networkx >=2.7 instead

Author: SherylHYX

README.md

@@ -246,7 +246,7 @@ If you notice anything unexpected, please open an [issue](https://github.com/She
 
 **Installation**
 
-Binaries are provided for Python version >= 3.7 and NetworkX version < 2.7.
+Binaries are provided for Python version >= 3.7 and NetworkX version >= 2.7.
 
 After installing [PyTorch](https://pytorch.org/get-started/locally/) and [PyG](https://pytorch-geometric.readthedocs.io/en/latest/notes/installation.html), simply run
 

docs/source/notes/installation.rst

@@ -2,7 +2,7 @@ Installation
 ============
 
 The installation of PyTorch Geometric Signed Directed requires the presence of certain prerequisites. These are described in great detail in the installation description of PyTorch Geometric. Please follow the instructions laid out `here <https://pytorch-geometric-signed-directed.readthedocs.io/en/latest/notes/installation.html>`_. You might also take a look at the `README file <https://github.com/SherylHYX/pytorch_geometric_signed_directed>`_ of the PyTorch Geometric Signed Directed repository.
-Binaries are provided for Python version >= 3.7, PyG version at least 2, and NetworkX version earlier than 2.7.
+Binaries are provided for Python version >= 3.7, PyG version at least 2, and NetworkX version no earlier than 2.7.
 
 After installing `PyTorch <https://pytorch.org/get-started/locally/>`_ and `PyG <https://pytorch-geometric.readthedocs.io/en/latest/notes/installation.html>`_, simply run
 

setup.py

@@ -11,7 +11,7 @@
     "scikit-learn",
     "torch_geometric",
     "numpy",
-    "networkx<2.7",
+    "networkx>=2.7",
     "scipy"
 ]
 

torch_geometric_signed_directed/data/signed/polarized_SSBM.py

@@ -115,7 +115,7 @@ def polarized_SSBM(total_n: int = 100, num_com: int = 3, N: int = 30, K: int = 2
         conflict_groups[perm[i]] = conflict_groups_old[i]
     # now fix the network connectedness and degree
     # first we fix connectedness
-    G = nx.from_scipy_sparse_matrix(large_A_p-large_A_n)
+    G = nx.from_scipy_sparse_array(large_A_p-large_A_n)
     largest_cc = max(nx.connected_components(G), key=len)
     A_p_new = sp.lil_matrix(large_A_p[list(largest_cc)][:, list(largest_cc)])
     A_n_new = sp.lil_matrix(large_A_n[list(largest_cc)][:, list(largest_cc)])

torch_geometric_signed_directed/utils/general/extract_network.py

@@ -21,13 +21,13 @@ def extract_network(A: sp.spmatrix, labels: Union[np.array, LongTensor, None] =
         * **A** (scipy sparse matrix) - Adjacency matrix after fixing degrees and obtaining a connected netework.
         * **labels** (numpy array) - Node labels after fixing degrees and obtaining a connected netework.
     """
-    G = nx.from_scipy_sparse_matrix(A, create_using=nx.DiGraph)
+    G = nx.from_scipy_sparse_array(A, create_using=nx.DiGraph)
     largest_cc = max(nx.weakly_connected_components(G))
     A_new = A[list(largest_cc)][:, list(largest_cc)]
     labels_new = None
     if labels is not None:
         labels_new = labels[list(largest_cc)]
-    G0 = nx.from_scipy_sparse_matrix(A_new, create_using=nx.DiGraph)
+    G0 = nx.from_scipy_sparse_array(A_new, create_using=nx.DiGraph)
     flag = True
     iter_num = 0
     keep = []
@@ -47,7 +47,7 @@ def extract_network(A: sp.spmatrix, labels: Union[np.array, LongTensor, None] =
                 lowest_degree -= 1
                 print('Nothing to keep, reducing lowest_degree by one to be {}!'.format(
                     lowest_degree))
-                G0 = nx.from_scipy_sparse_matrix(
+                G0 = nx.from_scipy_sparse_array(
                     A_new, create_using=nx.DiGraph)
                 break
 

torch_geometric_signed_directed/utils/general/link_split.py

@@ -238,7 +238,7 @@ def link_class_split(data: torch_geometric.data.Data, size: int = None, splits:
     A_undirected = to_scipy_sparse_matrix(undirect_edge_index)
     if maintain_connect:
         assert ratio == 1, "ratio should be 1.0 if maintain_connect=True"
-        G = nx.from_scipy_sparse_matrix(
+        G = nx.from_scipy_sparse_array(
             A_undirected, create_using=nx.Graph, edge_attribute='weight')
         mst = list(tree.minimum_spanning_edges(
             G, algorithm="kruskal", data=False))