This is an automated email from the ASF dual-hosted git repository.
jin pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/incubator-hugegraph-ai.git
The following commit(s) were added to refs/heads/main by this push:
new 672aeef fix: pylint in ml (#125)
672aeef is described below
commit 672aeef06fa1e3b6ebe96838e0f2368c7ce9f3ee
Author: SoJ <[email protected]>
AuthorDate: Fri Dec 6 13:00:00 2024 +0800
fix: pylint in ml (#125)
---
.../src/hugegraph_ml/data/hugegraph2dgl.py | 1 +
.../src/hugegraph_ml/examples/bgnn_example.py | 4 +-
.../src/hugegraph_ml/examples/bgrl_example.py | 2 +-
.../src/hugegraph_ml/examples/care_gnn_example.py | 3 --
.../examples/correct_and_smooth_example.py | 4 +-
.../src/hugegraph_ml/examples/gatne_example.py | 2 +-
hugegraph-ml/src/hugegraph_ml/models/arma.py | 6 +--
hugegraph-ml/src/hugegraph_ml/models/bgnn.py | 11 +-----
hugegraph-ml/src/hugegraph_ml/models/bgrl.py | 9 +----
hugegraph-ml/src/hugegraph_ml/models/care_gnn.py | 12 +++---
.../src/hugegraph_ml/models/cluster_gcn.py | 5 ++-
.../src/hugegraph_ml/models/correct_and_smooth.py | 2 +-
hugegraph-ml/src/hugegraph_ml/models/dagnn.py | 2 +-
hugegraph-ml/src/hugegraph_ml/models/deepergcn.py | 8 ++--
hugegraph-ml/src/hugegraph_ml/models/gatne.py | 17 +++-----
hugegraph-ml/src/hugegraph_ml/models/pgnn.py | 25 ++++++------
hugegraph-ml/src/hugegraph_ml/models/seal.py | 45 +++++++++-------------
.../hugegraph_ml/tasks/fraud_detector_caregnn.py | 13 +------
.../tasks/hetero_sample_embed_gatne.py | 7 ++--
.../src/hugegraph_ml/tasks/link_prediction_pgnn.py | 14 +++----
.../src/hugegraph_ml/tasks/link_prediction_seal.py | 35 ++++-------------
.../tasks/node_classify_with_sample.py | 3 +-
.../src/hugegraph_ml/utils/dgl2hugegraph_utils.py | 6 +--
23 files changed, 89 insertions(+), 147 deletions(-)
diff --git a/hugegraph-ml/src/hugegraph_ml/data/hugegraph2dgl.py
b/hugegraph-ml/src/hugegraph_ml/data/hugegraph2dgl.py
index 92ea00c..14f156d 100644
--- a/hugegraph-ml/src/hugegraph_ml/data/hugegraph2dgl.py
+++ b/hugegraph-ml/src/hugegraph_ml/data/hugegraph2dgl.py
@@ -16,6 +16,7 @@
# under the License.
# pylint: disable=too-many-branches
+# pylint: disable=C0304
import warnings
from typing import Optional, List
diff --git a/hugegraph-ml/src/hugegraph_ml/examples/bgnn_example.py
b/hugegraph-ml/src/hugegraph_ml/examples/bgnn_example.py
index 7c353f3..c395a1d 100644
--- a/hugegraph-ml/src/hugegraph_ml/examples/bgnn_example.py
+++ b/hugegraph-ml/src/hugegraph_ml/examples/bgnn_example.py
@@ -15,6 +15,8 @@
# specific language governing permissions and limitations
# under the License.
+# pylint: disable=C0103
+
from hugegraph_ml.models.bgnn import (
GNNModelDGL,
BGNNPredictor,
@@ -48,7 +50,7 @@ def bgnn_example():
gbdt_depth=6,
gbdt_lr=0.1,
)
