pyg-team/pytorch_geometric

11.2

Adjusted Score

11.2

Raw Score

100%

Time Factor

2026-07-06

Last Push

23.9K

Stars

Python

Language

196.1K

Lines of Code

1.4K

Files

2.2K

Pattern Hits

2026-07-14

Scan Date

0.02

HC Hit Rate

What These Metrics Mean

Adjusted Score: Primary synthetic code indicator. Raw score normalised per 1,000 lines of code and multiplied by the temporal discount factor. This is the definitive comparative metric — use it to rank repositories by AI authorship density.
Raw Score: The unmodified sum of all severity-weighted, context-multiplied pattern match scores before temporal discounting. Reflects the absolute signal strength independent of when the repository was last active.
Time Factor: The temporal discount multiplier (0–100%) applied to the raw score. Repositories last updated before ChatGPT's launch (Nov 2022) receive a 5% factor. Full signal is only assigned to repositories active in the post-adoption era (Jan 2024+).
Pattern Hits: Total count of individual pattern matches across all files and categories. A high hit count with a low score may indicate a very large codebase with isolated AI snippets; a low count with a high score indicates dense, concentrated AI signatures.
HC Hit Rate: High+Critical pattern hits per file, averaged across the repository. This orthogonal signal catches repositories where a few files are densely packed with high-severity AI tells — a strong indicator even when the normalised score appears moderate due to codebase size.
Lines of Code / Files: Total lines and files analysed. The scanner examines 94 file extensions. These denominators are used to normalise the score, enabling fair comparison between repositories of vastly different sizes.

Score History

This chart maps the temporal evolution of the adjusted synthetic code score across successive scan runs. An upward trajectory indicates ongoing incorporation of AI-generated code or expanding LLM-assisted scaffolding; a stable or declining trajectory may reflect active human refactoring, code removal, or the adoption of stricter authorship policies. The dashed secondary line (right axis) independently tracks total raw pattern hit count, which can diverge from the normalised score when codebase size changes significantly between scans.

Severity Breakdown

Classifies detected patterns by their diagnostic confidence and structural impact. CRITICAL patterns (coefficient 10) represent definitive synthetic signatures — hallucinated imports, explicit LLM attribution metadata — virtually never produced by human authors. HIGH (5) indicates strong structural tells such as cross-file repetition or cross-linguistic idioms. MEDIUM (2) covers recognisable conversational padding and AI-specific vocabulary. LOW (1) captures subtle indicators like tautological comments and generic boilerplate that require density to carry independent signal.

CRITICAL 4HIGH 22MEDIUM 72LOW 2087

Directory Score Breakdown

This horizontal bar chart decomposes the repository's raw synthetic code score by top-level directory, allowing you to pinpoint precisely which modules or components carry the highest AI authorship density. Directories with disproportionately high scores relative to their size warrant targeted manual review: concentrated AI signatures often trace back to mass-generated configuration layers, auto-ported test suites, LLM-scaffolded boilerplate classes, or entire subsystems authored under heavy copilot assistance. Use this view to prioritise your human code-review effort.

Pattern Findings

The scanner identified 2185 distinct pattern matches across 22 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a statistically significant AI authorship indicator. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.

Reading the findings table: The Severity column indicates the diagnostic confidence level (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — comment-context matches receive a ×1.5 weight because LLMs systematically over-annotate. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier as dense clusters constitute far stronger evidence of synthetic authorship than isolated hits.

