Add PyG smoke test and CI pipeline

.github/workflows/ci.yml

@@ -0,0 +1,40 @@
+name: CI
+
+on:
+  pull_request:
+  workflow_dispatch:
+
+jobs:
+  tests:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.10'
+
+      - name: Install PyTorch CPU
+        run: |
+          python -m pip install --upgrade pip
+          pip install torch==2.3.0 torchvision==0.18.0 torchaudio==2.3.0 --index-url https://download.pytorch.org/whl/cpu --no-cache-dir
+
+      - name: Install NVTX mock bindings
+        run: |
+          pip install git+https://github.com/YaoYinYing/nvtx-mock --force-reinstall
+          pip install nvtx
+
+      - name: Install graph dependencies
+        run: |
+          pip install "dgl<=2.4.0" -f https://data.dgl.ai/wheels/repo.html
+          pip install pyg_lib torch_scatter torch_sparse torch_cluster torch_spline_conv -f https://data.pyg.org/whl/torch-2.3.0+cpu.html
+          pip install torch_geometric
+
+      - name: Install project
+        run: |
+          pip install -e .
+
+      - name: Run smoke tests
+        run: |
+          python -m pytest -q tests/smoke

.gitignore

@@ -16,3 +16,4 @@ tests/tim_barrel_scaffold
 tests/tests_*
 examples/outputs/*
 examples/example_outputs/*
+.venv/

examples/minimal_graph_migration.py

@@ -0,0 +1,30 @@
+"""
+Minimal example derived from the example suite to exercise both DGL and
+PyTorch Geometric backends during the migration process.
+"""
+
+import torch
+import dgl
+from torch_geometric.data import Data
+
+
+def build_graphs():
+    src = torch.tensor([0, 1, 2, 2])
+    dst = torch.tensor([1, 2, 0, 3])
+    features = torch.arange(4, dtype=torch.float32).unsqueeze(-1)
+
+    dgl_graph = dgl.graph((src, dst))
+    dgl_graph.ndata["x"] = features
+
+    pyg_graph = Data(x=features, edge_index=torch.stack([src, dst], dim=0))
+    return dgl_graph, pyg_graph
+
+
+def main():
+    dgl_graph, pyg_graph = build_graphs()
+    print("DGL graph:", dgl_graph)
+    print("PyG graph:", pyg_graph)
+
+
+if __name__ == "__main__":
+    main()

tests/smoke/test_graph_backends.py

@@ -0,0 +1,22 @@
+import torch
+import dgl
+from torch_geometric.data import Data
+
+
+def test_graph_construction_parity():
+    src = torch.tensor([0, 1, 2, 2])
+    dst = torch.tensor([1, 2, 0, 3])
+    features = torch.arange(4, dtype=torch.float32).unsqueeze(-1)
+
+    dgl_graph = dgl.graph((src, dst))
+    dgl_graph.ndata["x"] = features
+
+    pyg_graph = Data(x=features, edge_index=torch.stack([src, dst], dim=0))
+
+    assert dgl_graph.num_nodes() == pyg_graph.num_nodes
+    assert dgl_graph.num_edges() == pyg_graph.num_edges
+
+    dgl_edges = torch.stack(dgl_graph.edges(order="eid"), dim=0)
+    pyg_edges = pyg_graph.edge_index
+    assert torch.equal(dgl_edges, pyg_edges)
+    assert torch.equal(dgl_graph.ndata["x"], pyg_graph.x)