- metrics = bgnn.fit(
+ _ = bgnn.fit(
g,
encoded_X,
y,
diff --git a/hugegraph-ml/src/hugegraph_ml/examples/bgrl_example.py
b/hugegraph-ml/src/hugegraph_ml/examples/bgrl_example.py
index 77c03f9..f73ab0c 100644
--- a/hugegraph-ml/src/hugegraph_ml/examples/bgrl_example.py
+++ b/hugegraph-ml/src/hugegraph_ml/examples/bgrl_example.py
@@ -16,7 +16,7 @@
# under the License.
from hugegraph_ml.data.hugegraph2dgl import HugeGraph2DGL
-from hugegraph_ml.models.bgrl import BGRL, GCN, MLP_Predictor,
CosineDecayScheduler, get_graph_drop_transform
+from hugegraph_ml.models.bgrl import BGRL, GCN, MLP_Predictor
from hugegraph_ml.models.mlp import MLPClassifier
from hugegraph_ml.tasks.node_classify import NodeClassify
from hugegraph_ml.tasks.node_embed import NodeEmbed
diff --git a/hugegraph-ml/src/hugegraph_ml/examples/care_gnn_example.py
b/hugegraph-ml/src/hugegraph_ml/examples/care_gnn_example.py
index e6fb52f..723d37d 100644
--- a/hugegraph-ml/src/hugegraph_ml/examples/care_gnn_example.py
+++ b/hugegraph-ml/src/hugegraph_ml/examples/care_gnn_example.py
@@ -18,11 +18,8 @@
from hugegraph_ml.data.hugegraph2dgl import HugeGraph2DGL
from hugegraph_ml.models.care_gnn import CAREGNN
from hugegraph_ml.tasks.fraud_detector_caregnn import DetectorCaregnn
-
-import dgl
import torch
-
def care_gnn_example(n_epochs=200):
hg2d = HugeGraph2DGL()
graph = hg2d.convert_hetero_graph(
diff --git
a/hugegraph-ml/src/hugegraph_ml/examples/correct_and_smooth_example.py
b/hugegraph-ml/src/hugegraph_ml/examples/correct_and_smooth_example.py
index aa09808..921fa94 100644
--- a/hugegraph-ml/src/hugegraph_ml/examples/correct_and_smooth_example.py
+++ b/hugegraph-ml/src/hugegraph_ml/examples/correct_and_smooth_example.py
@@ -16,10 +16,8 @@
# under the License.
from hugegraph_ml.data.hugegraph2dgl import HugeGraph2DGL
-from hugegraph_ml.models.correct_and_smooth import MLP, MLPLinear
+from hugegraph_ml.models.correct_and_smooth import MLP
from hugegraph_ml.tasks.node_classify import NodeClassify
-import argparse
-
def cs_example(n_epochs=200):
hg2d = HugeGraph2DGL()
diff --git a/hugegraph-ml/src/hugegraph_ml/examples/gatne_example.py
b/hugegraph-ml/src/hugegraph_ml/examples/gatne_example.py
index 6a994f0..a92abd1 100644
--- a/hugegraph-ml/src/hugegraph_ml/examples/gatne_example.py
+++ b/hugegraph-ml/src/hugegraph_ml/examples/gatne_example.py
@@ -17,7 +17,7 @@
from hugegraph_ml.data.hugegraph2dgl import HugeGraph2DGL
-from hugegraph_ml.models.gatne import DGLGATNE, NeighborSampler
+from hugegraph_ml.models.gatne import DGLGATNE
from hugegraph_ml.tasks.hetero_sample_embed_gatne import HeteroSampleEmbedGATNE
diff --git a/hugegraph-ml/src/hugegraph_ml/models/arma.py
b/hugegraph-ml/src/hugegraph_ml/models/arma.py
index 76ca096..dd8dc0f 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/arma.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/arma.py
@@ -15,7 +15,7 @@
# specific language governing permissions and limitations
# under the License.right (c) 2024 by jinsong, All Rights Reserved.
-# pylint: disable=E1101
+# pylint: disable=E1101,C0103
"""
auto-regressive moving average (ARMA)
@@ -28,11 +28,9 @@ DGL code:
https://github.com/dmlc/dgl/tree/master/examples/pytorch/arma
"""
import math
-
import dgl.function as fn
-
import torch
-import torch.nn as nn
+from torch import nn
import torch.nn.functional as F
diff --git a/hugegraph-ml/src/hugegraph_ml/models/bgnn.py
b/hugegraph-ml/src/hugegraph_ml/models/bgnn.py
index b137043..8015af8 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/bgnn.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/bgnn.py
