Functions — github.com/LFhase/GIA-HAO

↓ 17 callersMethodreset_parameters

(self)

models/model_pyg.py:33

↓ 10 callersMethod__init__

(self, in_channels, hidden_channels, out_channels, num_layers, dropout, layer_norm_first=Fals

models/model_pyg.py:16

↓ 7 callersMethodreset

(self)

attacks/utils.py:65

↓ 6 callersFunctioninit_feat

(num, features, device, style="sample", feat_lim_min=-1, feat_lim_max=1)

attacks/attack.py:9

↓ 6 callersFunctionrandom_injection

(adj, n_inject, n_edge_max, target_idx, device)

attacks/injection.py:99

↓ 5 callersFunctiongia_update_features

(attacker, model, adj_attack, features, features_attack, origin_labels, target_idx, homophily=None, hinge=Fals

attacks/attack.py:173

↓ 5 callersFunctionsep_test

(model, x, adj_t, y, target_idx, evaluator)

gnn_misg.py:92

↓ 4 callersFunctionadj_to_tensor

r""" Description ----------- Convert adjacency matrix in scipy sparse format to torch sparse tensor. Parameters ---------- a

attacks/utils.py:70

↓ 3 callersFunctiongcn_norm

(adj_t, order=-0.5, add_self_loops=True)

attacks/attack.py:54

↓ 3 callersFunctionnode_sim_estimate

estimate the mean and variance from the observed data points

attacks/attack.py:73

↓ 3 callersFunctionsample_mask

Create mask.

utils.py:390

↓ 2 callersFunctionagia_injection

(attacker, model, adj, n_inject, n_edge_max, features, features_attack, target_idx, origin_labels,

attacks/injection.py:464

↓ 2 callersMethodatt_coef

(self, features, adj)

models/model_pyg.py:612

↓ 2 callersMethodatt_coef

(self, features, adj)

models/model_pyg.py:768

↓ 2 callersMethodattack

(self, model, adj, features, target_idx, labels=None)

attacks/vanilla.py:47

↓ 2 callersFunctioneval_robustness

(model, features, adj, target_idx, labels, device, args, run)

gnn_misg.py:107

↓ 2 callersFunctiongcn_norm

(adj_t, order=-0.5, add_self_loops=True)

models/model_pyg.py:388

↓ 2 callersFunctiongenerate_grb_split

(data,mode='full')

load_graph.py:214

↓ 2 callersFunctionset_rand_seed

(rand_seed)

utils.py:247

↓ 2 callersFunctiontrain

(model, x, adj_t, y, train_idx, optimizer)

gnn_misg.py:36

↓ 1 callersFunctionatdgia_injection

(adj, n_inject, n_edge_max, origin_labels, current_pred, target_idx, device, self_connect_ratio=0

attacks/injection.py:219

↓ 1 callersFunctionatdgia_ranking_select

(adj, n_inject, n_edge_max, origin_labels, current_pred, target_idx, ratio=0.5, neg=False, weight1=0.9, weight

attacks/injection.py:340

↓ 1 callersFunctionedge_sim_analysis

(edge_index, features)

attacks/attack.py:26

↓ 1 callersFunctionfeat_normalize

r""" Description ----------- Feature normalization function. Parameters ---------- features : torch.FloatTensor Featu

utils.py:61

↓ 1 callersFunctionfeat_normalize

(feat, norm=None)

attacks/utils.py:202

↓ 1 callersFunctionflag

(model_forward, clean, y, args, optimizer, device, criterion)

attacks/flag.py:42

↓ 1 callersFunctiongenerate_percent_split

(dataset, seed=0, train_percent=10, val_percent=10)

load_graph.py:172

↓ 1 callersFunctionget_index_induc

r""" Description ----------- Get index under the inductive training setting. Parameters ---------- index_a : tuple T

utils.py:180

↓ 1 callersFunctioninductive_split

inductive split adjs for PyG graphs will automatically use relative ids for splitted graphs

utils.py:231

↓ 1 callersMethodinjection

(self, adj, n_inject, n_node, target_idx, target_node=None, mode='random-inter')

attacks/speit.py:112

↓ 1 callersMethodinjection

r""" Description ----------- Randomly inject nodes to target nodes. Parameters ---------- adj : scip

attacks/pgd.py:105

↓ 1 callersMethodinjection

r""" Description ----------- Randomly inject nodes to target nodes. Parameters ---------- adj : scip

