sort imports via isort

.github/workflows/linting.yml

@@ -65,6 +65,25 @@ jobs:
         run: |
           yamllint .
 
+  isort:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v2
+
+      - name: Set up Python
+        uses: actions/setup-python@v2
+        with:
+          python-version: 3.9
+
+      - name: Install dependencies
+        run: |
+          pip install isort
+
+      - name: Run linting
+        run: |
+          isort --check-only .
+
   pylint:
     runs-on: ubuntu-latest
 

.pre-commit-config.yaml

@@ -7,3 +7,9 @@ repos:
     hooks:
       - id: yapf
         name: Format code
+
+  - repo: https://github.com/pycqa/isort
+    rev: 5.10.1
+    hooks:
+      - id: isort
+        name: Sort imports

benchmark/citation/appnp.py

@@ -1,10 +1,11 @@
 import argparse
+
 import torch
-from torch.nn import Linear
 import torch.nn.functional as F
-from torch_geometric.nn import APPNP
-
 from citation import get_planetoid_dataset, random_planetoid_splits, run
+from torch.nn import Linear
+
+from torch_geometric.nn import APPNP
 
 parser = argparse.ArgumentParser()
 parser.add_argument('--dataset', type=str, required=True)

benchmark/citation/arma.py

@@ -1,10 +1,11 @@
 import argparse
+
 import torch
 import torch.nn.functional as F
-from torch_geometric.nn import ARMAConv
-
 from citation import get_planetoid_dataset, random_planetoid_splits, run
 
+from torch_geometric.nn import ARMAConv
+
 parser = argparse.ArgumentParser()
 parser.add_argument('--dataset', type=str, required=True)
 parser.add_argument('--random_splits', type=bool, default=False)

benchmark/citation/cheb.py

@@ -1,10 +1,11 @@
 import argparse
+
 import torch
 import torch.nn.functional as F
-from torch_geometric.nn import ChebConv
-
 from citation import get_planetoid_dataset, random_planetoid_splits, run
 
+from torch_geometric.nn import ChebConv
+
 parser = argparse.ArgumentParser()
 parser.add_argument('--dataset', type=str, required=True)
 parser.add_argument('--random_splits', type=bool, default=False)

benchmark/citation/datasets.py

@@ -1,7 +1,7 @@
 import os.path as osp
 
-from torch_geometric.datasets import Planetoid
 import torch_geometric.transforms as T
+from torch_geometric.datasets import Planetoid
 
 
 def get_planetoid_dataset(name, normalize_features=False, transform=None):

benchmark/citation/gat.py

@@ -1,10 +1,11 @@
 import argparse
+
 import torch
 import torch.nn.functional as F
-from torch_geometric.nn import GATConv
-
 from citation import get_planetoid_dataset, random_planetoid_splits, run
 
+from torch_geometric.nn import GATConv
+
 parser = argparse.ArgumentParser()
 parser.add_argument('--dataset', type=str, required=True)
 parser.add_argument('--random_splits', type=bool, default=False)

benchmark/citation/gcn.py

@@ -1,10 +1,11 @@
 import argparse
+
 import torch
 import torch.nn.functional as F
-from torch_geometric.nn import GCNConv
-
 from citation import get_planetoid_dataset, random_planetoid_splits, run
 
+from torch_geometric.nn import GCNConv
+
 parser = argparse.ArgumentParser()
 parser.add_argument('--dataset', type=str, required=True)
 parser.add_argument('--random_splits', type=bool, default=False)

benchmark/citation/sgc.py

@@ -1,10 +1,11 @@
 import argparse
+
 import torch
 import torch.nn.functional as F
-from torch_geometric.nn import SGConv
-
 from citation import get_planetoid_dataset, random_planetoid_splits, run
 
+from torch_geometric.nn import SGConv
+
 parser = argparse.ArgumentParser()
 parser.add_argument('--dataset', type=str, required=True)
 parser.add_argument('--random_splits', type=bool, default=False)

benchmark/kernel/asap.py

@@ -1,8 +1,9 @@
 import torch
 import torch.nn.functional as F
 from torch.nn import Linear
-from torch_geometric.nn import (ASAPooling, GraphConv, global_mean_pool,
-                                JumpingKnowledge)
+
+from torch_geometric.nn import (ASAPooling, GraphConv, JumpingKnowledge,
+                                global_mean_pool)
 
 
 class ASAP(torch.nn.Module):