@@ -15,7 +15,7 @@
# specific language governing permissions and limitations
# under the License.
-# pylint: disable=E1102,E0401,E0711,E0606,E0602
+# pylint:
disable=E1102,E0401,E0711,E0606,E0602,C0103,C0206,W0612,C0209,R1705,C0200,R1735,W0201
"""
Boost-GNN (BGNN)
@@ -335,14 +335,7 @@ class BGNNPredictor:
train_metric, val_metric, test_metric = metrics[metric_name][-1]
if epoch and epoch % logging_epochs == 0:
pbar.set_description(
- "Epoch {:05d} | Loss {:.3f} | Loss {:.3f}/{:.3f}/{:.3f} | Time
{:.4f}".format(
- epoch,
- loss,
- train_metric,
- val_metric,
- test_metric,
- epoch_time,
- )
+ f"Epoch {epoch:05d} | Loss {loss:.3f} | Loss
{train_metric:.3f}/{val_metric:.3f}/{test_metric:.3f} | Time {epoch_time:.4f}"
)
def fit(
diff --git a/hugegraph-ml/src/hugegraph_ml/models/bgrl.py
b/hugegraph-ml/src/hugegraph_ml/models/bgrl.py
index ed87a43..60b01a0 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/bgrl.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/bgrl.py
@@ -15,7 +15,7 @@
# specific language governing permissions and limitations
# under the License.right (c) 2024 by jinsong, All Rights Reserved.
-# pylint: disable=E1102
+# pylint: disable=E1102,C0103,R1705.R1734
"""
Bootstrapped Graph Latents (BGRL)
@@ -143,7 +143,6 @@ class BGRL(nn.Module):
target_x = transform_2(graph)
online_x, target_x = dgl.add_self_loop(online_x),
dgl.add_self_loop(target_x)
online_feats, target_feats = online_x.ndata["feat"],
target_x.ndata["feat"]
-
# forward online network
online_y1 = self.online_encoder(online_x, online_feats)
# prediction
@@ -151,7 +150,6 @@ class BGRL(nn.Module):
# forward target network
with torch.no_grad():
target_y1 = self.target_encoder(target_x, target_feats).detach()
-
# forward online network 2
online_y2 = self.online_encoder(target_x, target_feats)
# prediction
@@ -159,7 +157,6 @@ class BGRL(nn.Module):
# forward target network
with torch.no_grad():
target_y2 = self.target_encoder(online_x, online_feats).detach()
-
loss = (
2
- cosine_similarity(online_q1, target_y1.detach(), dim=-1).mean()
@@ -238,9 +235,7 @@ class CosineDecayScheduler:
)
else:
raise ValueError(
- "Step ({}) > total number of steps ({}).".format(
- step, self.total_steps
- )
+ f"Step ({step}) > total number of steps ({self.total_steps})."
)
def get_graph_drop_transform(drop_edge_p, feat_mask_p):
diff --git a/hugegraph-ml/src/hugegraph_ml/models/care_gnn.py
b/hugegraph-ml/src/hugegraph_ml/models/care_gnn.py
index 84e850f..994513e 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/care_gnn.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/care_gnn.py
@@ -15,7 +15,7 @@
# specific language governing permissions and limitations
# under the License.
-# pylint: disable=E1101
+# pylint: disable=E1101,C0103
"""
CAmouflage-REsistant GNN (CARE-GNN)
@@ -30,7 +30,7 @@ DGL code:
https://github.com/dmlc/dgl/tree/master/examples/pytorch/caregnn
import dgl.function as fn
import numpy as np
import torch as th
-import torch.nn as nn
+from torch import nn
class CAREConv(nn.Module):
@@ -100,7 +100,7 @@ class CAREConv(nn.Module):
g.ndata["h"] = feat
hr = {}
- for i, etype in enumerate(g.canonical_etypes):
+ for _, etype in enumerate(g.canonical_etypes):
g.apply_edges(self._calc_distance, etype=etype)
self.dist[etype] = g.edges[etype].data["d"]
sampled_edges = self._top_p_sampling(g[etype], self.p[etype])
@@ -109,10 +109,10 @@ class CAREConv(nn.Module):
g.send_and_recv(
sampled_edges,
fn.copy_u("h", "m"),
- fn.mean("m", "h_%s" % etype[1]),
+ fn.mean("m", f"h_{etype[1]}"),
etype=etype,
)
- hr[etype] = g.ndata["h_%s" % etype[1]]
+ hr[etype] = g.ndata[f"h_{etype[1]}"]
if self.activation is not None:
hr[etype] = self.activation(hr[etype])
@@ -175,7 +175,7 @@ class CAREGNN(nn.Module):
)
# Hidden layers with n - 2 layers
- for i in range(self.num_layers - 2):
+ for _ in range(self.num_layers - 2):
self.layers.append(
CAREConv(
self.hid_dim,
diff --git a/hugegraph-ml/src/hugegraph_ml/models/cluster_gcn.py
b/hugegraph-ml/src/hugegraph_ml/models/cluster_gcn.py
index 09fd04a..231e2bb 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/cluster_gcn.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/cluster_gcn.py
@@ -27,10 +27,11 @@ Author's code:
https://github.com/google-research/google-research/tree/master/cl
DGL code: https://github.com/dmlc/dgl/tree/master/examples/pytorch/cluster_gcn
"""
-import dgl.nn as dglnn
-import torch.nn as nn
+from torch import nn
import torch.nn.functional as F
+import dgl.nn as dglnn
+
class SAGE(nn.Module):
def __init__(self, in_feats, n_hidden, n_classes):
super().__init__()
diff --git a/hugegraph-ml/src/hugegraph_ml/models/correct_and_smooth.py
b/hugegraph-ml/src/hugegraph_ml/models/correct_and_smooth.py
index 70b8f9e..50481c6 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/correct_and_smooth.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/correct_and_smooth.py