Unused Imports885 hits · 714 pts

Severity	File	Line	Context
LOW	benchmark/citation/__init__.py	1	CODE
LOW	benchmark/citation/__init__.py	2	CODE
LOW	benchmark/citation/__init__.py	2	CODE
LOW	benchmark/runtime/__init__.py	1	CODE
LOW	benchmark/utils/__init__.py	1	CODE
LOW	benchmark/utils/__init__.py	2	CODE
LOW	benchmark/utils/__init__.py	2	CODE
LOW	benchmark/utils/__init__.py	3	CODE
LOW	benchmark/utils/__init__.py	4	CODE
LOW	benchmark/utils/__init__.py	5	CODE
LOW	benchmark/utils/__init__.py	5	CODE
LOW	benchmark/utils/__init__.py	6	CODE
LOW	benchmark/utils/utils.py	19	CODE
LOW	benchmark/multi_gpu/training/training_benchmark_xpu.py	5	CODE
LOW	benchmark/points/__init__.py	1	CODE
LOW	benchmark/points/__init__.py	2	CODE
LOW	benchmark/kernel/__init__.py	1	CODE
LOW	benchmark/kernel/__init__.py	2	CODE
LOW	test/testing/test_decorators.py	7	CODE
LOW	test/testing/test_decorators.py	13	CODE
LOW	test/testing/test_decorators.py	19	CODE
LOW	graphgym/main.py	4	CODE
LOW	graphgym/custom_graphgym/__init__.py	1	CODE
LOW	graphgym/custom_graphgym/__init__.py	2	CODE
LOW	graphgym/custom_graphgym/__init__.py	3	CODE
LOW	graphgym/custom_graphgym/__init__.py	4	CODE
LOW	graphgym/custom_graphgym/__init__.py	5	CODE
LOW	graphgym/custom_graphgym/__init__.py	6	CODE
LOW	graphgym/custom_graphgym/__init__.py	7	CODE
LOW	graphgym/custom_graphgym/__init__.py	8	CODE
LOW	graphgym/custom_graphgym/__init__.py	9	CODE
LOW	graphgym/custom_graphgym/__init__.py	10	CODE
LOW	graphgym/custom_graphgym/__init__.py	11	CODE
LOW	graphgym/custom_graphgym/__init__.py	12	CODE
LOW	graphgym/custom_graphgym/__init__.py	13	CODE
LOW	graphgym/custom_graphgym/network/example.py	5	CODE
LOW	torch_geometric/experimental.py	9	CODE
LOW	torch_geometric/__init__.py	6	CODE
LOW	torch_geometric/__init__.py	6	CODE
LOW	torch_geometric/__init__.py	7	CODE
LOW	torch_geometric/__init__.py	7	CODE
LOW	torch_geometric/__init__.py	11	CODE
LOW	torch_geometric/__init__.py	12	CODE
LOW	torch_geometric/__init__.py	12	CODE
LOW	torch_geometric/__init__.py	13	CODE
LOW	torch_geometric/__init__.py	13	CODE
LOW	torch_geometric/__init__.py	13	CODE
LOW	torch_geometric/__init__.py	14	CODE
LOW	torch_geometric/__init__.py	15	CODE
LOW	torch_geometric/__init__.py	15	CODE
LOW	torch_geometric/__init__.py	15	CODE
LOW	torch_geometric/__init__.py	27	CODE
LOW	torch_geometric/__init__.py	27	CODE
LOW	torch_geometric/__init__.py	27	CODE
LOW	torch_geometric/config_store.py	22	CODE
LOW	torch_geometric/typing.py	166	CODE
LOW	torch_geometric/typing.py	311	CODE
LOW	torch_geometric/visualization/__init__.py	3	CODE
LOW	torch_geometric/visualization/__init__.py	3	CODE
LOW	torch_geometric/visualization/__init__.py	4	CODE
825 more matches not shown…

