ensure sorted sets before averaging in pick points (#226)

should finally fix #189

.gitea/ISSUE_TEMPLATE/BUG_TEMPLATE.yml

@@ -17,7 +17,7 @@ body:
       description: For which version have you observed this behavior?
       placeholder: You find the program version in "Help/About"
     validations:
-      required: true
+      required: false
   - type: textarea
     attributes:
       label: Expected behavior

src/nmreval/data/points.py

@@ -320,7 +320,11 @@ class Points:
             pts = []
 
         _tmp_x = self._x[self.mask]
-        _tmp_y = self._y[self.mask]
+        x_order = np.argsort(_tmp_x)
+        _tmp_x = _tmp_x[x_order]
+        _tmp_y = self._y[self.mask][x_order]
+        _tmp_yerr = self._y_err[self.mask][x_order]
+
         if idx is not None:
             for idx_i in idx:
                 if isinstance(idx_i, tuple):
@@ -338,7 +342,7 @@ class Points:
                     right_b = int(min(len(self), x_idx + avg_range[1] + 1))
 
                 if left_b < right_b:
-                    pts.append([_tmp_x[x_idx], *self._average(avg_mode, x_idx, left_b, right_b)])
+                    pts.append([_tmp_x[x_idx], *self._average(_tmp_x, _tmp_y, _tmp_yerr, avg_mode, x_idx, left_b, right_b)])
                 else:
                     pts.append([_tmp_x[x_idx], _tmp_y[x_idx], self._y_err[x_idx]])
 
@@ -358,28 +362,37 @@ class Points:
             left_b = int(max(0, x_idx - avg_range[0]))
             right_b = int(min(len(self), x_idx + avg_range[1] + 1))
 
-            pts.append([_tmp_x[x_idx], *self._average(avg_mode, x_idx, left_b, right_b)])
+            pts.append([_tmp_x[x_idx], *self._average(_tmp_x, _tmp_y, _tmp_yerr, avg_mode, x_idx, left_b, right_b)])
 
         return pts
 
-    def _average(self, mode: str, idx, left: int, right: int) -> tuple[float, float]:
+    @staticmethod
+    def _average(
+            x: np.ndarray,
+            y: np.ndarray,
+            y_err: np.ndarray,
+            mode: str,
+            idx: int,
+            left: int,
+            right: int,
+    ) -> tuple[float, float]:
         if mode == 'mean':
-            y_mean = np.mean(self._y[self.mask][left:right].real)
-            y_err = np.linalg.norm(self._y_err[self.mask][left:right]) / (right - left)
+            y_mean = np.mean(y[left:right].real)
+            y_err_mean = np.linalg.norm(y_err[left:right]) / (right - left)
 
         elif mode == 'sum':
-            y_mean = np.sum(self._y[self.mask][left:right].real)
-            y_err = np.linalg.norm(self._y_err[self.mask][left:right])
+            y_mean = np.sum(y[left:right].real)
+            y_err_mean = np.linalg.norm(y_err[left:right])
 
         elif mode == 'integral':
-            y_mean = simpson(self._y[self.mask][left:right].real, x=self._x[left:right])
-            y_err = np.linalg.norm(cumulative_trapezoid(self._y_err[self.mask][left:right].real, x=self._x[left:right]))
+            y_mean = simpson(y[left:right].real, x=x[left:right])
+            y_err_mean = np.linalg.norm(cumulative_trapezoid(y_err[left:right].real, x=x[left:right]))
 
         else:
-            y_mean = self._y[self.mask][idx].real
-            y_err = self._y_err[self.mask][idx]
+            y_mean = y[idx].real
+            y_err_mean = y_err[idx]
 
-        return y_mean, y_err
+        return y_mean, y_err_mean
 
     def concatenate(self, other):
         """