@@ -15,7 +15,7 @@
# specific language governing permissions and limitations
# under the License.
-# pylint: disable=E1101,E1102
+# pylint: disable=E1101,E1102,W0223,R1705
"""
Correct and Smooth (C&S)
diff --git a/hugegraph-ml/src/hugegraph_ml/models/dagnn.py
b/hugegraph-ml/src/hugegraph_ml/models/dagnn.py
index 8afd8db..77cb1ce 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/dagnn.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/dagnn.py
@@ -15,7 +15,7 @@
# specific language governing permissions and limitations
# under the License.
-# pylint: disable=E1101
+# pylint: disable=E1101,C0103
"""
Deep Adaptive Graph Neural Network (DAGNN)
diff --git a/hugegraph-ml/src/hugegraph_ml/models/deepergcn.py
b/hugegraph-ml/src/hugegraph_ml/models/deepergcn.py
index 0781ded..26b41fc 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/deepergcn.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/deepergcn.py
@@ -27,12 +27,12 @@ Author's code: https://github.com/lightaime/deep_gcns_torch
DGL code: https://github.com/dmlc/dgl/tree/master/examples/pytorch/deepergcn
"""
-import dgl.function as fn
-import torch.nn as nn
+import torch
+from torch import nn
import torch.nn.functional as F
+
+import dgl.function as fn
from dgl.nn.functional import edge_softmax
-from dgl.nn.pytorch.glob import AvgPooling
-import torch
# pylint: disable=E1101,E0401
diff --git a/hugegraph-ml/src/hugegraph_ml/models/gatne.py
b/hugegraph-ml/src/hugegraph_ml/models/gatne.py
index b0fe329..69cac2d 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/gatne.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/gatne.py
@@ -15,7 +15,7 @@
# specific language governing permissions and limitations
# under the License.right (c) 2024 by jinsong, All Rights Reserved.
-# pylint: disable=E1101
+# pylint: disable=E1101,R0205,C0200,R1732
"""
General Attributed Multiplex HeTerogeneous Network Embedding (GATNE)
@@ -28,24 +28,19 @@ DGL code:
https://github.com/dmlc/dgl/tree/master/examples/pytorch/GATNE-T
"""
import math
-import os
-import sys
import time
-from collections import defaultdict
+import multiprocessing
+from functools import partial, reduce
import numpy as np
+
import torch
from torch import nn
import torch.nn.functional as F
-from numpy import random
from torch.nn.parameter import Parameter
-from tqdm.auto import tqdm
import dgl
import dgl.function as fn
-import multiprocessing
-from functools import partial, reduce, wraps
-
class NeighborSampler(object):
def __init__(self, g, num_fanouts):
@@ -232,7 +227,7 @@ def generate_pairs(all_walks, window_size, num_workers):
# for each node, choose the first neighbor and second neighbor of it to
form pairs
# Get all worker processes
start_time = time.time()
- print("We are generating pairs with {} cores.".format(num_workers))
+ print(f"We are generating pairs with {num_workers} cores.")