Hyper-Verbose Identifiers653 hits · 627 pts

Severity	File	Line	Snippet	Context
LOW	benchmark/utils/utils.py	39	def get_dataset_with_transformation(name, root, use_sparse_tensor=False,	CODE
LOW	benchmark/kernel/train_eval.py	20	def cross_validation_with_val_set(dataset, model, folds, epochs, batch_size,	CODE
LOW	test/test_warnings.py	15	def test_no_warn_if_compiling(_):	CODE
LOW	test/test_onnx.py	32	def test_safe_onnx_export_ci_resilient() -> None:	CODE
LOW	test/test_onnx.py	62	def test_safe_onnx_export_success() -> None:	CODE
LOW	test/test_onnx.py	107	def test_safe_onnx_export_with_skip_on_error() -> None:	CODE
LOW	test/test_onnx.py	134	def test_serde_error_patterns() -> None:	CODE
LOW	test/test_onnx.py	173	def test_non_serde_error_reraise() -> None:	CODE
LOW	test/test_onnx.py	303	def test_pytest_environment_detection() -> None:	CODE
LOW	test/test_config_mixin.py	116	def test_config_mixin_compound() -> None:	CODE
LOW	test/test_edge_index.py	1375	def test_compile_create_edge_index():	CODE
LOW	test/test_inspector.py	102	def test_get_params_from_method_call() -> None:	CODE
LOW	test/visualization/test_graph_visualization.py	12	def test_visualize_graph_via_graphviz(tmp_path, backend):	CODE
LOW	test/visualization/test_graph_visualization.py	26	def test_visualize_graph_via_graphviz_with_node_labels(tmp_path, backend):	CODE
LOW	test/visualization/test_graph_visualization.py	41	def test_visualize_graph_via_networkx(tmp_path, backend):	CODE
LOW	test/llm/conftest.py	8	def pytest_collection_modifyitems(items):	CODE
LOW	test/llm/test_rag_loader.py	97	def test_initialization_basic(self):	CODE
LOW	test/llm/test_rag_loader.py	108	def test_initialization_with_all_params(self):	CODE
LOW	test/llm/test_rag_loader.py	140	def test_basic_query_without_vector_retriever(self):	CODE
LOW	test/llm/test_rag_loader.py	156	def test_query_with_vector_retriever(self):	CODE
LOW	test/llm/test_rag_loader.py	177	def test_query_with_subgraph_filter(self):	CODE
LOW	test/llm/test_rag_loader.py	204	def test_rag_loader_integration(tmp_path):	CODE
LOW	test/llm/test_large_graph_indexer.py	130	def _sorted_tensors_are_close(tensor1, tensor2):	CODE
LOW	test/llm/test_txt2qa.py	67	def test_route_after_validation_continue(self):	CODE
LOW	test/llm/test_txt2qa.py	73	def test_route_after_validation_feedback(self):	CODE
LOW	test/llm/test_txt2qa.py	95	def test_generate_single_query_nim(tmp_path):	CODE
LOW	test/llm/utils/test_rag_backend_utils.py	101	def create_mock_graph_and_triples():	CODE
LOW	test/llm/utils/test_rag_backend_utils.py	121	def test_apply_retrieval_via_pcst_isolated_node():	CODE
LOW	test/llm/utils/test_rag_backend_utils.py	171	def test_create_graph_basic_functionality(self):	CODE
LOW	test/llm/utils/test_rag_backend_utils.py	192	def test_create_graph_empty_triples(self):	CODE
LOW	test/llm/utils/test_rag_backend_utils.py	219	def test_create_backend_data_load(self):	CODE
LOW	test/llm/utils/test_rag_feature_store.py	43	def test_retrieve_seed_nodes_single_query(self):	CODE
LOW	test/llm/utils/test_rag_feature_store.py	78	def test_retrieve_seed_nodes_multiple_queries(self):	CODE
LOW	test/llm/utils/test_rag_feature_store.py	112	def test_load_subgraph_valid_sample(self, induced):	CODE
LOW	test/llm/utils/test_rag_graph_store.py	18	def test_sample_subgraph_with_valid_tensor_input():	CODE
LOW⚡	test/llm/models/test_glem.py	10	def test_deal_nan_tensor_replaces_nans():	CODE
LOW⚡	test/llm/models/test_glem.py	20	def test_deal_nan_non_tensor_passthrough():	CODE
LOW⚡	test/llm/models/test_glem.py	26	def test_deal_nan_tensor_dtypes(dtype):	CODE
LOW	test/llm/models/test_glem.py	46	def test_deal_nan_is_non_mutating():	CODE
LOW	test/llm/models/test_glem.py	76	def test_glem_bert_tiny_branch(monkeypatch):	CODE
LOW	test/llm/models/test_glem.py	101	def test_glem_auto_model_branch(monkeypatch):	CODE
LOW	test/llm/models/test_llm.py	127	def test_llm_variable_lengths(dummy_llm):	CODE
LOW	test/llm/models/test_sentence_transformer.py	17	def test_sentence_transformer(batch_size, pooling_strategy, device, verbose):	CODE
LOW	test/llm/models/test_g_retriever.py	63	def test_g_retriever_many_tokens() -> None:	CODE
LOW	test/llm/models/test_g_retriever.py	160	def test_gretriever_prefix_embedding_injection(batch_size):	CODE
LOW	test/llm/models/test_txt2kg.py	22	def test_parse_n_check_triples_formats():	CODE
LOW	test/llm/models/test_txt2kg.py	29	def test_chunk_text_simple_sentence():	CODE
LOW	test/llm/models/test_txt2kg.py	44	def test_local_lm_integration(monkeypatch):	CODE
LOW	test/llm/models/test_txt2kg.py	83	def test_merge_triples_deterministically_basic():	CODE
LOW	test/llm/models/test_txt2kg.py	101	def test_merge_triples_deterministically_unicode_and_nonstring():	CODE
LOW⚡	test/llm/models/test_txt2kg.py	119	def test_merge_triples_deterministically_empty():	CODE
LOW⚡	test/llm/models/test_txt2kg.py	127	def test_merge_triples_deterministically_singleton():	CODE
LOW⚡	test/llm/models/test_txt2kg.py	135	def test_chunk_to_triples_str_cloud(monkeypatch):	CODE
LOW	test/llm/models/test_txt2kg.py	191	def test_extract_relevant_triples_cloud(monkeypatch):	CODE
LOW	test/llm/models/test_txt2kg.py	202	def test_multiproc_helper_success(monkeypatch):	CODE
LOW	test/llm/models/test_txt2kg.py	233	def test_multiproc_helper_retry(monkeypatch):	CODE
LOW	test/llm/models/test_txt2kg.py	292	def test_add_doc_empty_text_cpu(kg_cpu):	CODE
LOW	test/llm/models/test_txt2kg.py	308	def test_add_doc_nonempty_text_placeholder(kg_cpu, monkeypatch):	CODE
LOW	test/explain/conftest.py	139	def _check_explanation_hetero(	CODE
LOW	test/explain/test_explanation.py	13	def create_random_explanation(	CODE
593 more matches not shown…