attacks/rnd.py:64

↓ 1 callersFunctionload_cora

(self_loop=True)

load_graph.py:59

↓ 1 callersFunctionload_dblp

(pyg=True, self_loop=True)

load_graph.py:138

↓ 1 callersFunctionload_grb

(name, mode="full", self_loop=True)

load_graph.py:36

↓ 1 callersFunctionload_heter_g

(dataset_name, pyg=False, splits_file_path=None, train_percentage=0.48, val_percentage=0.32, embedding_mode=No

load_graph.py:271

↓ 1 callersFunctionload_ogb

(name, self_loop=True, undirected=True)

load_graph.py:81

↓ 1 callersFunctionload_reddit

(self_loop=True)

load_graph.py:70

↓ 1 callersFunctionmain

()

gnn_misg.py:243

↓ 1 callersFunctionmeta_injection

(attacker, model, adj, n_inject, n_edge_max, features, features_attack, target_idx, origin_labels,

attacks/injection.py:590

↓ 1 callersFunctionnode_sim_analysis

(adj, x)

attacks/attack.py:66

↓ 1 callersFunctionnormalize_adj

Symmetrically normalize adjacency matrix.

utils.py:476

↓ 1 callersMethodopt_adj_attack

(self, model, adj, features, features_attack, target_idx, origin_labels, n_epoch=1e9, homophily=None)

attacks/agia.py:94

↓ 1 callersFunctionparse_index_file

Parse index file.

utils.py:382

↓ 1 callersFunctionprune_graph

(adj_test, target_idx, k)

utils.py:17

↓ 1 callersFunctionrandom_class_injection

(adj, n_inject, n_edge_max, origin_labels, target_idx, device, not_full=False)

attacks/injection.py:52

↓ 1 callersFunctionreproduction_info

()

gnn_misg.py:235

↓ 1 callersMethodreset_parameters

(self)

models/model_pyg.py:71

↓ 1 callersFunctionsmooth_update_features

(attacker, model, adj_attack, features, features_attack, origin_labels, target_idx, homophily=None, n_inject_c

attacks/attack.py:242

↓ 1 callersFunctiontarget_select

(model,adj,features,labels,target_idx,num)

utils.py:25

↓ 1 callersFunctiontdgia_injection

(adj, n_inject, n_edge_max, origin_labels, current_pred, target_idx, device, self_connect_ratio=0

attacks/injection.py:133

↓ 1 callersFunctiontest

(model, x, adj_t, y, split_idx, evaluator)

gnn_misg.py:70

↓ 1 callersFunctiontrain_flag

(model, x, adj_t, y, train_idx, optimizer, device, args)

gnn_misg.py:51

↓ 1 callersMethodupdate_features

r""" Description ----------- Adversarial feature generation of injected nodes. Parameters ----------

attacks/speit.py:257

↓ 1 callersMethodupdate_features

r""" Description ----------- Update features of injected nodes. Parameters ---------- model : torch.

attacks/pgd.py:161

FunctionGCNAdjNorm

(adj, order=-0.5)

models/model_pyg.py:370

Method__call__

r""" Parameters ---------- score : float Value of attack acore.

attacks/utils.py:46

Method__init__

(self, train_mask, val_mask, test_mask)

load_graph.py:167

Method__init__

(self, *args, **kwargs)

utils.py:340

Method__init__

(self, epsilon, n_epoch, n_inject_max, n_e

attacks/speit.py:50

Method__init__

r""" Parameters ---------- patience : int, optional Number of epoch to wait if no further improvement. Default: `

attacks/utils.py:29

Method__init__

(self, epsilon, n_epoch, n_inject_max, n_e

attacks/gia.py:23

Method__init__

(self, epsilon, n_epoch, a_epoch, n_inject

attacks/seqgia.py:23

Method__init__

(self, epsilon, n_epoch, n_inject_max, n_e

attacks/pgd.py:47

Method__init__

(self, epsilon, n_epoch, n_inject_max, n_e

attacks/vanilla.py:18

Method__init__

(self, epsilon, n_epoch, a_epoch, n_inject

attacks/agia.py:24

Method__init__

(self, epsilon, n_epoch, n_inject_max, n_e

attacks/rnd.py:19

Method__init__

(self, in_channels, hidden_channels, out_channels, num_layers, dropout, layer_norm_first=Fals