benchmark/kernel/datasets.py

@@ -1,9 +1,10 @@
 import os.path as osp
 
 import torch
+
+import torch_geometric.transforms as T
 from torch_geometric.datasets import TUDataset
 from torch_geometric.utils import degree
-import torch_geometric.transforms as T
 
 
 class NormalizedDegree(object):

benchmark/kernel/diff_pool.py

@@ -3,7 +3,8 @@
 import torch
 import torch.nn.functional as F
 from torch.nn import Linear
-from torch_geometric.nn import DenseSAGEConv, dense_diff_pool, JumpingKnowledge
+
+from torch_geometric.nn import DenseSAGEConv, JumpingKnowledge, dense_diff_pool
 
 
 class Block(torch.nn.Module):

benchmark/kernel/edge_pool.py

@@ -1,8 +1,9 @@
 import torch
 import torch.nn.functional as F
 from torch.nn import Linear
-from torch_geometric.nn import (GraphConv, EdgePooling, global_mean_pool,
-                                JumpingKnowledge)
+
+from torch_geometric.nn import (EdgePooling, GraphConv, JumpingKnowledge,
+                                global_mean_pool)
 
 
 class EdgePool(torch.nn.Module):

benchmark/kernel/gcn.py

@@ -1,7 +1,8 @@
 import torch
 import torch.nn.functional as F
 from torch.nn import Linear
-from torch_geometric.nn import GCNConv, global_mean_pool, JumpingKnowledge
+
+from torch_geometric.nn import GCNConv, JumpingKnowledge, global_mean_pool
 
 
 class GCN(torch.nn.Module):

benchmark/kernel/gin.py

@@ -1,7 +1,9 @@
 import torch
 import torch.nn.functional as F
-from torch.nn import Linear, Sequential, ReLU, BatchNorm1d as BN
-from torch_geometric.nn import GINConv, global_mean_pool, JumpingKnowledge
+from torch.nn import BatchNorm1d as BN
+from torch.nn import Linear, ReLU, Sequential
+
+from torch_geometric.nn import GINConv, JumpingKnowledge, global_mean_pool
 
 
 class GIN0(torch.nn.Module):

benchmark/kernel/global_attention.py

@@ -1,7 +1,8 @@
 import torch
 import torch.nn.functional as F
 from torch.nn import Linear
-from torch_geometric.nn import SAGEConv, GlobalAttention
+
+from torch_geometric.nn import GlobalAttention, SAGEConv
 
 
 class GlobalAttentionNet(torch.nn.Module):

benchmark/kernel/graclus.py

@@ -1,9 +1,10 @@
 import torch
 import torch.nn.functional as F
 from torch.nn import Linear
+
 from torch_geometric.data import Batch
-from torch_geometric.nn import (GraphConv, graclus, max_pool, global_mean_pool,
-                                JumpingKnowledge)
+from torch_geometric.nn import (GraphConv, JumpingKnowledge, global_mean_pool,
+                                graclus, max_pool)
 
 
 class Graclus(torch.nn.Module):

benchmark/kernel/graph_sage.py

@@ -1,7 +1,8 @@
 import torch
 import torch.nn.functional as F
 from torch.nn import Linear
-from torch_geometric.nn import SAGEConv, global_mean_pool, JumpingKnowledge
+
+from torch_geometric.nn import JumpingKnowledge, SAGEConv, global_mean_pool
 
 
 class GraphSAGE(torch.nn.Module):

benchmark/kernel/main.py

@@ -1,21 +1,20 @@
+import argparse
 from itertools import product
 
-import argparse
+from asap import ASAP
 from datasets import get_dataset
-from train_eval import cross_validation_with_val_set
-
-from gcn import GCN, GCNWithJK
-from graph_sage import GraphSAGE, GraphSAGEWithJK
-from gin import GIN0, GIN0WithJK, GIN, GINWithJK
-from graclus import Graclus
-from top_k import TopK
-from sag_pool import SAGPool
 from diff_pool import DiffPool
 from edge_pool import EdgePool
+from gcn import GCN, GCNWithJK
+from gin import GIN, GIN0, GIN0WithJK, GINWithJK
 from global_attention import GlobalAttentionNet
+from graclus import Graclus
+from graph_sage import GraphSAGE, GraphSAGEWithJK
+from sag_pool import SAGPool
 from set2set import Set2SetNet
 from sort_pool import SortPool
-from asap import ASAP
+from top_k import TopK
+from train_eval import cross_validation_with_val_set
 
