feat(LDA): autodetect components from PCADataModel

chemfusekit/lda.py

@@ -10,6 +10,7 @@
 from chemfusekit.__utils import graph_output, run_split_test
 from chemfusekit.__utils import print_confusion_matrix, print_table, GraphMode
 from .__base import BaseDataModel, BaseClassifier, BaseSettings
+from .pca import PCADataModel
 
 
 class LDASettings(BaseSettings):
@@ -23,19 +24,20 @@ def __init__(self, components: int = 3, output: GraphMode = GraphMode.NONE, test
 
 class LDA(BaseClassifier):
     '''Class to store the data, methods and artifacts for Linear Discriminant Analysis'''
-    def __init__(self, settings: LDASettings, data_model: BaseDataModel):
-        super().__init__(settings, data_model)
+    def __init__(self, settings: LDASettings, data: BaseDataModel):
+        super().__init__(settings, data)
         self.settings = settings
-        self.x_data = data_model.x_data
-        self.x_train = data_model.x_train
-        self.y = data_model.y
+        self.data = data
+        # Self-detect components if the data is from PCA
+        if isinstance(data, PCADataModel):
+            self.settings.components = data.components - 1
 
     def lda(self):
         '''Performs Linear Discriminant Analysis'''
 
         lda = LD(n_components=self.settings.components) # N-1 where N are the classes
-        scores_lda = lda.fit(self.x_data, self.y).transform(self.x_data)
-        pred = lda.predict(self.x_data)
+        scores_lda = lda.fit(self.data.x_data, self.y).transform(self.data.x_data)
+        pred = lda.predict(self.data.x_data)
 
         print_table(
             [f"LV{i+1}" for i in range(scores_lda.shape[1])],
@@ -72,7 +74,7 @@ def lda(self):
             self.settings.output
         )
 
-        pred = lda.predict(self.x_data)
+        pred = lda.predict(self.data.x_data)
         print_confusion_matrix(
             y1=self.y,
             y2=pred,
@@ -81,13 +83,13 @@ def lda(self):
         )
 
         lv_cols = [f'LV{i+1}' for i in range(self.settings.components)]
-        scores = pd.DataFrame(data = scores_lda, columns = lv_cols) # latent variables
-        scores.index = self.x_data.index
+        scores = pd.DataFrame(data=scores_lda, columns=lv_cols)     # latent variables
+        scores.index = self.data.x_data.index
         y_dataframe = pd.DataFrame(self.y, columns=['Substance'])
 
         scores = pd.concat([scores, y_dataframe], axis = 1)
 
-        # Store the traiend model
+        # Store the trained model
         self.model = lda
 
         # Show graphs if required by the user
@@ -102,7 +104,7 @@ def lda(self):
         if self.settings.test_split:
             run_split_test(
                 scores.drop('Substance', axis=1).values,
-                self.y,
+                self.data.y,
                 LD(n_components=self.settings.components),
                 mode=self.settings.output
             )

chemfusekit/pca.py

@@ -16,9 +16,11 @@
 
 class PCADataModel(BaseDataModel):
     '''Data model for the PCA outputs.'''
-    def __init__(self, x_data: pd.DataFrame, x_train: pd.DataFrame, y: np.ndarray, array_scores: np.ndarray):
+    def __init__(self, x_data: pd.DataFrame, x_train: pd.DataFrame, y: np.ndarray, array_scores: np.ndarray,
+                 components: int):
         super().__init__(x_data, x_train, y)
         self.array_scores = array_scores
+        self.components = components
 
 
 class PCASettings:
@@ -276,5 +278,6 @@ def export_data(self) -> PCADataModel:
             self.data.x_data,
             self.data.x_train,
             self.data.y,
-            self.array_scores
+            self.array_scores,
+            self.components
         )

docs/docs/pca/pcadatamodel.md

@@ -11,7 +11,7 @@ It inherits from the [`BaseDataModel`](../base/basedatamodel.md).
 ## Syntax
 
 ```python
-PCAModel(x_data: pd.DataFrame, x_train: pd.DataFrame, y: np.ndarray, array_scores: np.ndarray)
+PCAModel(x_data: pd.DataFrame, x_train: pd.DataFrame, y: np.ndarray, array_scores: np.ndarray, components: int)
 ```
 
 ## Fields and constructor parameters
@@ -20,6 +20,9 @@ The first two are `Pandas` `DataFrame` objects:
 - `x_data`
 - `x_train`
 
-The last two are `NumPy` `ndarray`s:
+The second two are `NumPy` `ndarray`s:
 - `y`
 - `array_scores`
+
+The last is an integer:
+- `components`

docs/versioned_docs/version-2.0.0/pca/pcadatamodel.md

@@ -11,7 +11,7 @@ It inherits from the [`BaseDataModel`](../base/basedatamodel.md).
 ## Syntax
 
 ```python
-PCAModel(x_data: pd.DataFrame, x_train: pd.DataFrame, y: np.ndarray, array_scores: np.ndarray)
+PCAModel(x_data: pd.DataFrame, x_train: pd.DataFrame, y: np.ndarray, array_scores: np.ndarray, components: int)
 ```
 
 ## Fields and constructor parameters
@@ -20,6 +20,9 @@ The first two are `Pandas` `DataFrame` objects:
 - `x_data`
 - `x_train`
 
-The last two are `NumPy` `ndarray`s:
+The second two are `NumPy` `ndarray`s:
 - `y`
 - `array_scores`
+
+The last is an integer:
+- `components`

examples/pca_lda_notebook.ipynb

@@ -112,7 +112,10 @@
     "\n",
     "# Print the number of components and the statistics\n",
     "print(f\"\\nNumber of components: {pca.components}\\n\")\n",
-    "pca.pca_stats()"
+    "pca.pca_stats()\n",
+    "\n",
+    "# Export data from PCA\n",
+    "pca_data = pca.export_data()"
    ]
   },
   {
@@ -134,13 +137,12 @@
     "from chemfusekit.lda import LDASettings, LDA, GraphMode\n",
     "\n",
     "settings = LDASettings(\n",
-    "    components=(pca.components - 1),    # one less component than the number determined by PCA\n",
-    "    output=GraphMode.GRAPHIC,   # graphs will be printed as pictures\n",
-    "    test_split=True # Run split test\n",
+    "    output=GraphMode.GRAPHIC    # Graphs will be printed\n",
+    "    test_split=True     # Run split test\n",
     ")\n",
     "\n",
     "# Initialize and run the LDA class\n",
-    "lda = LDA(settings, lldf.fused_data)\n",
+    "lda = LDA(settings, pca_data)   # components will be determined automatically from the PCA data\n",
     "lda.lda()"
    ]
   },