models/model_pyg.py:53

Method__init__

(self, in_channels, hidden_channels, out_channels, num_layers, dropout, layer_norm_first=Fals

models/model_pyg.py:107

Method__init__

(self, in_channels, hidden_channels, out_channels, num_layers, dropout, layer_norm_first=Fals

models/model_pyg.py:126

Method__init__

(self, in_channels, hidden_channels, out_channels, num_layers, dropout, layer_norm_first=Fals

models/model_pyg.py:164

Method__init__

(self, in_channels, hidden_channels, out_channels, num_layers, dropout, layer_norm_first=Fals

models/model_pyg.py:203

Method__init__

(self, in_channels: Union[int, Tuple[int, int]], out_channels: int, heads: int = 1, threshold

models/model_pyg.py:250

Method__init__

(self, in_channels, hidden_channels, out_channels, num_layers, dropout)

models/model_pyg.py:418

Method__init__

(self, in_features, out_features, act0=F.elu, act1=F.relu, initial=False, dropout=0.5)

models/model_pyg.py:496

Method__init__

(self, in_channels, hidden_channels, out_channels, num_layers, dropout, layer_norm_first=Fals

models/model_pyg.py:561

Method__init__

(self, in_channels, hidden_channels, out_channels, num_layers, dropout, layer_norm_first=Fals

models/model_pyg.py:677

Functionadj_preprocess

r""" Description ----------- Preprocess the adjacency matrix. Parameters ---------- adj : scipy.sparse.csr.csr_matrix or a t

attacks/utils.py:125

Methodattack

(self, model, adj, features, target_idx, labels=None)

attacks/speit.py:81

Methodattack

(self, model, adj, features, target_idx, labels=None)

attacks/gia.py:55

Methodattack

(self, model, adj, features, target_idx, labels=None)

attacks/seqgia.py:69

Methodattack

(self, model, adj, features, target_idx, labels=None)

attacks/pgd.py:76

Methodattack

(self, model, adj, features, target_idx, labels=None)

attacks/agia.py:68

Methodattack

(self, model, adj, features, target_idx, labels=None)

attacks/rnd.py:48

Functioncheck_feat_range

r""" Description ----------- Check if the generated features are within the limited range. Parameters ---------- features :

attacks/utils.py:591

Functioncheck_symmetry

r""" Description ----------- Check if the adjacency matrix is symmetric. Parameters ---------- adj : scipy.sparse.csr.csr_ma

attacks/utils.py:568

Methodcon_forward

(self,x,adj_t,layers=-1)

models/model_pyg.py:92

Functiondeg_estimate

(adj, num, style='sample')

attacks/attack.py:92

Functiondice_injection

(adj, n_inject, n_edge_max, origin_labels, target_idx, device)

attacks/injection.py:11

Functiondownload

r""" Description ----------- Download dataset from URL. Parameters ---------- url : str URL to the dataset. save

attacks/utils.py:428

Functionedge_sim_estimate

estimate the mean and variance from the observed data points

attacks/attack.py:35

Functioneval_acc

r""" Description ----------- Accuracy metric for node classification. Parameters ---------- pred : torch.Tensor Outp

attacks/metric.py:7

Functioneval_f1multilabel

r""" Description ----------- F1 score for multi-label node classification. Parameters ---------- pred : torch.Tensor

attacks/metric.py:85

Functioneval_rocauc

r""" Description ----------- ROC-AUC score for multi-label node classification. Parameters ---------- pred : torch.Tensor

attacks/metric.py:40

Functionextra_misg_ids

sample misclassified training samples and save to args.misg_path

utils.py:255

Functionfeat_preprocess

r""" Description ----------- Preprocess the features. Parameters ---------- features : torch.Tensor or numpy.array F

attacks/utils.py:179

Functionfix_seed

r""" Description ----------- Fix random process by a seed. Parameters ---------- seed : int, optional Random seed. D

attacks/utils.py:255

Functionflag_biased

(model_forward, perturb_shape, y, args, optimizer, device, criterion, training_idx)

attacks/flag.py:9

Methodforward

(self, x, adj_t)

models/model_pyg.py:40

Methodforward

(self, x, adj_t, layers=-1)

models/model_pyg.py:77

Functions151 in github.com/LFhase/GIA-HAO