndim, nparams, lnew):

autofit/optimize/non_linear/non_linear.py

@@ -137,8 +137,11 @@ def config(self, attribute_name, attribute_type=str):
         attribute
             An attribute for the key with the specified type.
         """
-        return conf.instance.non_linear.get(
-            self.__class__.__name__, attribute_name, attribute_type
+        return conf.instance.non_linear.config_for(
+            self.__class__.__name__).get(
+            "general",
+            attribute_name,
+            attribute_type
         )
 
     def __eq__(self, other):

test_autofit/test_files/config/non_linear/DownhillSimplex.ini

@@ -0,0 +1,8 @@
+[general]
+xtol = 1e-4
+ftol = 1e-4
+maxiter = None
+maxfun = None
+full_output = 0
+disp = 1
+retall = 0

test_autofit/test_files/config/non_linear/GridSearch.ini

@@ -0,0 +1,3 @@
+[general]
+number_of_cores = 3
+step_size = 0.1

test_autofit/test_files/config/non_linear.ini → test_autofit/test_files/config/non_linear/MultiNest.ini

@@ -1,4 +1,4 @@
-[MultiNest]
+[general]
 importance_nested_sampling = True
 multimodal = True
 const_efficiency_mode = False
@@ -21,17 +21,4 @@ init_MPI = False
 stagger_resampling_likelihood = False
 stagger_resampling_value = 1000.0
 terminate_at_acceptance_ratio = False
-acceptance_ratio_threshold = 1.0
-
-[DownhillSimplex]
-xtol = 1e-4
-ftol = 1e-4
-maxiter = None
-maxfun = None
-full_output = 0
-disp = 1
-retall = 0
-
-[GridSearch]
-number_of_cores = 3
-step_size = 0.1
+acceptance_ratio_threshold = 1.0