Created nice wrapper functions around the njit functions for convenience

jive/solver/icholprecon.py

 import numpy as np
 import scipy.sparse as spsp
-import scipy.sparse.linalg as spspla
 from numba import njit
 
 class ICholPrecon:
@@ -13,17 +12,15 @@ class ICholPrecon:
 
     def update(self, sourcematrix):
         self._sourcematrix = sourcematrix
-        self._L = spsp.tril(self._sourcematrix, format='csr')
-        ichol_data = incomplete_cholesky(self._L.data, self._L.indices, self._L.indptr)
-        self._L.data = ichol_data
+        self._L = incomplete_cholesky(self._sourcematrix)
         self._LT = self._L.T.tocsr()
 
     def dot(self, lhs):
         return self._L @ (self._L.T @ lhs)
 
     def solve(self, rhs):
-        tmp = solve_triangular(self._L.data, self._L.indices, self._L.indptr, rhs, lower=True)
-        return solve_triangular(self._LT.data, self._LT.indices, self._LT.indptr, tmp, lower=False)
+        tmp = solve_triangular(self._L, rhs, lower=True)
+        return solve_triangular(self._LT, tmp, lower=False)
 
     def get_matrix(self):
         return self._L @ self._L.T
@@ -33,6 +30,27 @@ def declare(factory):
     factory.declare_precon('ichol', ICholPrecon)
 
 
+#####################
+# wrapper functions #
+#####################
+
+def incomplete_cholesky(A):
+    if not spsp.isspmatrix_csr(A):
+        raise ValueError('A has to be a sparse matrix in csr format')
+
+    L = spsp.tril(A, format='csr')
+    L.data = incomplete_cholesky_jit(L.data, L.indices, L.indptr)
+
+    return L
+
+
+def solve_triangular(A, b, lower=True):
+    if not spsp.isspmatrix_csr(A):
+        raise ValueError('A has to be a sparse matrix in csr format')
+
+    return solve_triangular_jit(A.data, A.indices, A.indptr, b, lower)
+
+
 ##########################
 # numba helper functions #
 ##########################
@@ -73,7 +91,7 @@ def idxs2rowscols(indices, indptr):
 
 
 @njit
-def incomplete_cholesky(data, indices, indptr):
+def incomplete_cholesky_jit(data, indices, indptr):
     L_data = data.copy()
     rows, cols = idxs2rowscols(indices, indptr)
 
@@ -126,7 +144,7 @@ def incomplete_cholesky(data, indices, indptr):
 
 
 @njit
-def solve_triangular(data, indices, indptr, b, lower=True):
+def solve_triangular_jit(data, indices, indptr, b, lower=True):
     x = np.zeros_like(b)
 
     if lower: