Fix documentation

Prior to this commit, but as of the very previous one, the documentation
did not accurately reflect the form imposed on the filter the method is
looking for. The leading coefficient of the denominator polynomial is
fixed at 1e0. This is a crucial difference that fixes a bug recorded in
another commit.

src/dkpy/uncertainty_bound.py

@@ -128,13 +128,13 @@ def _pre_J(x: np.ndarray) -> np.ndarray:
         x : np.ndarray
             Coefficients of the biproper rational function of given order that
             is the uncertainty weight. If order is equal to n and
-                x = [a_n, ..., a_0, b_n, ..., b_0],
+                x = [a_n, ..., a_0, b_n-1, ..., b_0],
             then
                 w_E(x) = a(s) / b(s),
             where
                 a(s) = a_n s^n + ... + a_1 s + a_0
             and
-                b(s) = 1e0 s^n + ... + b_1 s + b_0.
+                b(s) = 1e0 s^n + b_n-1 s^n-1 + ... + b_1 s + b_0.
             Must be of size 2 * (order + 1) - 1.
 
         Returns
@@ -162,13 +162,13 @@ def _J(x: np.ndarray) -> float:
         x : np.ndarray
             Coefficients of the biproper rational function of given order that
             is the uncertainty weight. If order is equal to n and
-                x = [a_n, ..., a_0, b_n, ..., b_0],
+                x = [a_n, ..., a_0, b_n-1, ..., b_0],
             then
                 w_E(x) = a(s) / b(s),
             where
                 a(s) = a_n s^n + ... + a_1 s + a_0
             and
-                b(s) = 1e0 s^n + ... + b_1 s + b_0.
+                b(s) = 1e0 s^n + b_n-1 s^n-1 + ... + b_1 s + b_0.
             Must be of size 2 * (order + 1) - 1.
 
         Returns