Update range slider when the parameters change

distribution_zoo/cont_uni/normal.py

@@ -12,28 +12,43 @@ class Normal(BaseDistribution):
     display_name = 'Normal'
     range_min = -50.0
     range_max = 50.0
-    param_range_start = -10.0
-    param_range_end = 10.0
-    param_mean = 0.0
-    param_std = 1.0
+    param_mean = st.session_state['normal_mean'] if 'normal_mean' in st.session_state else 0.0
+    param_std = st.session_state['normal_std'] if 'normal_std' in st.session_state else 1.0
+    param_range_start = None
+    param_range_end = None
 
     def __init__(self):
         super().__init__()
 
     def sliders(self):
 
+        if 'normal_range' not in st.session_state:
+            self.update_range()
+
+        # This slider's initial value is set from st.session_state['normal_mean'], set with update_range()
         self.param_range_start, self.param_range_end = st.sidebar.slider(
-            'Range', min_value=self.range_min, max_value=self.range_max, value=(-10.0, 10.0), step=0.1, key='normal_range'
+            'Range', min_value=self.range_min, max_value=self.range_max, step=0.1, key='normal_range'
         )
 
         self.param_mean = st.sidebar.slider(
-            r'Mean ($\mu$)', min_value=-30.0, max_value=30.0, value=0.0, step=0.1, key='normal_mean'
+            r'Mean ($\mu$)', min_value=-16.0, max_value=16.0, value=self.param_mean, step=0.1, key='normal_mean',
+            on_change=self.update_range
         )
 
         self.param_std = st.sidebar.slider(
-            r'Standard deviation ($\sigma$)', min_value=0.1, max_value=20.0, value=1.0, step=0.1, key='normal_std'
+            r'Standard deviation ($\sigma$)', min_value=0.1, max_value=8.0, value=self.param_std, step=0.1,
+            key='normal_std', on_change=self.update_range
         )
 
+    def update_range(self):
+
+        mean = st.session_state['normal_mean'] if 'normal_mean' in st.session_state else self.param_mean
+        std = st.session_state['normal_std'] if 'normal_std' in st.session_state else self.param_std
+
+        new_lower = round(stats.norm(loc=mean, scale=std).ppf(0.0001), 1)
+        new_upper = round(stats.norm(loc=mean, scale=std).ppf(0.9999), 1)
+        st.session_state['normal_range'] = (new_lower, new_upper)
+
     def plot(self):
 
         x = np.linspace(self.range_min, self.range_max, 1000)