modify tutorial notebooks

setup.py

@@ -4,7 +4,7 @@
 from setuptools import setup, find_packages
 with open("README.md", "r") as fh:
     long_description = fh.read()
-__version__ = "0.1.29"
+__version__ = "0.1.30"
 
 setup(
     name='emsigma',

sigma/gui/gui.py

@@ -5,7 +5,7 @@
 import random
 import numpy as np
 import pandas as pd
-from typing import List
+from typing import List, Dict
 import hyperspy.api as hs
 from matplotlib import pyplot as plt
 import matplotlib as mpl
@@ -367,28 +367,42 @@ def dropdown_b_eventhandler(change):
     display(plots_output)
 
 
-def view_pixel_distributions(sem, norm_list=[], peak="Fe_Ka", cmap="viridis"):
-    out = widgets.Output()
-    with out:
+def view_pixel_distributions(sem, norm_list=[], cmap="viridis"):
+    peak_options = sem.feature_list
+    dropdown_peaks = widgets.Dropdown(options=peak_options, description="Element:")
+
+    plots_output = widgets.Output()
+
+    with plots_output:
         fig = visual.plot_pixel_distributions(
-            sem=sem, norm_list=norm_list, peak=peak, cmap=cmap
-        )
+            sem=sem, norm_list=norm_list, peak=dropdown_peaks.value, cmap=cmap
+            )
         plt.show()
-
-    out_box = widgets.Box([out])
+        save_fig(fig)
+
+    def dropdown_option_eventhandler(change):
+        plots_output.clear_output()
+        with plots_output:
+            fig = visual.plot_pixel_distributions(
+            sem=sem, norm_list=norm_list, peak=dropdown_peaks.value, cmap=cmap
+            )
+            plt.show()
+            save_fig(fig)
+
+    dropdown_peaks.observe(dropdown_option_eventhandler, names="value")
+    out_box = widgets.VBox([dropdown_peaks, plots_output])    
     display(out_box)
-    save_fig(fig)
 
 
 def view_intensity_maps(edx, element_list):
     pick_color(visual.plot_intensity_maps, edx=edx, element_list=element_list)
 
 
 def view_bic(
-    latent,
-    n_components=20,
-    model="BayesianGaussianMixture",
-    model_args={"random_state": 6},
+    latent: np.ndarray,
+    n_components: int = 20,
+    model: str = "BayesianGaussianMixture",
+    model_args: Dict = {"random_state": 6},
 ):
     bic_list = PixelSegmenter.bic(latent, n_components, model, model_args)
     fig = go.Figure(

sigma/src/dim_reduction.py

@@ -1,6 +1,7 @@
 import os
 import torch
 import datetime
+from typing import Dict
 import numpy as np
 import torch.nn as nn
 from torch.optim import Adam
@@ -73,22 +74,22 @@ def set_up_results_dirs(self):
     def run_model(
         self,
         num_epochs: int,
-        patience: int,
         batch_size: int,
-        learning_rate=1e-4,
-        weight_decay=0.0,
-        task="train_eval",
+        patience: int = 50,
+        learning_rate: float = 1e-4,
+        weight_decay: float = 0.0,
+        task: str = "train_eval",
         noise_added=None,
-        criterion="MSE",
-        KLD_lambda=1e-4,
-        print_latent=False,
-        lr_scheduler_args={
-            "factor": 0.5,
-            "verbose": True,
-            "patience": 5,
-            "threshold": 1e-2,
-            "min_lr": 1e-7,
-        },
+        criterion: str = "MSE",
+        KLD_lambda:float = 1e-4,
+        print_latent: bool = False,
+        lr_scheduler_args: Dict={
+                            "factor": 0.5,
+                            "verbose": True,
+                            "patience": 5,
+                            "threshold": 1e-2,
+                            "min_lr": 1e-7,
+                        },
     ):
         # Loss functions
         if criterion == "MSE":

sigma/src/segmentation.py

@@ -4,6 +4,7 @@
 from sigma.utils.load import SEMDataset
 from sigma.utils.visualisation import make_colormap
 
+from typing import Dict
 import hyperspy.api as hs
 import numpy as np
 import pandas as pd
@@ -23,11 +24,11 @@
 class PixelSegmenter(object):
     def __init__(
         self,
-        latent: np,
-        dataset_norm: np,
+        latent: np.ndarray,
+        dataset_norm: np.ndarray,
         sem: SEMDataset,
-        method="BayesianGaussianMixture",
-        method_args={"n_components": 8, "random_state": 4},
+        method: str = "BayesianGaussianMixture",
+        method_args: Dict = {"n_components": 8, "random_state": 4},
     ):
 
         self.latent = latent