Testing

.gitignore

+__pycache__/*
+model.*
\ No newline at end of file

DataAnalyzer.py

@@ -26,7 +26,6 @@ class DataAnalyzer:
     def applyPCA(self, dataframe: pd.DataFrame):
         pca = PCA(n_components=self.n_components)
         fitted = pca.fit_transform(dataframe[self.metrics])
-
         dataframe[self.axis[:self.n_components]] = pd.DataFrame(fitted, index=dataframe.index)
     
     def applyKMeans(self, dataframe: pd.DataFrame, n_clusters: int) -> KMeans:
@@ -45,26 +44,30 @@ class DataAnalyzer:
     def plot(self, dataframe: pd.DataFrame, labels: List[int], centroids: List[List[float]]):
         if self.n_components == 2:
             ax = plt.subplot(1, 1, 1)
-            scat = ax.scatter(dataframe['x'], dataframe['y'], c=labels, cmap=self.cmap, label=labels)
+            # scat = ax.scatter(dataframe['x'], dataframe['y'], c=labels, cmap=self.cmap, label=labels)
         else:
             fig = plt.figure()
             ax = Axes3D(fig)
-            scat = ax.scatter(dataframe['x'], dataframe['y'], dataframe['z'], c=labels, cmap=self.cmap, label=labels)
-
-        # legend = ax.legend(*scat.legend_elements(), loc='upper right', title='Clusters')
-        # ax.add_artist(legend)
+            # scat = ax.scatter(dataframe['x'], dataframe['y'], dataframe['z'], c=labels, cmap=self.cmap, label=labels)
 
+        #legend = ax.legend(*scat.legend_elements(), loc='upper right', title='Clusters')
+        #ax.add_artist(legend)
+        # , marker='°'
         arr = np.arange(len(centroids))
-        ax.scatter(centroids[:, 0], centroids[:, 1], c=arr, cmap=self.cmap, marker='*', s=500, lw=1.5, edgecolor=(0, 0, 0, 1))
+        ax.scatter(centroids[:, 0], centroids[:, 1], c=arr, cmap=self.cmap, marker='*', s=400, lw=1.5, edgecolor=(0, 0, 0, 1))
+
+        # ax.set_xlim([-1000000, 8000000])
+        # ax.set_ylim([0, 2600])
+        # if (self.n_components == 3):
+        #     ax.set_zlim([-80, 80])
 
         ax.set_xticklabels([])
         ax.set_yticklabels([])
         if (self.n_components == 3):
             ax.set_zticklabels([])
-
         return ax, plt
 
-    def addPatientPoint(self, dataframe: pd.DataFrame, ax: Axes, processInstanceId: str, lastOrKM: bool):
+    def addPatientPoint(self, dataframe: pd.DataFrame, ax: Axes, processInstanceId: str, lastOrKM: bool, kmeans: KMeans):
         #If lastOrKM is true it plots the cluster of all patient's data
         patientData = dataframe[dataframe['processInstanceId'] == processInstanceId]
 
@@ -76,15 +79,26 @@ class DataAnalyzer:
             pointToPlot = patientData.tail(1)[self.axis[:self.n_components]].to_numpy()[0]
 
         if self.n_components == 2:
-            ax.scatter(pointToPlot[0], pointToPlot[1], marker='*', c='black', s=450)
+            ax.scatter(pointToPlot[0], pointToPlot[1], c=kmeans.predict(pointToPlot.reshape(1, -1)), s=250, marker='x')
         else:
-            ax.scatter(pointToPlot[0], pointToPlot[1], pointToPlot[2], marker='*', c='black', s=450)
+            ax.scatter(pointToPlot[0], pointToPlot[1], pointToPlot[2], c=kmeans.predict(pointToPlot.reshape(1, -1)), s=250, marker='x', cmap=self.cmap)
         
         return pointToPlot
 
-    
+    def addAllPatientPoints(self, dataframe: pd.DataFrame, ax: Axes, processInstanceId: str):
+        patientData = dataframe[dataframe['processInstanceId'] == processInstanceId]
 
+        data = patientData[self.axis[:self.n_components]].to_numpy()
         
+        if self.n_components == 2:
+            x = data[:, 0]
+            y = data[:, 1]
 
+            ax.quiver(x[:-1:4], y[:-1:4], x[1::4] - x[:-1:4], y[1::4] - y[:-1:4])
+            #ax.plot(data[:, 0], data[:, 1], c='green')
+        else:
+            x = data[:, 0]
+            y = data[:, 1]
+            z = data[:, 2]
 
-
+            ax.quiver(x[:-1], y[:-1], z[:-1], x[1:] - x[:-1], y[1:] - y[:-1], z[1:] - z[:-1], length=10, normalize=True, arrow_length_ratio=0.9)

main.py

@@ -21,8 +21,8 @@ def main(n_dimensions: int, n_clusters: int, processInstanceId: str, method: str
 
     method = True if method == 'cluster' else False
 
-    point = da.addPatientPoint(df, ax, processInstanceId, method)
-    # da.addAllPatientPoints(df, ax, processInstanceId)
+    # point = da.addPatientPoint(df, ax, processInstanceId, True, kmeans)
+    da.addAllPatientPoints(df, ax, processInstanceId)
     #distance = kmeans.transform(point.reshape(1, -1))
     #normalizedDistance = (distance / np.linalg.norm(distance, ord=2, axis=1, keepdims=True))[0]
     # out = {