fix(pca): field and property issues

chemfusekit/__base.py

@@ -170,7 +170,6 @@ class BaseReducer(BaseActionClass):
 
     def __init__(self, settings: BaseSettings, data: BaseDataModel):
         super().__init__(settings, data)
-        self.rescaled_data = None
 
     @abstractmethod
     def export_data(self) -> BaseDataModel:
@@ -193,4 +192,8 @@ def reduce(self, data: BaseDataModel) -> BaseDataModel:
             x_train=x_train,
             y=data.y
         )
-
+
+    @property
+    @abstractmethod
+    def rescaled_data(self) -> BaseDataModel:
+        pass

chemfusekit/lda.py

@@ -1,5 +1,6 @@
 '''Linear Discriminant Analysis module'''
 from copy import copy
+from functools import cached_property
 from typing import Optional
 
 import joblib
@@ -8,7 +9,6 @@
 
 from sklearn.discriminant_analysis import LinearDiscriminantAnalysis as LD
 
-from chemfusekit.lldf import LLDFDataModel
 from chemfusekit.__utils import graph_output, run_split_test
 from chemfusekit.__utils import print_confusion_matrix, print_table, GraphMode
 from .__base import BaseDataModel, BaseClassifier, BaseClassifierSettings, BaseReducer
@@ -139,3 +139,25 @@ def export_data(self) -> LDADataModel:
             y=self.data.y,
             components=self.settings.components
         )
+
+    @cached_property
+    def rescaled_data(self) -> BaseDataModel:
+        if self.model is None:
+            settings_backup = copy(self.settings)
+            self.settings.output = GraphMode.NONE
+            self.settings.test_split = False
+            self.lda()
+            self.settings = settings_backup
+
+        x_data = pd.DataFrame(self.model.transform(self.data.x_data))
+        y_dataframe = pd.DataFrame(self.data.y, columns=['Substance'])
+        x_train = pd.concat(
+            [y_dataframe, x_data],
+            axis=1
+        )
+
+        return BaseDataModel(
+            x_data,
+            x_train,
+            self.data.y
+        )

chemfusekit/pca.py

@@ -1,4 +1,5 @@
 '''Principal Component Analysis Module'''
+import functools
 from copy import copy
 from functools import cached_property
 from typing import Optional
@@ -48,7 +49,7 @@ class PCA(BaseReducer):
     def __init__(self, settings: PCASettings, data: BaseDataModel):
         super().__init__(settings, data)
         self.components = 0
-        self.pca_model: Optional[PC] = None
+        self.model: Optional[PC] = None
         self.array_scores: Optional[np.ndarray] = None
 
     def pca(self):
@@ -108,10 +109,10 @@ def pca(self):
             plt.ylabel('Cumulative Proportional Variance Explained')
             plt.show()
 
-        # Run PCA producing the pca_model with a proper number of components
+        # Run PCA producing the model with a proper number of components
         pca = PC(n_components=self.components)
-        self.pca_model = pca
-        self.pca_model.fit(x_data)
+        self.model = pca
+        self.model.fit(x_data)
 
     def pca_stats(self):
         '''Produces PCA-related statistics.'''
@@ -120,28 +121,28 @@ def pca_stats(self):
 
         # Prepare the Scores dataframe (and concatenate the original 'Region' variable)
         pc_cols = [f"PC{i+1}" for i in range(self.components)]
-        scores = pd.DataFrame(data=self.pca_model.fit_transform(x_data), columns=pc_cols)
+        scores = pd.DataFrame(data=self.model.fit_transform(x_data), columns=pc_cols)
         scores.index = x_data.index
-        scores = pd.concat([scores, x_train.Substance], axis = 1)
+        scores = pd.concat([scores, x_train.Substance], axis=1)
 
         print_table(
             pc_cols + ['Substance'],
-            [scores.iloc[:,i] for i in range(scores.shape[1])],
+            [scores.iloc[:, i] for i in range(scores.shape[1])],
             "PCA scores for each component",
             self.settings.output
         )
 
         # Prepare the loadings dataframe
         loadings = pd.DataFrame(
-            self.pca_model.components_.T,
+            self.model.components_.T,
             columns=pc_cols,
             index=x_data.columns
         )
         loadings["Attributes"] = loadings.index
 
         print_table(
             pc_cols + ['Retention Time'],
-            [loadings.iloc[:,i] for i in range(loadings.shape[1])],
+            [loadings.iloc[:, i] for i in range(loadings.shape[1])],
             "PCA Loadings",
             self.settings.output
         )
@@ -177,9 +178,9 @@ def pca_stats(self):
             fig.show()
 
