-made all model files use the config file, added defaults to said con…

…fig file

adaboost_model.py

@@ -1,9 +1,17 @@
 __author__ = 'hao yuan'
 from sklearn.ensemble import AdaBoostRegressor
 from sklearn.tree import DecisionTreeRegressor
+import configuration_parser
 
 
 def get():
-    model = AdaBoostRegressor(DecisionTreeRegressor(max_depth=12),
-             n_estimators=275)
-    return model
+    config = configuration_parser.parse()
+    estimators = config.getint(__name__, 'estimators')
+    lr = config.getfloat(__name__, 'learning rate')
+    loss = config.get(__name__, 'loss function')
+    exec('max_depth=' + config.get(__name__, 'max_depth'), locals(), globals())
+    min_samples_split = config.getint(__name__, 'min_samples_split')
+    min_samples_leaf = config.getint(__name__, 'min_samples_leaf')
+    return AdaBoostRegressor(DecisionTreeRegressor(max_depth=max_depth,min_samples_split=min_samples_split,
+                                                   min_samples_leaf=min_samples_leaf),n_estimators=estimators,loss=loss,
+                                                   learning_rate=lr)

default.conf

@@ -25,13 +25,40 @@ test_cases = ${default:test_cases}
 [LeaveOutAlloyCV]
 save_path = ../DBTT/graphs/leaveoutAlloy/{}.png
 
+[dtr_model]
+max_depth = 5
+min_samples_split = 2
+min_samples_leaf = 1
+split criterion = mse
+
 [gkrr_model]
 alpha = 0.00139
 coef0 = 1
 degree = 3
 gamma = 0.518
 kernel = rbf
 
+[lkrr_model]
+alpha = 0.00518
+gamma = 0.518
+kernel = laplacian
+
+[randomforest_model]
+estimators = 100
+max_depth = 5
+min_samples_split = 2
+min_samples_leaf = 1
+max_leaf_nodes = None
+jobs = 1
+
+[adaboost_model]
+estimators = 275
+max_depth = 12
+min_samples_split = 2
+min_samples_leaf = 1
+learning rate = 1
+loss function = linear
+
 #minmax, size, transfer_function are the verbatim arguments for neurolab.net.newff()
 #training_algorithm is the verbatim 'support train fcn' for neurolab.train omitting 'train_'
 #see: https://pythonhosted.org/neurolab/lib.html#module-neurolab.net

dtr_model.py

@@ -0,0 +1,11 @@
+import configuration_parser
+import sklearn.tree as tree
+
+def get():
+	config = configuration_parser.parse()
+	exec('max_depth=' + config.get(__name__, 'max_depth'),locals(),globals())
+	min_samples_split = config.getint(__name__, 'min_samples_split')
+	min_samples_leaf = config.getint(__name__, 'min_samples_leaf')
+	criterion = config.get(__name__, 'split criterion')
+	return tree.DecisionTreeRegressor(criterion=criterion,max_depth=max_depth,min_samples_leaf=min_samples_leaf,
+	                                  min_samples_split=min_samples_split)

lkrr_model.py

@@ -1,8 +1,12 @@
 from sklearn.kernel_ridge import KernelRidge
+import configuration_parser
 
 __author__ = 'haotian'
 
 
 def get():
-
-    return KernelRidge(alpha= .00518, gamma = .518, kernel='laplacian')
+    config = configuration_parser.parse()
+    alpha = config.getfloat(__name__, 'alpha')
+    gamma = config.getfloat(__name__, 'gamma')
+    kernel = config.get(__name__, 'kernel')
+    return KernelRidge(alpha=alpha,gamma=gamma,kernel=kernel)

randomforest_model.py

@@ -1,15 +1,43 @@
 from sklearn.ensemble import RandomForestRegressor
-
+import configuration_parser
+import ast
 __author__ = 'hao yuan'
 
 
 def get():
-    model = RandomForestRegressor(n_estimators=100,
-                               max_features='auto',
-                               max_depth=5,
-                               min_samples_split=2,
-                               min_samples_leaf=1,
-                               min_weight_fraction_leaf=0,
-                               max_leaf_nodes=None,
-                               n_jobs=1)
-    return model
+    config = configuration_parser.parse()
+    estimators = config.getint(__name__, 'estimators')
+    exec('max_depth = ' + config.get(__name__, 'max_depth'),locals(),globals())
+    min_samples_split = config.getint(__name__, 'min_samples_split')
+    min_samples_leaf = config.getint(__name__, 'min_samples_leaf')
+    exec('max_leaf_nodes=' + config.get(__name__, 'max_leaf_nodes'),locals(),globals())
+    jobs = config.getint(__name__, 'jobs')
+
+    model = RandomForestRegressor(n_estimators=estimators,
+                               max_depth=max_depth,
+                               min_samples_split=min_samples_split,
+                               min_samples_leaf=min_samples_leaf,
+                               max_leaf_nodes=max_leaf_nodes,
+                               n_jobs=jobs)
+    return model
+
+def __executeStringNoneOrNumber__(readFromConfig,varAssignment):
+    try:
+        int(readFromConfig)
+        return varAssignment + '=' + readFromConfig  # a number
+    except:
+        try:
+            float(readFromConfig)
+            return varAssignment + '=' + readFromConfig  # a number
+        except:
+            try:
+                if exec(readFromConfig) == None: return varAssignment + '=' + readFromConfig  # None
+            except:
+                return varAssignment + '= \' ' + readFromConfig + '\''  # string
+
+def __executeStringOrInt__(readFromConfig,varAssignment):
+    try:
+        int(readFromConfig)
+        return varAssignment + '=' + readFromConfig  # a number
+    except:
+        return varAssignment + '= \' ' + readFromConfig + '\''  # string