 parser = argparse.ArgumentParser()
 parser.add_argument('--epochs', type=int, default=100)

benchmark/kernel/main_performance.py

@@ -2,17 +2,15 @@
 from itertools import product
 
 import torch
-
-from torch_geometric import seed_everything
-from torch_geometric.loader import DataLoader
-from torch_geometric.profile import profileit, timeit, get_stats_summary
-
 from datasets import get_dataset
-from train_eval import train, eval_acc
-
 from gcn import GCN
-from graph_sage import GraphSAGE
 from gin import GIN
+from graph_sage import GraphSAGE
+from train_eval import eval_acc, train
+
+from torch_geometric import seed_everything
+from torch_geometric.loader import DataLoader
+from torch_geometric.profile import get_stats_summary, profileit, timeit
 
 seed_everything(0)
 

benchmark/kernel/sag_pool.py

@@ -1,8 +1,9 @@
 import torch
 import torch.nn.functional as F
 from torch.nn import Linear
-from torch_geometric.nn import (GraphConv, SAGPooling, global_mean_pool,
-                                JumpingKnowledge)
+
+from torch_geometric.nn import (GraphConv, JumpingKnowledge, SAGPooling,
+                                global_mean_pool)
 
 
 class SAGPool(torch.nn.Module):

benchmark/kernel/set2set.py

@@ -1,6 +1,7 @@
 import torch
 import torch.nn.functional as F
 from torch.nn import Linear
+
 from torch_geometric.nn import SAGEConv, Set2Set
 
 

benchmark/kernel/sort_pool.py

@@ -1,6 +1,7 @@
 import torch
 import torch.nn.functional as F
-from torch.nn import Linear, Conv1d
+from torch.nn import Conv1d, Linear
+
 from torch_geometric.nn import SAGEConv, global_sort_pool
 
 

benchmark/kernel/top_k.py

@@ -1,8 +1,9 @@
 import torch
 import torch.nn.functional as F
 from torch.nn import Linear
-from torch_geometric.nn import (GraphConv, TopKPooling, global_mean_pool,
-                                JumpingKnowledge)
+
+from torch_geometric.nn import (GraphConv, JumpingKnowledge, TopKPooling,
+                                global_mean_pool)
 
 
 class TopK(torch.nn.Module):

benchmark/kernel/train_eval.py

@@ -2,10 +2,12 @@
 
 import torch
 import torch.nn.functional as F
+from sklearn.model_selection import StratifiedKFold
 from torch import tensor
 from torch.optim import Adam
-from sklearn.model_selection import StratifiedKFold
-from torch_geometric.loader import DataLoader, DenseDataLoader as DenseLoader
+
+from torch_geometric.loader import DataLoader
+from torch_geometric.loader import DenseDataLoader as DenseLoader
 
 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
 

benchmark/points/datasets.py

@@ -1,7 +1,7 @@
 import os.path as osp
 
-from torch_geometric.datasets import ModelNet
 import torch_geometric.transforms as T
+from torch_geometric.datasets import ModelNet
 
 
 def get_dataset(num_points):

benchmark/points/edge_cnn.py

@@ -1,11 +1,14 @@
 import argparse
+
 import torch
 import torch.nn.functional as F
-from torch.nn import Sequential as Seq, Linear as Lin, ReLU
-from torch_geometric.nn import DynamicEdgeConv, global_max_pool
-
 from points.datasets import get_dataset
 from points.train_eval import run
+from torch.nn import Linear as Lin
+from torch.nn import ReLU
+from torch.nn import Sequential as Seq
+
+from torch_geometric.nn import DynamicEdgeConv, global_max_pool
 
 parser = argparse.ArgumentParser()
 parser.add_argument('--epochs', type=int, default=200)

benchmark/points/mpnn.py

@@ -1,11 +1,14 @@
 import argparse
+
 import torch
 import torch.nn.functional as F
-from torch.nn import Sequential as Seq, Linear as Lin, ReLU
-from torch_geometric.nn import NNConv, radius_graph, fps, global_mean_pool
-
 from points.datasets import get_dataset
 from points.train_eval import run
+from torch.nn import Linear as Lin
+from torch.nn import ReLU
+from torch.nn import Sequential as Seq
+
+from torch_geometric.nn import NNConv, fps, global_mean_pool, radius_graph
 
 parser = argparse.ArgumentParser()
 parser.add_argument('--epochs', type=int, default=200)