# Start all worker processes
pool = multiprocessing.Pool(processes=num_workers)
@@ -259,7 +254,7 @@ def generate_pairs(all_walks, window_size, num_workers):
pool.close()
end_time = time.time()
- print("Generate pairs end, use {}s.".format(end_time - start_time))
+ print(f"Generate pairs end, use {end_time - start_time}s.")
return np.array([list(pair) for pair in set(pairs)])
diff --git a/hugegraph-ml/src/hugegraph_ml/models/pgnn.py
b/hugegraph-ml/src/hugegraph_ml/models/pgnn.py
index 1c6dcb5..8f51d0c 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/pgnn.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/pgnn.py
@@ -15,7 +15,7 @@
# specific language governing permissions and limitations
# under the License.
-# pylint: disable=E1101,E0401
+# pylint: disable=E1101,E0401,C0103,R1732,C0200,R1705
"""
Position-aware Graph Neural Networks (P-GNN)
@@ -27,19 +27,20 @@ Author's code: https://github.com/JiaxuanYou/P-GNN
DGL code: https://github.com/dmlc/dgl/tree/master/examples/pytorch/P-GNN
"""
-import dgl.function as fn
-import torch
-from torch import nn
-import torch.nn.functional as F
import multiprocessing as mp
import random
from multiprocessing import get_context
+import torch
+from torch import nn
+import torch.nn.functional as F
+
import networkx as nx
import numpy as np
from tqdm.auto import tqdm
from sklearn.metrics import roc_auc_score
+import dgl.function as fn
class PGNN_layer(nn.Module):
def __init__(self, input_dim, output_dim):
@@ -162,9 +163,9 @@ def split_edges(p, edges, data, non_train_ratio=0.2):
data.update(
{
- "{}_edges_train".format(p): edges[:, :split1], # 80%
- "{}_edges_val".format(p): edges[:, split1:split2], # 10%
- "{}_edges_test".format(p): edges[:, split2:], # 10%
+ f"{p}_edges_train": edges[:, :split1], # 80%
+ f"{p}_edges_val": edges[:, split1:split2], # 10%
+ f"p{}_edges_test": edges[:, split2:], # 10%
}
)
@@ -393,8 +394,8 @@ def preselect_anchor(data, num_workers=4):
def get_loss(p, data, out, loss_func, device, get_auc=True):
edge_mask = np.concatenate(
(
- data["positive_edges_{}".format(p)],
- data["negative_edges_{}".format(p)],
+ data[f"positive_edges_{p}"],
+ data[f"negative_edges_{p}"],
),
axis=-1,
)
@@ -410,13 +411,13 @@ def get_loss(p, data, out, loss_func, device,
get_auc=True):
label_positive = torch.ones(
[
- data["positive_edges_{}".format(p)].shape[1],
+ data[f"positive_edges_{p}"].shape[1],
],
dtype=pred.dtype,
)
label_negative = torch.zeros(
[
- data["negative_edges_{}".format(p)].shape[1],
+ data[f"negative_edges_{p}"].shape[1],
],
dtype=pred.dtype,
)
diff --git a/hugegraph-ml/src/hugegraph_ml/models/seal.py
b/hugegraph-ml/src/hugegraph_ml/models/seal.py
index 3c167c5..43d9a89 100644
--- a/hugegraph-ml/src/hugegraph_ml/models/seal.py
+++ b/hugegraph-ml/src/hugegraph_ml/models/seal.py
@@ -15,7 +15,7 @@
# specific language governing permissions and limitations
# under the License.
-# pylint: disable=E0401
+# pylint: disable=E0401,R1719,C0103,R0205,R1721.R1705,R0205
"""
SEAL
@@ -27,33 +27,28 @@ Author's code: https://github.com/muhanzhang/SEAL
DGL code: https://github.com/dmlc/dgl/tree/master/examples/pytorch/seal
"""
+import argparse
+import os.path as osp
+from copy import deepcopy
+import logging
+import os
+import time
+
import torch
-import torch.nn as nn
+from torch import nn
import torch.nn.functional as F
-
-from dgl.nn.pytorch import GraphConv, SAGEConv, SortPooling, SumPooling
-import argparse
+from torch.utils.data import DataLoader, Dataset
import dgl
+from dgl import add_self_loop, NID
+from dgl.dataloading.negative_sampler import Uniform
+from dgl.nn.pytorch import GraphConv, SAGEConv, SortPooling, SumPooling
import numpy as np
-import pandas as pd
from ogb.linkproppred import DglLinkPropPredDataset, Evaluator
from scipy.sparse.csgraph import shortest_path
-
-import os.path as osp
-from copy import deepcopy
-
-from dgl import add_self_loop, DGLGraph, NID
-from dgl.dataloading.negative_sampler import Uniform
-from torch.utils.data import DataLoader, Dataset
from tqdm import tqdm
-import logging
-import os
-import time
-
-
class GCN(nn.Module):
"""
GCN Model
@@ -581,7 +576,7 @@ class SEALSampler(object):
sample_nodes = [target_nodes]
frontiers = target_nodes
- for i in range(self.hop):
+ for _ in range(self.hop):
frontiers = self.graph.out_edges(frontiers)[1]
frontiers = torch.unique(frontiers)
sample_nodes.append(frontiers)
@@ -622,7 +617,7 @@ class SEALSampler(object):
subgraph_list = []
labels_list = []
edge_dataset = EdgeDataSet(edges, labels,
transform=self.sample_subgraph)
- self.print_fn("Using {} workers in sampling
job.".format(self.num_workers))
+ self.print_fn(f"Using {self.num_workers} workers in sampling job.")
