Move CalculateScoresBetweenAllIonmodes from plotting_wrapper_function…

…s.py to training_wrapper_functions.py and use fingerprint settings
matchms · Oct 30, 2024 · caf0804 · caf0804
1 parent 6af69ee
commit caf0804
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 22 deletions.
diff --git a/ms2deepscore/wrapper_functions/plotting_wrapper_functions.py b/ms2deepscore/wrapper_functions/plotting_wrapper_functions.py
@@ -7,32 +7,23 @@
 from ms2deepscore.benchmarking.plot_loss_per_bin import plot_loss_per_bin_multiple_benchmarks
 
 
-def create_plots_between_ionmodes(positive_validation_spectra,
-                                  negative_validation_spectra,
-                                  model_file_name: str,
+def create_plots_between_ionmodes(scores_between_all_ionmodes: CalculateScoresBetweenAllIonmodes,
                                   results_folder: str,
                                   nr_of_bins=50):
     """
     Creates comparison plots between different ion modes.
 
     Parameters
     ----------
-    positive_validation_spectra:
-        The validation spectra in positive ionisation mode
-    negative_validation_spectra:
-        The validation spectra in negative ionistation mode
-    model_file_name:
-        MS2Deepscore model file name
+    scores_between_all_ionmodes:
+        CalculateScoresBetweenAllIonmodes object containing predictions and tanimoto scores between ionmodes.
     results_folder:
         Directory where the plots will be saved.
     nr_of_bins:
         Nr of tanimoto score bins to use for plotting.
     """
     os.makedirs(results_folder, exist_ok=True)
 
-    scores_between_all_ionmodes = CalculateScoresBetweenAllIonmodes(model_file_name, positive_validation_spectra,
-                                                                    negative_validation_spectra)
-
     plot_loss_per_bin_multiple_benchmarks(scores_between_all_ionmodes.list_of_predictions_and_tanimoto_scores(),
                                           nr_of_bins=nr_of_bins,
                                           loss_type="RMSE")

diff --git a/ms2deepscore/wrapper_functions/training_wrapper_functions.py b/ms2deepscore/wrapper_functions/training_wrapper_functions.py
@@ -64,14 +64,16 @@ def train_ms2deepscore_wrapper(spectra_file_path,
     ms2ds_history_plot_file_name = os.path.join(results_folder, settings.history_plot_file_name)
     plot_history(history["losses"], history["val_losses"], ms2ds_history_plot_file_name)
 
+    scores_between_all_ionmodes = CalculateScoresBetweenAllIonmodes(
+        os.path.join(stored_training_data.trained_models_folder, model_directory_name, settings.model_file_name),
+        stored_training_data.load_training_data("positive", "validation"),
+        stored_training_data.load_training_data("negative", "validation"),
+        settings.fingerprint_type, settings.fingerprint_nbits)
+
     # Create performance plots for validation spectra
-    create_plots_between_ionmodes(positive_validation_spectra=stored_training_data.load_training_data("positive", "validation"),
-                                  negative_validation_spectra=stored_training_data.load_training_data("negative", "validation"),
+    create_plots_between_ionmodes(scores_between_all_ionmodes,
                                   results_folder=os.path.join(stored_training_data.trained_models_folder,
                                                               model_directory_name, "benchmarking_results"),
-                                  model_file_name=os.path.join(stored_training_data.trained_models_folder,
-                                                               model_directory_name,
-                                                               settings.model_file_name),
                                   nr_of_bins=50)
     return model_directory_name
 
@@ -179,8 +181,8 @@ def parameter_search(
 
         scores_between_all_ionmodes = CalculateScoresBetweenAllIonmodes(
             model_file_name=os.path.join(stored_training_data.trained_models_folder,
-                                                   model_directory_name,
-                                                   settings.model_file_name),
+                                         model_directory_name,
+                                         settings.model_file_name),
             positive_validation_spectra=positive_validation_spectra,
             negative_validation_spectra=negative_validation_spectra)
 
@@ -192,9 +194,9 @@ def parameter_search(
         for loss_type in loss_types:
             losses_per_ionmode = {}
             for predictions_and_tanimoto_scores in scores_between_all_ionmodes.list_of_predictions_and_tanimoto_scores():
-                    _, losses = predictions_and_tanimoto_scores.get_average_loss_per_bin_per_inchikey_pair(loss_type,
-                                                                                                           settings.same_prob_bins)
-                    losses_per_ionmode[predictions_and_tanimoto_scores.label] = losses
+                _, losses = predictions_and_tanimoto_scores.get_average_loss_per_bin_per_inchikey_pair(loss_type,
+                                                                                                       settings.same_prob_bins)
+                losses_per_ionmode[predictions_and_tanimoto_scores.label] = losses
             combination_results["losses"][loss_type] = losses_per_ionmode
 
         # Store results