Deep Nesting245 hits · 238 pts

Severity	File	Line	Context
LOW	benchmark/citation/train_eval.py	43	CODE
LOW	benchmark/citation/train_eval.py	121	CODE
LOW	benchmark/training/training_benchmark.py	98	CODE
LOW	benchmark/loader/neighbor_loader.py	16	CODE
LOW	benchmark/points/train_eval.py	58	CODE
LOW	benchmark/inference/inference_benchmark.py	44	CODE
LOW	benchmark/kernel/main_performance.py	75	CODE
LOW	benchmark/kernel/main_performance.py	117	CODE
LOW	benchmark/kernel/train_eval.py	20	CODE
LOW	test/test_onnx.py	32	CODE
LOW	test/test_onnx.py	62	CODE
LOW	test/test_onnx.py	107	CODE
LOW	test/test_onnx.py	134	CODE
LOW	test/test_onnx.py	173	CODE
LOW	test/test_onnx.py	197	CODE
LOW	test/test_onnx.py	233	CODE
LOW	test/test_onnx.py	272	CODE
LOW	test/test_onnx.py	303	CODE
LOW	test/test_onnx.py	335	CODE
LOW	test/test_onnx.py	385	CODE
LOW	test/explain/conftest.py	106	CODE
LOW	test/explain/conftest.py	138	CODE
LOW	test/explain/conftest.py	86	CODE
LOW	test/explain/conftest.py	107	CODE
LOW	test/explain/conftest.py	139	CODE
LOW	test/explain/conftest.py	289	CODE
LOW	test/explain/test_hetero_explanation.py	13	CODE
LOW	test/explain/algorithm/test_captum.py	65	CODE
LOW	test/explain/algorithm/test_attention_explainer.py	69	CODE
LOW	test/explain/algorithm/test_attention_explainer.py	145	CODE
LOW	test/explain/algorithm/test_graphmask_explainer.py	28	CODE
LOW	test/explain/algorithm/test_gnn_explainer.py	38	CODE
LOW	test/explain/algorithm/test_captum_hetero.py	35	CODE
LOW	test/explain/algorithm/test_captum_explainer.py	54	CODE
LOW	test/distributed/test_dist_neighbor_loader.py	82	CODE
LOW	test/utils/conftest.py	35	CODE
LOW	test/sampler/test_sampler_neighbor_sampler.py	63	CODE
LOW	graphgym/configs_gen.py	90	CODE
LOW	graphgym/configs_gen.py	124	CODE
LOW	graphgym/configs_gen.py	166	CODE
LOW	torch_geometric/experimental.py	95	CODE
LOW	torch_geometric/experimental.py	100	CODE
LOW	torch_geometric/experimental.py	115	CODE
LOW	torch_geometric/index.py	696	CODE
LOW	torch_geometric/edge_index.py	1477	CODE
LOW	torch_geometric/edge_index.py	1849	CODE
LOW	torch_geometric/inspector.py	514	CODE
LOW	torch_geometric/inspector.py	207	CODE
LOW	torch_geometric/inspector.py	336	CODE
LOW	torch_geometric/config_store.py	196	CODE
LOW	torch_geometric/config_store.py	30	CODE
LOW	torch_geometric/config_store.py	118	CODE
LOW	torch_geometric/config_store.py	35	CODE
LOW	torch_geometric/config_store.py	123	CODE
LOW	torch_geometric/hash_tensor.py	39	CODE
LOW	torch_geometric/llm/models/qa_gen.py	1384	CODE
LOW	torch_geometric/llm/models/qa_gen.py	1715	CODE
LOW	torch_geometric/llm/models/qa_gen.py	2239	CODE
LOW	torch_geometric/llm/models/qa_gen.py	2350	CODE
LOW	torch_geometric/llm/models/qa_gen.py	2414	CODE
185 more matches not shown…

Cross-File Repetition18 hits · 90 pts

Severity	File	Snippet	Context
HIGH	torch_geometric/nn/attention/performer.py	forward pass. args: x (torch.tensor): node feature tensor :math:`\mathbf{x} \in \mathbb{r}^{b \times n \times f}`, with	STRING