sampler = DataLoader(
edge_dataset,
batch_size=32,
@@ -709,9 +704,7 @@ class SEALData(object):
path = osp.join(
self.save_dir or "",
- "{}_{}_{}-hop_{}-subsample.bin".format(
- self.prefix, split_type, self.hop, subsample_ratio
- ),
+
"f{self.prefix}_{split_type}_{self.hop}-hop_{subsample_ratio}-subsample.bin"
)
if osp.exists(path):
@@ -720,15 +713,15 @@ class SEALData(object):
dataset = GraphDataSet(graph_list, data["labels"])
else:
- self.print_fn("Processed {} files not exist.".format(split_type))
+ self.print_fn(f"Processed {split_type} files not exist.")
edges, labels = self.generator(split_type)
- self.print_fn("Generate {} edges totally.".format(edges.size(0)))
+ self.print_fn(f"Generate {edges.size(0)} edges totally.")
graph_list, labels = self.sampler(edges, labels)
dataset = GraphDataSet(graph_list, labels)
dgl.save_graphs(path, graph_list, {"labels": labels})
- self.print_fn("Save preprocessed subgraph to {}".format(path))
+ self.print_fn(f"Save preprocessed subgraph to {path}")
return dataset
diff --git a/hugegraph-ml/src/hugegraph_ml/tasks/fraud_detector_caregnn.py
b/hugegraph-ml/src/hugegraph_ml/tasks/fraud_detector_caregnn.py
index 632cfe3..53518a6 100644
--- a/hugegraph-ml/src/hugegraph_ml/tasks/fraud_detector_caregnn.py
+++ b/hugegraph-ml/src/hugegraph_ml/tasks/fraud_detector_caregnn.py
@@ -15,7 +15,7 @@
# specific language governing permissions and limitations
# under the License.
-# pylint: disable=E0401
+# pylint: disable=E0401,C0301
import torch
from torch import nn
@@ -51,7 +51,6 @@ class DetectorCaregnn:
torch.nonzero(train_mask,
as_tuple=False).squeeze(1).to(self._device)
)
val_idx = torch.nonzero(val_mask,
as_tuple=False).squeeze(1).to(self._device)
- test_idx = torch.nonzero(test_mask,
as_tuple=False).squeeze(1).to(self._device)
rl_idx = torch.nonzero(
train_mask.to(self._device) & labels.bool(), as_tuple=False
).squeeze(1)
@@ -89,15 +88,7 @@ class DetectorCaregnn:
tr_loss.backward()
optimizer.step()
print(
- "Epoch {}, Train: Recall: {:.4f} AUC: {:.4f} Loss: {:.4f} |
Val: Recall: {:.4f} AUC: {:.4f} Loss: {:.4f}".format(
- epoch,
- tr_recall,
- tr_auc,
- tr_loss.item(),
- val_recall,
- val_auc,
- val_loss.item(),
- )
+ f"Epoch {epoch}, Train: Recall: {tr_recall:.4f} AUC:
{tr_auc:.4f} Loss: {tr_loss.item():.4f} | Val: Recall: {val_recall:.4f} AUC:
{val_auc:.4f} Loss: {val_loss.item():.4f}"
)
self._model.RLModule(self.graph, epoch, rl_idx)