         # Get PCA scores
-        t = scores.iloc[:,0:self.components]
+        t = scores.iloc[:, 0:self.components]
         # Get PCA loadings
-        p = loadings.iloc[:,0:self.components]
+        p = loadings.iloc[:, 0:self.components]
         # Calculate error array
         err = x_data - np.dot(t,p.T)
         # Calculate Q-residuals (sum over the rows of the error array)
@@ -206,7 +207,7 @@ def mean_confidence_interval(data, confidence=self.settings.confidence_level):
         # Create a dataframe using only T2 and Q-residuals
         hot_q_data = pd.DataFrame(
             {'T2': tsq, 'Qres': q, 'Substance': x_train.Substance},
-            index = x_data.index
+            index=x_data.index
         )
 
         if self.settings.output is GraphMode.GRAPHIC:
@@ -216,7 +217,7 @@ def mean_confidence_interval(data, confidence=self.settings.confidence_level):
                 x="T2",
                 y="Qres",
                 hover_data={'Sample': (hot_q_data.index)},
-                color = "Substance"
+                color="Substance"
             )
             fig.add_hline(y=abs(q_conf),line_dash="dot", line_color='Red')
             fig.add_vline(x=tsq_conf,line_dash="dot", line_color='Red')
@@ -264,7 +265,7 @@ def mean_confidence_interval(data, confidence=self.settings.confidence_level):
 
         print_table(
             pc_cols,
-            [array_scores[:,i] for i in range(array_scores.shape[1])],
+            [array_scores[:, i] for i in range(array_scores.shape[1])],
             "Array without 'Substance' column",
             self.settings.output
         )
@@ -273,7 +274,7 @@ def mean_confidence_interval(data, confidence=self.settings.confidence_level):
 
     def export_data(self) -> PCADataModel:
         '''Export data artifacts.'''
-        if self.pca_model is None or self.array_scores is None:
+        if self.model is None or self.array_scores is None:
             raise RuntimeError("Run both pca() and pca_stats() methods before exporting data!")
 
         return PCADataModel(
@@ -286,15 +287,14 @@ def export_data(self) -> PCADataModel:
 
     @cached_property
     def rescaled_data(self) -> PCADataModel:
-        if self.array_scores is None:
+        if self.model is None:
             settings_backup = copy(self.settings)
             self.settings.output = GraphMode.NONE
-            if self.pca_model is None:
-                self.pca()
+            self.pca()
             self.pca_stats()
             self.settings = settings_backup
 
-        x_data = pd.DataFrame(self.pca_model.transform(self.data.x_data))
+        x_data = pd.DataFrame(self.model.transform(self.data.x_data))
         y_dataframe = pd.DataFrame(self.data.y, columns=['Substance'])
         x_train = pd.concat(
             [y_dataframe, x_data],
@@ -324,7 +324,7 @@ def from_file(cls, settings: PCASettings, model_path: str):
             np.asarray(pd.DataFrame)
         )
         class_instance = PCA(settings, data)
-        class_instance.pca_model = model
+        class_instance.model = model
         return class_instance
 
     def import_model(self, import_path: str):
@@ -336,7 +336,7 @@ def import_model(self, import_path: str):
 
     def export_model(self, export_path: str):
         '''Exports the underlying sklearn PCA model to a file.'''
-        if self.pca_model is not None:
-            joblib.dump(self.pca_model, export_path)
+        if self.model is not None:
+            joblib.dump(self.model, export_path)
         else:
             raise RuntimeError("You haven't trained the model yet! You cannot export it now.")

docs/docs/pca/pca.md

@@ -23,7 +23,7 @@ PCA(settings: PCASettings, data: BaseDataModel)
 
 - `fused_data`: object of type [`LLDF`](../lldf/lldf-class.md). Contains the data to be analyzed.
 - `components`: Number of components for the PCA analysis. Defaults to 0.
-- `pca_model`: A `PCA` model from `scikit-learn`. Defaults to `None`.
+- `model`: A `PCA` model from `scikit-learn`. Defaults to `None`.
 - `settings`: object of type [`PCASettings`](./pcasettings.md). Contains the settings for
   the `PCA` object. 
 

docs/versioned_docs/version-2.5.0/pca/pca.md

@@ -23,7 +23,7 @@ PCA(settings: PCASettings, data: BaseDataModel)
 
 - `fused_data`: object of type [`LLDF`](../lldf/lldf-class.md). Contains the data to be analyzed.
 - `components`: Number of components for the PCA analysis. Defaults to 0.
-- `pca_model`: A `PCA` model from `scikit-learn`. Defaults to `None`.
+- `model`: A `PCA` model from `scikit-learn`. Defaults to `None`.
 - `settings`: object of type [`PCASettings`](./pcasettings.md). Contains the settings for
   the `PCA` object. 
 

tests/test_pca.py

@@ -192,7 +192,7 @@ def test_pca_import_export(self):
         pca2 = PCA.from_file(pca_settings, 'pca_model.sklearn')
 
         # Assert the equality of the two models
-        self.assertEqual(pca.pca_model.get_params(), pca2.pca_model.get_params())
+        self.assertEqual(pca.model.get_params(), pca2.model.get_params())
 
     def test_pca_reduce(self):
         '''Test case for data dimensionality reduction.'''