Add docstring

src/xspline/bspl.py

@@ -7,12 +7,24 @@
 
 
 def cache_bspl(function: RawFunction) -> RawFunction:
+    """Cache implementation for bspline basis functions, to avoid repetitively
+    evaluate functions.
+
+    Parameters
+    ----------
+    function
+        Raw value, derivative and definite integral functions.
+
+    Returns
+    -------
+    describe
+        Cached version of the raw functions.
+
+    """
     cache = {}
 
     def wrapper_function(*args, **kwargs) -> NDArray:
-        key = tuple(
-            tuple(x.ravel()) if isinstance(x, np.ndarray) else x for x in args
-        )
+        key = tuple(tuple(x.ravel()) if isinstance(x, np.ndarray) else x for x in args)
         if key in cache:
             return cache[key]
         result = function(*args, **kwargs)
@@ -28,6 +40,22 @@ def cache_clear():
 
 @cache_bspl
 def bspl_val(params: BsplParams, x: NDArray) -> NDArray:
+    """Value of the bspline function.
+
+    Parameters
+    ----------
+    params
+        Bspline function parameters as a tuple including, knots, degree and the
+        index of the spline basis.
+    x
+        Data points.
+
+    Returns
+    -------
+    describe
+        Function value of the bspline function.
+
+    """
     # knots, degree, and index
     t, k, i = params
 
@@ -44,10 +72,10 @@ def bspl_val(params: BsplParams, x: NDArray) -> NDArray:
 
         if t[ii[0]] != t[ii[2]]:
             n0 = bspl_val((t, k - 1, i), x)
-            val0 = (x - t[ii[0]])*n0/(t[ii[2]] - t[ii[0]])
+            val0 = (x - t[ii[0]]) * n0 / (t[ii[2]] - t[ii[0]])
         if t[ii[1]] != t[ii[3]]:
             n1 = bspl_val((t, k - 1, i + 1), x)
-            val1 = (t[ii[3]] - x)*n1/(t[ii[3]] - t[ii[1]])
+            val1 = (t[ii[3]] - x) * n1 / (t[ii[3]] - t[ii[1]])
 
         val = val0 + val1
 
@@ -56,6 +84,22 @@ def bspl_val(params: BsplParams, x: NDArray) -> NDArray:
 
 @cache_bspl
 def bspl_der(params: BsplParams, x: NDArray, order: int) -> NDArray:
+    """Derivative of the bspline function.
+
+    Parameters
+    ----------
+    params
+        Bspline function parameters as a tuple including, knots, degree and the
+        index of the spline basis.
+    x
+        Data points.
+
+    Returns
+    -------
+    describe
+        Derivative of the bspline function.
+
+    """
     # knots, degree, and index
     t, k, i = params
 
@@ -72,10 +116,10 @@ def bspl_der(params: BsplParams, x: NDArray, order: int) -> NDArray:
 
     if t[ii[0]] != t[ii[2]]:
         n0 = bspl_der((t, k - 1, i), x, order - 1)
-        val0 = k*n0/(t[ii[2]] - t[ii[0]])
+        val0 = k * n0 / (t[ii[2]] - t[ii[0]])
     if t[ii[1]] != t[ii[3]]:
         n1 = bspl_der((t, k - 1, i + 1), x, order - 1)
-        val1 = k*n1/(t[ii[3]] - t[ii[1]])
+        val1 = k * n1 / (t[ii[3]] - t[ii[1]])
 
     val = val0 - val1
 
@@ -84,6 +128,22 @@ def bspl_der(params: BsplParams, x: NDArray, order: int) -> NDArray:
 
 @cache_bspl
 def bspl_int(params: BsplParams, x: NDArray, order: int) -> NDArray:
+    """Definite integral of the bspline function.
+
+    Parameters
+    ----------
+    params
+        Bspline function parameters as a tuple including, knots, degree and the
+        index of the spline basis.
+    x
+        Data points.
+
+    Returns
+    -------
+    describe
+        Definite integral of the bspline function.
+
+    """
     # knots, degree, and index
     t, k, i = params
 
