Merge branch 'Snopt push origin mainoff-line'

Author: gbg141

configs/transforms/liftings/graph2simplicial/line_lifting.yaml

@@ -0,0 +1,3 @@
+transform_type: "lifting"
+transform_name: "SimplicialLineLifting"
+feature_lifting: ProjectionSum

modules/data/utils/utils.py

@@ -50,16 +50,16 @@ def get_complex_connectivity(complex, max_rank, signed=False):
                 )
             except ValueError:  # noqa: PERF203
                 if connectivity_info == "incidence":
-                    connectivity[f"{connectivity_info}_{rank_idx}"] = (
-                        generate_zero_sparse_connectivity(
-                            m=practical_shape[rank_idx - 1], n=practical_shape[rank_idx]
-                        )
+                    connectivity[
+                        f"{connectivity_info}_{rank_idx}"
+                    ] = generate_zero_sparse_connectivity(
+                        m=practical_shape[rank_idx - 1], n=practical_shape[rank_idx]
                     )
                 else:
-                    connectivity[f"{connectivity_info}_{rank_idx}"] = (
-                        generate_zero_sparse_connectivity(
-                            m=practical_shape[rank_idx], n=practical_shape[rank_idx]
-                        )
+                    connectivity[
+                        f"{connectivity_info}_{rank_idx}"
+                    ] = generate_zero_sparse_connectivity(
+                        m=practical_shape[rank_idx], n=practical_shape[rank_idx]
                     )
     connectivity["shape"] = practical_shape
     return connectivity

modules/transforms/data_transform.py

@@ -20,6 +20,9 @@
 from modules.transforms.liftings.graph2simplicial.graph_induced_lifting import (
     SimplicialGraphInducedLifting,
 )
+from modules.transforms.liftings.graph2simplicial.line_lifting import (
+    SimplicialLineLifting,
+)
 from modules.transforms.liftings.graph2simplicial.vietoris_rips_lifting import (
     SimplicialVietorisRipsLifting,
 )
@@ -29,6 +32,7 @@
     "HypergraphKNNLifting": HypergraphKNNLifting,
     # Graph -> Simplicial Complex
     "SimplicialCliqueLifting": SimplicialCliqueLifting,
+    "SimplicialLineLifting": SimplicialLineLifting,
     "SimplicialVietorisRipsLifting": SimplicialVietorisRipsLifting,
     "SimplicialGraphInducedLifting": SimplicialGraphInducedLifting,
     # Graph -> Cell Complex

modules/transforms/liftings/graph2simplicial/line_lifting.py

@@ -0,0 +1,81 @@
+import networkx as nx
+import torch_geometric
+from toponetx.classes import SimplicialComplex
+
+from modules.transforms.liftings.graph2simplicial.base import (
+    Graph2SimplicialLifting,
+)
+
+
+class SimplicialLineLifting(Graph2SimplicialLifting):
+    r"""Lifts graphs to a simplicial complex domain by considering line simplicial complex.
+
+    Line simplicial complex is a clique complex of the line graph. Line graph is a graph, in which
+    the vertices are the edges in the initial graph, and two vertices are adjacent if the corresponding
+    edges are adjacent in the initial graph.
+
+    Parameters
+    ----------
+    **kwargs : optional
+        Additional arguments for the class.
+    """
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+
+    def lift_topology(self, data: torch_geometric.data.Data) -> dict:
+        r"""Lifts the topology of a graph to simplicial domain via line simplicial complex construction.
+
+        Parameters
+        ----------
+        data : torch_geometric.data.Data
+            The input data to be lifted.
+
+        Returns
+        ----------
+        dict
+            The lifted topology.
+        """
+
+        graph = self._generate_graph_from_data(data)
+        line_graph = nx.line_graph(graph)
+
+        node_features = {
+            node: ((data.x[node[0], :] + data.x[node[1], :]) / 2)
+            for node in list(line_graph.nodes)
+        }
+
+        cliques = nx.find_cliques(line_graph)
+        simplices = list(cliques)  # list(map(lambda x: set(x), cliques))
+
+        # we need to rename simplices here since now vertices are named as pairs
+        self.rename_vertices_dict = {
+            node: i for i, node in enumerate(line_graph.nodes)
+        }
+        self.rename_vertices_dict_inverse = {
+            i: node for node, i in self.rename_vertices_dict.items()
+        }
+        renamed_line_graph = nx.relabel_nodes(
+            line_graph, self.rename_vertices_dict
+        )
+
+        renamed_simplices = [
+            {self.rename_vertices_dict[vertex] for vertex in simplex}
+            for simplex in simplices
+        ]
+
+        renamed_node_features = {
+            self.rename_vertices_dict[node]: value
+            for node, value in node_features.items()
+        }
+
+        simplicial_complex = SimplicialComplex(simplices=renamed_simplices)
+        self.complex_dim = simplicial_complex.dim
+
+        simplicial_complex.set_simplex_attributes(
+            renamed_node_features, name="features"
+        )
+
+        return self._get_lifted_topology(
+            simplicial_complex, renamed_line_graph
+        )