HIGH	torch_geometric/nn/attention/polynormer.py	forward pass. args: x (torch.tensor): node feature tensor :math:`\mathbf{x} \in \mathbb{r}^{b \times n \times f}`, with	STRING
HIGH	torch_geometric/nn/attention/sgformer.py	forward pass. args: x (torch.tensor): node feature tensor :math:`\mathbf{x} \in \mathbb{r}^{b \times n \times f}`, with	STRING
HIGH	torch_geometric/nn/dense/dense_gat_conv.py	forward pass. args: x (torch.tensor): node feature tensor :math:`\mathbf{x} \in \mathbb{r}^{b \times n \times f}`, with	STRING
HIGH	torch_geometric/nn/dense/dense_gcn_conv.py	forward pass. args: x (torch.tensor): node feature tensor :math:`\mathbf{x} \in \mathbb{r}^{b \times n \times f}`, with	STRING
HIGH	torch_geometric/nn/dense/dense_gin_conv.py	forward pass. args: x (torch.tensor): node feature tensor :math:`\mathbf{x} \in \mathbb{r}^{b \times n \times f}`, with	STRING
HIGH	torch_geometric/nn/models/deep_graph_infomax.py	evaluates latent space quality via a logistic regression downstream task.	STRING
HIGH	torch_geometric/nn/models/node2vec.py	evaluates latent space quality via a logistic regression downstream task.	STRING
HIGH	torch_geometric/nn/models/metapath2vec.py	evaluates latent space quality via a logistic regression downstream task.	STRING
HIGH	torch_geometric/nn/conv/gatv2_conv.py	runs the forward pass of the module. args: x (torch.tensor or (torch.tensor, torch.tensor)): the input node features. ed	STRING
HIGH	torch_geometric/nn/conv/transformer_conv.py	runs the forward pass of the module. args: x (torch.tensor or (torch.tensor, torch.tensor)): the input node features. ed	STRING
HIGH	torch_geometric/nn/conv/fa_conv.py	runs the forward pass of the module. args: x (torch.tensor or (torch.tensor, torch.tensor)): the input node features. ed	STRING
HIGH	torch_geometric/nn/conv/gat_conv.py	runs the forward pass of the module. args: x (torch.tensor or (torch.tensor, torch.tensor)): the input node features. ed	STRING
HIGH	torch_geometric/nn/norm/pair_norm.py	forward pass. args: x (torch.tensor): the source tensor. batch (torch.tensor, optional): the batch vector :math:`\mathbf	STRING
HIGH	torch_geometric/nn/norm/graph_size_norm.py	forward pass. args: x (torch.tensor): the source tensor. batch (torch.tensor, optional): the batch vector :math:`\mathbf	STRING
HIGH	torch_geometric/nn/norm/instance_norm.py	forward pass. args: x (torch.tensor): the source tensor. batch (torch.tensor, optional): the batch vector :math:`\mathbf	STRING
HIGH	torch_geometric/nn/norm/graph_norm.py	forward pass. args: x (torch.tensor): the source tensor. batch (torch.tensor, optional): the batch vector :math:`\mathbf	STRING
HIGH	torch_geometric/nn/norm/layer_norm.py	forward pass. args: x (torch.tensor): the source tensor. batch (torch.tensor, optional): the batch vector :math:`\mathbf	STRING