@@ -101,21 +161,25 @@ def bspl_int(params: BsplParams, x: NDArray, order: int) -> NDArray:
 
         if t[ii[0]] != t[ii[2]]:
             val0 = (
-                (x - t[ii[0]])*bspl_int((t, k - 1, i), x, order) +
-                order*bspl_int((t, k - 1, i), x, order - 1)
-            )/(t[ii[2]] - t[ii[0]])
+                (x - t[ii[0]]) * bspl_int((t, k - 1, i), x, order)
+                + order * bspl_int((t, k - 1, i), x, order - 1)
+            ) / (t[ii[2]] - t[ii[0]])
         if t[ii[1]] != t[ii[3]]:
             val1 = (
-                (t[ii[3]] - x)*bspl_int((t, k - 1, i + 1), x, order) -
-                order*bspl_int((t, k - 1, i + 1), x, order - 1)
-            )/(t[ii[3]] - t[ii[1]])
+                (t[ii[3]] - x) * bspl_int((t, k - 1, i + 1), x, order)
+                - order * bspl_int((t, k - 1, i + 1), x, order - 1)
+            ) / (t[ii[3]] - t[ii[1]])
 
         val = val0 + val1
 
     return val
 
 
 def clear_bspl_cache() -> None:
+    """Clear all cache of the value, derivative and definite integral for
+    bspline function.
+
+    """
     bspl_val.cache_clear()
     bspl_der.cache_clear()
     bspl_int.cache_clear()
@@ -149,6 +213,19 @@ def __init__(self, params: BsplParams) -> None:
 
 
 def get_bspl_funs(knots: tuple[float, ...], degree: int) -> tuple[Bspl]:
-    return tuple(
-        Bspl((knots, degree, i)) for i in range(-degree, len(knots) - 1)
-    )
+    """Create the bspline basis functions give knots and degree.
+
+    Parameters
+    ----------
+    knots
+        Bspline knots.
+    degree
+        Bspline degree.
+
+    Returns
+    -------
+    describe
+        A full set of bspline functions.
+
+    """
+    return tuple(Bspl((knots, degree, i)) for i in range(-degree, len(knots) - 1))

src/xspline/indi.py

@@ -1,13 +1,28 @@
 from math import factorial
 
 import numpy as np
-from numpy.typing import NDArray
 
-from xspline.typing import IndiParams
+from xspline.typing import IndiParams, NDArray
 from xspline.xfunction import BundleXFunction
 
 
 def indi_val(params: IndiParams, x: NDArray) -> NDArray:
+    """Indicator value function,
+
+    Parameters
+    ----------
+    params
+        Indicator parameters as a tuple consists of lower and upper bound of
+        the interval corresponding to the indicator function.
+    x
+        Data points.
+
+    Returns
+    -------
+    describe
+        Indicator function value.
+
+    """
     # lower and upper bounds
     lb, ub = params
 
@@ -22,21 +37,54 @@ def indi_val(params: IndiParams, x: NDArray) -> NDArray:
 
 
 def indi_der(params: IndiParams, x: NDArray, order: int) -> NDArray:
+    """Indicator derivative function. Since indicator function is a piecewise
+    constant function, its derivative will always be zero.
+
+    Parameters
+    ----------
+    params
+        Indicator parameters as a tuple consists of lower and upper bound of
+        the interval corresponding to the indicator function.
+    x
+        Data points.
+
+    Returns
+    -------
+    describe
+        Indicator deviative value.
+
+    """
     return np.zeros(x.size, dtype=x.dtype)
 
 
 def indi_int(params: IndiParams, x: NDArray, order: int) -> NDArray:
+    """Indicator definite integral function. It is a piecewise polynomial
+    function.
+
+    Parameters
+    ----------
+    params
+        Indicator parameters as a tuple consists of lower and upper bound of
+        the interval corresponding to the indicator function.
+    x
+        Data points.
+
+    Returns
+    -------
+    describe
+        Indicator definite integral value.
+
+    """
     # lower and upper bounds
     lb, ub = params
 