diff --git a/hugegraph-ml/src/hugegraph_ml/tasks/hetero_sample_embed_gatne.py
b/hugegraph-ml/src/hugegraph_ml/tasks/hetero_sample_embed_gatne.py
index 62af5d2..91d00ac 100644
--- a/hugegraph-ml/src/hugegraph_ml/tasks/hetero_sample_embed_gatne.py
+++ b/hugegraph-ml/src/hugegraph_ml/tasks/hetero_sample_embed_gatne.py
@@ -16,11 +16,11 @@
# under the License.
+import random
import dgl
import torch
from torch import nn
from tqdm.auto import tqdm
-import random
from hugegraph_ml.models.gatne import (
construct_typenodes_from_graph,
generate_pairs,
@@ -58,7 +58,7 @@ class HeteroSampleEmbedGATNE:
all_walks = []
for i in range(edge_type_count):
nodes = torch.LongTensor(type_nodes[i] *
num_walks).to(self._device)
- traces, types = dgl.sampling.random_walk(
+ traces, _ = dgl.sampling.random_walk(
self.graph,
nodes,
metapath=[self.graph.etypes[i]] * (neighbor_samples - 1),
@@ -84,14 +84,13 @@ class HeteroSampleEmbedGATNE:
lr=lr,
)
- tensors = []
for epoch in range(n_epochs):
self._model.train()
random.shuffle(train_pairs)
data_iter = tqdm(
train_dataloader,
- desc="epoch %d" % (epoch),
+ desc=f"epoch {epoch}",
total=(len(train_pairs) + (batch_size - 1)) // batch_size,
)
avg_loss = 0.0
diff --git a/hugegraph-ml/src/hugegraph_ml/tasks/link_prediction_pgnn.py
b/hugegraph-ml/src/hugegraph_ml/tasks/link_prediction_pgnn.py
index 8b03cbd..13e761e 100644
--- a/hugegraph-ml/src/hugegraph_ml/tasks/link_prediction_pgnn.py
+++ b/hugegraph-ml/src/hugegraph_ml/tasks/link_prediction_pgnn.py
@@ -18,8 +18,6 @@
import torch
import dgl
from torch import nn
-from tqdm import trange
-import numpy as np
from hugegraph_ml.models.pgnn import (
get_dataset,
preselect_anchor,
@@ -85,10 +83,10 @@ class LinkPredictionPGNN:
if epoch % 100 == 0:
print(
epoch,
- "Loss {:.4f}".format(loss_train),
- "Train AUC: {:.4f}".format(auc_train),
- "Val AUC: {:.4f}".format(auc_val),
- "Test AUC: {:.4f}".format(auc_test),
- "Best Val AUC: {:.4f}".format(best_auc_val),
- "Best Test AUC: {:.4f}".format(best_auc_test),
+ f"Loss {loss_train:.4f}",
+ f"Train AUC: {auc_train:.4f}",
+ f"Val AUC: {auc_val:.4f}",
+ f"Test AUC: {auc_test:.4f}",
+ f"Best Val AUC: {best_auc_val:.4f}",
+ f"Best Test AUC: {best_auc_test:.4f}",
)
