Merge pull request #209 from EhsanGharibNezhad/main

update clean-codes from main

TelescopeML/IO_utils.py

@@ -100,8 +100,8 @@ def load_or_dump_trained_model_CNN(
     # weights_path = self.create_generic_path(f'{indicator}_weights')
 
     if load_or_dump == 'dump':
-        trained_model.trained_model.save(path_architecture)
-        trained_model.trained_model.save_weights(path_weights)
+        trained_model.trained_ML_model.save(path_architecture)
+        trained_model.trained_ML_model.save_weights(path_weights)
 
         with open(path_history, 'wb') as file:
             pk.dump(trained_model.history.history, file)

TelescopeML/StatVisAnalyzer.py

@@ -1154,7 +1154,14 @@ def plot_scatter_x_y (x, y,
     # Show the plot
     show(p)
 
-def plot_filtere_data(dataset, filter_bounds, feature_to_plot, title_label, wl_synthetic,output_names, __reference_data__):
+def plot_filtere_data(dataset,
+                      filter_bounds,
+                      feature_to_plot,
+                      title_label,
+                      wl_synthetic,
+                      output_names,
+                      __reference_data__,
+                      __save_plots__=False):
     """
     Plot a DataFrame with a single x-axis (using column names) and multiple y-axes.
 
@@ -1189,7 +1196,7 @@ def plot_filtere_data(dataset, filter_bounds, feature_to_plot, title_label, wl_s
 
     # print(filtered_data.T[col][:4].values[0])
     ax.set_xlabel('Wavelength [$\mu$m]', fontsize = 12)
-    ax.set_ylabel(r'F$_{\nu}$  [erg/cm$^2$/s/Hz]', fontsize = 12)
+    ax.set_ylabel(r'TOA F$_{\nu}^{\rm Syn}$  [erg/cm$^2$/s/Hz]', fontsize = 12)
     dict_features = {'temperature': 'Effective Temperature', 'gravity': 'Gravity', 'metallicity': 'Metallicity',
                      'c_o_ratio': 'Carbon-to-oxygen ratio'}
     ax.set_title(dict_features[feature_to_plot] + " " + title_label, fontsize = 14)
@@ -1208,8 +1215,9 @@ def plot_filtere_data(dataset, filter_bounds, feature_to_plot, title_label, wl_s
     dict_features = {'temperature': 'T$_{eff}$ [K]', 'gravity': 'log$g$', 'metallicity': '[M/H]', 'c_o_ratio': 'C/O'}
     cbar.set_label(dict_features[feature_to_plot], fontsize = 12)
 
-    # plt.savefig(os.path.join(__reference_data__, 'figures', feature_to_plot + "_trainin_examples.pdf"), dpi=500,
-    #             bbox_inches='tight')
+    if __save_plots__:
+        plt.savefig(os.path.join(__reference_data__, 'figures', feature_to_plot + "_training_examples.pdf"), dpi=500,
+                    bbox_inches='tight')
 
     plt.show()
 
@@ -1325,7 +1333,8 @@ def plot_ML_model_loss(trained_ML_model_history=None, title=None):
 
 def plot_boxplot(data,
                  title=None, xlabel='Wavelength [$\mu$m]', ylabel='Scaled Values',
-                 xticks_list=None, fig_size=(14, 3)):
+                 xticks_list=None, fig_size=(14, 3),
+                 __save_plots__=False):
     """
     Make a boxplot with the scaled features.
 
@@ -1363,6 +1372,11 @@ def plot_boxplot(data,
         ax.set_xticks(xtick_positions[::i])
         ax.set_xticklabels(xticks_list[::i])
 
+    # if __save_plots__:
+    #     plt.savefig(os.path.join(__reference_data__, 'figures', feature_to_plot + "_training_examples.pdf"),
+    #                 dpi=500,
+    #                 bbox_inches='tight')
+
     plt.tight_layout()
     plt.show()
 

TelescopeML/__version__.py

@@ -1,3 +1,7 @@
 # %% VERSIONS
 # Default/Latest/Current version
-__version__ = '0.0.3'
+__version__ = '0.0.4'
+# Cleaned the codes and included more specific names for the parameters
+# Edited the notebooks and transferred all funcs into the StatVisAnalyzer module
+# Edited the name of functions in StatVisAnalyzer
+# Added a module named DeepTuner to BOHB tuning the DNN hyperparameters

setup.cfg

@@ -1,7 +1,7 @@
 [metadata]
 
 name = TelescopeML
-version = 0.0.3
+version = 0.0.4
 summary = Deep Convolutional Neural Networks and Machine Learning Models for Analyzing Stellar and Exoplanetary Telescope Spectra
 description_file = README.md
 keywords = Astronomy, Stars, Telescope, CNN, Deep Learning, Machine Learning