Self-Referential Comments31 hits · 86 pts

Severity	File	Line	Snippet	Context
MEDIUM	test/llm/utils/test_rag_feature_store.py	116	# Create a mock SamplerOutput	COMMENT
MEDIUM⚡	test/llm/models/test_glem.py	27	# Create a tensor with one NaN value	COMMENT
MEDIUM	test/explain/conftest.py	268	# Create a HeteroConv model	COMMENT
MEDIUM	test/explain/algorithm/test_attention_explainer.py	44	# Create a single BaseGNN that uses all three attention mechanisms	COMMENT
MEDIUM	test/explain/algorithm/test_attention_explainer.py	105	# Create a dictionary of attention-based convolutions for each edge	COMMENT
MEDIUM	test/explain/algorithm/test_attention_explainer.py	122	# Create the HeteroConv layer	COMMENT
MEDIUM	test/explain/algorithm/test_attention_explainer.py	251	# Create the hetero attention model	COMMENT
MEDIUM	test/explain/algorithm/test_attention_explainer.py	254	# Create the explainer	COMMENT
MEDIUM	test/explain/algorithm/test_attention_explainer.py	290	# Create the hetero conv attention model	COMMENT
MEDIUM	test/explain/algorithm/test_attention_explainer.py	293	# Create the explainer	COMMENT
MEDIUM	test/utils/test_negative_sampling.py	33	# Create a fully-connected graph:	COMMENT
MEDIUM	torch_geometric/llm/utils/vectorrag.py	121	# Create a new DocumentRetriever with the loaded data	COMMENT
MEDIUM	torch_geometric/nn/to_hetero_with_bases_transformer.py	537	# This method is used to split the output tensors into individual type-wise	STRING
MEDIUM	torch_geometric/nn/aggr/base.py	47	# Define the boundary indices for three sets:	COMMENT
MEDIUM	torch_geometric/datasets/web_qsp_dataset.py	124	# Create an iterator for each element's graph in the current split	STRING
MEDIUM	torch_geometric/datasets/ged_dataset.py	208	# Create a one-hot encoded feature matrix denoting the atom	STRING
MEDIUM	torch_geometric/data/remote_backend_utils.py	1	# This file defines a set of utilities for remote backends (backends that are	COMMENT
MEDIUM	torch_geometric/data/in_memory_dataset.py	232	# Create the on-disk dataset ==========================================	STRING
MEDIUM	torch_geometric/data/database.py	307	# Create the table (if it does not exist) by mapping the Python schema	STRING
MEDIUM	torch_geometric/data/hetero_data.py	63	# Create an edge type "(author, writes, paper)" and building the	COMMENT
MEDIUM	torch_geometric/data/lightning/datamodule.py	127	# Define a new `NeighborSampler` to be re-used across data loaders:	COMMENT
MEDIUM	examples/llm/relbench_gretriever.py	88	# Define the projector for heterogeneous node features	COMMENT
MEDIUM	examples/llm/txt2qa.py	158	# Create the graph	COMMENT
MEDIUM	examples/llm/g_retriever.py	62	# Create a dictionary mapping parameter names to their requires_grad status	COMMENT
MEDIUM	examples/llm/txt2qa_config/text_config_nim.yaml	2	# This file contains all the settings for the QA generation workflow	COMMENT
MEDIUM	examples/llm/txt2qa_config/text_config_vllm.yaml	2	# This file contains all the settings for the QA generation workflow	COMMENT
MEDIUM	examples/distributed/kuzu/papers_100M/train.py	32	# The below code sets up the remote backend of Kùzu for PyG.	COMMENT
MEDIUM	…d/graphlearn_for_pytorch/dist_train_sage_supervised.py	126	# Define the model and optimizer.	COMMENT
MEDIUM	examples/quiver/multi_gpu_quiver.py	131	# The below code enable Quiver for PyG.	COMMENT
MEDIUM⚡	examples/quiver/single_gpu_quiver.py	28	# The below code enable Quiver for PyG.	COMMENT
MEDIUM	examples/pytorch_ignite/gin.py	127	# Create a tensorboard logger to write logs:	COMMENT

