fix problems at the edges of the isochrones grid

Previously the detection of that was basically incorrecty and
occasionally results outside the mass or feh grid could have been returned

py/minimint/mist_interpolator.py

@@ -473,13 +473,13 @@ def __get_eep_coeffs(self, mass, logage, feh):
         l2feh = l1feh + 1
         l1mass = np.searchsorted(self.umass, mass) - 1
         l2mass = l1mass + 1
-        bad = np.zeros(N, dtype=bool)
-        bad = bad | (l2mass >= len(self.umass)) | (l2feh >= len(
+        bads0 = np.zeros(N, dtype=bool)
+        bads0 = bads0 | (l2mass >= len(self.umass)) | (l2feh >= len(
             self.ufeh)) | (l1mass < 0) | (l1feh < 0)
-        l1mass[bad] = 0
-        l2mass[bad] = 1
-        l1feh[bad] = 0
-        l2feh[bad] = 1
+        l1mass[bads0] = 0
+        l2mass[bads0] = 1
+        l1feh[bads0] = 0
+        l2feh[bads0] = 1
 
         x = (feh - self.ufeh[l1feh]) / (self.ufeh[l2feh] - self.ufeh[l1feh]
                                         )  # from 0 to 1
@@ -499,7 +499,7 @@ def FF(curi, xind):
                 C4[xind] * self.logage_grid[l2feh[xind], l2mass[xind], curi])
 
         lefts, rights, bads = _binary_search(mass, logage, self.neep, FF)
-        bads = bads | (rights >= self.neep)
+        bads = bads0 | bads | (rights >= self.neep)
         lefts[bads] = 0
         rights[bads] = 1
         LV, RV = [(C1 * self.logage_grid[l1feh, l1mass, _] +