fix some weird reversions to csr_matrix

graspologic/embed/omni.py

@@ -52,7 +52,7 @@ def _get_omnibus_matrix_sparse(matrices: List[csr_array]) -> csr_array:
         # row
         rows.append(hstack(current_row))
 
-    return vstack(rows, format="csr")
+    return csr_array(vstack(rows, format="csr"))
 
 
 def _get_laplacian_matrices(
@@ -97,7 +97,7 @@ def _get_omni_matrix(
     out : 2d-array
         Array of shape (n_vertices * n_graphs, n_vertices * n_graphs)
     """
-    if isspmatrix_csr(graphs[0]):
+    if isinstance(graphs[0], csr_array):
         return _get_omnibus_matrix_sparse(graphs)  # type: ignore
 
     shape = graphs[0].shape

graspologic/utils/utils.py

@@ -44,7 +44,7 @@ def average_matrices(
     if isinstance(matrices[0], np.ndarray):
         return np.mean(matrices, axis=0)  # type: ignore
     elif isspmatrix_csr(matrices[0]):
-        return sum(matrices) / len(matrices)
+        return np.sum(matrices) / len(matrices)
 
     raise TypeError(f"Unexpected type {matrices}")
 
@@ -325,7 +325,9 @@ def symmetrize(
            [1, 1, 1]])
     """
     # graph = import_graph(graph)
-    sparse = isspmatrix_csr(graph)
+
+    sparse = isinstance(graph, csr_array)
+
     pac = scipy.sparse if sparse else np
 
     if method == "triu":
@@ -340,6 +342,11 @@ def symmetrize(
 
     dia = diags(graph.diagonal()) if sparse else np.diag(np.diag(graph))
     graph = graph + graph.T - dia
+
+    if sparse:
+        # some scipy funcs still output a sparse matrix
+        graph = csr_array(graph)
+
     return graph
 
 
@@ -360,7 +367,11 @@ def remove_loops(graph: GraphRepresentation) -> Union[np.ndarray, csr_array]:
     """
     graph = import_graph(graph)
 
-    dia = diags(graph.diagonal()) if isspmatrix_csr(graph) else np.diag(np.diag(graph))
+    dia = (
+        csr_array(diags(graph.diagonal()))
+        if isinstance(graph, csr_array)
+        else np.diag(np.diag(graph))
+    )
 
     graph = graph - dia
 
@@ -460,23 +471,28 @@ def to_laplacian(
         in_root = 1 / np.sqrt(in_degree)  # this is 10x faster than ** -0.5
         out_root = 1 / np.sqrt(out_degree)
 
-    diag = diags if isspmatrix_csr(graph) else np.diag
+    sparse = isinstance(graph, csr_array)
+
+    diag = diags if sparse else np.diag
 
     in_root[np.isinf(in_root)] = 0
     out_root[np.isinf(out_root)] = 0
 
     in_root = diag(in_root)  # just change to sparse diag for sparse support
     out_root = diag(out_root)
 
+    if sparse:
+        # for some reason scipy is still returning in csr_matrix form, not what we want
+        in_root = csr_array(in_root)
+        out_root = csr_array(out_root)
+
     if form == "I-DAD":
         L = diag(in_degree) - A
         L = in_root @ L @ in_root
     elif form == "DAD" or form == "R-DAD":
         L = out_root @ A @ in_root
     if is_symmetric(A):
-        return symmetrize(
-            L, method="avg"
-        )  # sometimes machine prec. makes this necessary
+        L = symmetrize(L, method="avg")  # sometimes machine prec. makes this necessary
     return L