test/transforms/liftings/graph2simplicial/test_line_lifting.py

@@ -0,0 +1,84 @@
+"""Test the message passing module."""
+
+import torch
+import torch_geometric
+
+from modules.transforms.liftings.graph2simplicial.line_lifting import (
+    SimplicialLineLifting,
+)
+
+
+def create_test_graph():
+    num_nodes = 5
+    x = [1] * num_nodes
+    edge_index = torch.tensor(
+        [[0, 1, 2, 3, 4, 1, 2], [1, 2, 3, 4, 0, 4, 3]], dtype=torch.long
+    )  # [[0, 0, 1, 1, 2, 2, 3], [1, 4, 2, 3, 3, 4, 4]]
+    y = [0, 0, 1, 1, 0]
+
+    return torch_geometric.data.Data(
+        x=torch.tensor(x).float().reshape(-1, 1),
+        edge_index=edge_index,  # torch.Tensor(edge_index, dtype=torch.long),
+        num_nodes=num_nodes,
+        y=torch.tensor(y),
+    )
+
+
+class TestSimplicialLineLifting:
+    """Test the SimplicialLineLifting class."""
+
+    def setup_method(self):
+        # Load the graph
+        self.data = create_test_graph()  # load_manual_graph()
+
+        # Initialise the SimplicialCliqueLifting class
+        self.lifting_signed = SimplicialLineLifting(signed=True)
+        self.lifting_unsigned = SimplicialLineLifting(signed=False)
+
+    def test_lift_topology(self):
+        """Test the lift_topology method."""
+
+        # Test the lift_topology method
+        lifted_data_signed = self.lifting_signed.forward(self.data.clone())
+        lifted_data_unsigned = self.lifting_unsigned.forward(self.data.clone())
+
+        expected_incidence_1 = torch.tensor(
+            [
+                [-1.0, -1.0, -1.0, -1.0, 0.0, 0.0, 0.0, 0.0, 0.0],
+                [1.0, 0.0, 0.0, 0.0, -1.0, -1.0, 0.0, 0.0, 0.0],
+                [0.0, 1.0, 0.0, 0.0, 1.0, 0.0, -1.0, 0.0, 0.0],
+                [0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, -1.0, -1.0],
+                [0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 1.0, 0.0],
+                [0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 1.0],
+            ]
+        )
+
+        print(lifted_data_signed.incidence_1.to_dense())
+
+        assert (
+            abs(expected_incidence_1) == lifted_data_unsigned.incidence_1.to_dense()
+        ).all(), "Something is wrong with unsigned incidence_1 (nodes to edges)."
+        assert (
+            expected_incidence_1 == lifted_data_signed.incidence_1.to_dense()
+        ).all(), "Something is wrong with signed incidence_1 (nodes to edges)."
+
+        expected_incidence_2 = torch.tensor(
+            [
+                [1.0, 0.0, 0.0],
+                [-1.0, 1.0, 0.0],
+                [0.0, 0.0, 1.0],
+                [0.0, -1.0, -1.0],
+                [1.0, 0.0, 0.0],
+                [0.0, 0.0, 0.0],
+                [0.0, 1.0, 0.0],
+                [0.0, 0.0, 0.0],
+                [0.0, 0.0, 1.0],
+            ]
+        )
+
+        assert (
+            abs(expected_incidence_2) == lifted_data_unsigned.incidence_2.to_dense()
+        ).all(), "Something is wrong with unsigned incidence_2 (edges to triangles)."
+        assert (
+            expected_incidence_2 == lifted_data_signed.incidence_2.to_dense()
+        ).all(), "Something is wrong with signed incidence_2 (edges to triangles)."