diff --git a/hugegraph-ml/src/hugegraph_ml/tasks/link_prediction_seal.py
b/hugegraph-ml/src/hugegraph_ml/tasks/link_prediction_seal.py
index c307cc1..1294942 100644
--- a/hugegraph-ml/src/hugegraph_ml/tasks/link_prediction_seal.py
+++ b/hugegraph-ml/src/hugegraph_ml/tasks/link_prediction_seal.py
@@ -15,20 +15,16 @@
# specific language governing permissions and limitations
# under the License.
+# pylint: disable=R1728
-from typing import Literal
-
+import time
import torch
+from torch.nn import BCEWithLogitsLoss
from dgl import DGLGraph, NID, EID
-from torch import nn
-from tqdm import tqdm
from dgl.dataloading import GraphDataLoader
-from torch.nn import BCEWithLogitsLoss
-import time
+from tqdm import tqdm
import numpy as np
-from hugegraph_ml.models.seal import SEALData, DGCNN, evaluate_hits
-from hugegraph_ml.utils.early_stopping import EarlyStopping
-
+from hugegraph_ml.models.seal import SEALData, evaluate_hits
class LinkPredictionSeal:
def __init__(self, graph: DGLGraph, split_edge, model):
@@ -55,8 +51,6 @@ class LinkPredictionSeal:
num_workers=32,
print_fn=print,
)
- node_attribute = seal_data.ndata["feat"]
- edge_weight = seal_data.edata["weight"].float()
train_data = seal_data("train")
val_data = seal_data("valid")
test_data = seal_data("test")
@@ -103,9 +97,7 @@ class LinkPredictionSeal:
optimizer = torch.optim.Adam(parameters, lr=lr)
loss_fn = BCEWithLogitsLoss()
print(
- "Total parameters: {}".format(
- sum([p.numel() for p in self._model.parameters()])
- )
+ f"Total parameters: {sum([p.numel() for p in
self._model.parameters()])}"
)
# train and evaluate loop
@@ -135,16 +127,7 @@ class LinkPredictionSeal:
)
evaluate_time = time.time()
print(
- "Epoch-{}, train loss: {:.4f}, hits@{}: val-{:.4f},
test-{:.4f}, "
- "cost time: train-{:.1f}s, total-{:.1f}s".format(
- epoch,
- loss,
- 50,
- val_metric,
- test_metric,
- train_time - start_time,
- evaluate_time - start_time,
- )
+ f"Epoch-{epoch}, train loss: {loss:.4f}, hits@{50}:
val-{val_metric:.4f}, test-{test_metric:.4f}, cost time: train-{train_time -
start_time:.1f}s, total-{evaluate_time - start_time:.1f}s"
)
summary_val.append(val_metric)
summary_test.append(test_metric)
@@ -152,9 +135,7 @@ class LinkPredictionSeal:
print("Experiment Results:")
print(
- "Best hits@{}: {:.4f}, epoch: {}".format(
- 50, np.max(summary_test), np.argmax(summary_test)
- )
+ f"Best hits@{50}: {np.max(summary_test):.4f}, epoch:
{np.argmax(summary_test)}"
)
@torch.no_grad()
diff --git a/hugegraph-ml/src/hugegraph_ml/tasks/node_classify_with_sample.py
b/hugegraph-ml/src/hugegraph_ml/tasks/node_classify_with_sample.py
index 393cd09..c60d36c 100644
--- a/hugegraph-ml/src/hugegraph_ml/tasks/node_classify_with_sample.py
+++ b/hugegraph-ml/src/hugegraph_ml/tasks/node_classify_with_sample.py
@@ -15,6 +15,7 @@
# specific language governing permissions and limitations
# under the License.
+# pylint: disable=C0301
from typing import Literal
@@ -140,7 +141,7 @@ class NodeClassifyWithSample:
test_labels = []
total_loss = 0
with torch.no_grad():
- for it, sg in enumerate(self.dataloader):
+ for _, sg in enumerate(self.dataloader):
sg_feats = feats[sg.ndata["_ID"]]
sg_labels = labels[sg.ndata["_ID"]]
sg_test_msak = test_mask[sg.ndata["_ID"]].bool()
diff --git a/hugegraph-ml/src/hugegraph_ml/utils/dgl2hugegraph_utils.py
b/hugegraph-ml/src/hugegraph_ml/utils/dgl2hugegraph_utils.py
index cdc4ea3..1d652df 100644
--- a/hugegraph-ml/src/hugegraph_ml/utils/dgl2hugegraph_utils.py
+++ b/hugegraph-ml/src/hugegraph_ml/utils/dgl2hugegraph_utils.py
@@ -17,10 +17,11 @@
# pylint: disable=too-many-branches
# pylint: disable=too-many-statements
+# pylint: disable=E0401,C0302,C0103,W1514,R1735,R1734,C0206
import os
from typing import Optional
-
+import json
import dgl
import numpy as np
import scipy
@@ -31,7 +32,6 @@ from dgl.data.utils import _get_dgl_url, download, load_graphs
import networkx as nx
from ogb.linkproppred import DglLinkPropPredDataset
import pandas as pd
-import json
from pyhugegraph.api.graph import GraphManager
from pyhugegraph.api.schema import SchemaManager
from pyhugegraph.client import PyHugeClient
@@ -381,7 +381,6 @@ def import_graph_from_nx(
# create property schema
# check props and create vertex label
vertex_label = f"{dataset_name}_vertex"
- props_value = {}
client_schema.vertexLabel(vertex_label).useAutomaticId().ifNotExist().create()
# add vertices for batch (note MAX_BATCH_NUM)
idx_to_vertex_id = {}
@@ -532,7 +531,6 @@ def import_graph_from_ogb(
else:
raise ValueError("dataset not supported")
graph_dgl = dataset_dgl[0]
- split_edges = dataset_dgl.get_edge_split()
client: PyHugeClient = PyHugeClient(
ip=ip, port=port, graph=graph, user=user, pwd=pwd,
graphspace=graphspace