     val = np.zeros(x.size, dtype=x.dtype)
     ind0 = (x >= lb[0]) & (x <= ub[0])
     ind1 = x > ub[0]
-    val[ind0] = (x[ind0] - lb[0])**(-order)/factorial(-order)
+    val[ind0] = (x[ind0] - lb[0]) ** (-order) / factorial(-order)
     for i in range(-order):
-        val[ind1] += (
-            ((ub[0] - lb[0])**(-order - i)/factorial(-order - i)) *
-            ((x[ind1] - ub[0])**i/factorial(i))
+        val[ind1] += ((ub[0] - lb[0]) ** (-order - i) / factorial(-order - i)) * (
+            (x[ind1] - ub[0]) ** i / factorial(i)
         )
     return val
 

src/xspline/poly.py

@@ -1,21 +1,70 @@
 from math import factorial
 
 import numpy as np
-from numpy.typing import NDArray
 
-from xspline.typing import PolyParams
+from xspline.typing import PolyParams, NDArray
 from xspline.xfunction import BundleXFunction, XFunction
 
 
 def poly_val(params: PolyParams, x: NDArray) -> NDArray:
+    """Polynomial value function.
+
+    Parameters
+    ----------
+    params
+        Polynomial coefficients.
+    x
+        Data points.
+
+    Returns
+    -------
+    describe
+        Polynomial function values.
+
+    """
     return np.polyval(params, x)
 
 
 def poly_der(params: PolyParams, x: NDArray, order: int) -> NDArray:
+    """Polynomial derivative function.
+
+    Parameters
+    ----------
+    params
+        Polynomial coefficients.
+    x
+        Data points.
+    order
+        Order of differentiation.
+
+    Returns
+    -------
+    describe
+        Polynomial derivative values.
+
+    """
     return np.polyval(np.polyder(params, order), x)
 
 
 def poly_int(params: PolyParams, x: NDArray, order: int) -> NDArray:
+    """Polynomial definite integral function.
+
+    Parameters
+    ----------
+    params
+        Polynomial coefficients.
+    x
+        Data points.
+    order
+        Order of integration. Here we use negative integer.
+
+    Returns
+    -------
+    describe
+        Polynomial definite integral values.
+
+    """
+
     return np.polyval(np.polyint(params, -order), x)
 
 
@@ -46,16 +95,50 @@ def __init__(self, params: PolyParams) -> None:
 
 
 def get_poly_params(fun: XFunction, x: float, degree: int) -> tuple[float, ...]:
+    """Solve polynomial (taylor) coefficients provided the ``XFunction``.
+
+    Parameters
+    ----------
+    fun
+        Provided ``XFunction`` to be approximated.
+    x
+        The point where we want to approximate ``XFunction`` by the polynomial.
+    degree
+        Degree of the approximation polynomial.
+
+    Returns
+    -------
+    describe
+        The approximation polynomial coefficients.
+
+    """
     if degree < 0:
         return (0.0,)
     rhs = np.array([fun(x, order=i) for i in range(degree, -1, -1)])
     mat = np.zeros((degree + 1, degree + 1))
     for i in range(degree + 1):
         for j in range(i + 1):
-            mat[i, j] = factorial(degree - j) / factorial(i - j) * x**(i - j)
+            mat[i, j] = factorial(degree - j) / factorial(i - j) * x ** (i - j)
     return tuple(np.linalg.solve(mat, rhs))
 
 
 def get_poly_fun(fun: XFunction, x: float, degree: int) -> Poly:
+    """Get the approximation polynomial function.
+
+    Parameters
+    ----------
+    fun
+        Provided ``XFunction`` to be approximated.
+    x
+        The point where we want to approximate ``XFunction`` by the polynomial.
+    degree
+        Degree of the approximation polynomial.
+
+    Returns
+    -------
+    describe
+        Instance of the ``Poly`` class to approximate provided ``XFunction``.
+
+    """
     params = get_poly_params(fun, x, degree)
     return Poly(params)