AI Structural Patterns81 hits · 80 pts

Severity	File	Line	Context
LOW	benchmark/runtime/dgl/rgcn.py	66	CODE
LOW	benchmark/runtime/dgl/rgcn.py	135	CODE
LOW	test/test_edge_index.py	1307	CODE
LOW	test/test_edge_index.py	1332	CODE
LOW	test/test_edge_index.py	1350	CODE
LOW	torch_geometric/typing.py	177	CODE
LOW	torch_geometric/visualization/graph.py	155	CODE
LOW	torch_geometric/llm/utils/backend_utils.py	319	CODE
LOW	torch_geometric/llm/models/sentence_transformer.py	169	CODE
LOW	torch_geometric/llm/models/qa_gen.py	2414	CODE
LOW	torch_geometric/llm/models/qa_gen.py	168	CODE
LOW	torch_geometric/llm/models/glem.py	46	CODE
LOW	torch_geometric/llm/models/protein_mpnn.py	188	CODE
LOW	…rch_geometric/explain/algorithm/graphmask_explainer.py	83	CODE
LOW	…rch_geometric/explain/algorithm/graphmask_explainer.py	140	CODE
LOW	torch_geometric/nn/attention/performer.py	50	CODE
LOW	torch_geometric/nn/attention/performer.py	63	CODE
LOW	torch_geometric/nn/attention/performer.py	95	CODE
LOW	torch_geometric/nn/attention/performer.py	173	CODE
LOW	torch_geometric/nn/attention/qformer.py	66	CODE
LOW	torch_geometric/nn/dense/mincut_pool.py	117	CODE
LOW	torch_geometric/nn/models/lpformer.py	720	CODE
LOW	torch_geometric/nn/models/graph_mixer.py	133	CODE
LOW	torch_geometric/nn/models/gpse.py	494	CODE
LOW	torch_geometric/nn/models/visnet.py	753	CODE
LOW	torch_geometric/nn/models/visnet.py	1079	CODE
LOW	torch_geometric/nn/models/mlp.py	78	CODE
LOW	torch_geometric/nn/models/dimenet.py	126	CODE
LOW	torch_geometric/nn/models/dimenet.py	501	CODE
LOW	torch_geometric/nn/models/dimenet.py	768	CODE
LOW	torch_geometric/nn/models/basic_gnn.py	70	CODE
LOW	torch_geometric/nn/models/polynormer.py	45	CODE
LOW	torch_geometric/nn/models/attentive_fp.py	50	CODE
LOW	torch_geometric/nn/models/schnet.py	93	CODE
LOW	torch_geometric/nn/conv/gatv2_conv.py	133	CODE
LOW	torch_geometric/nn/conv/general_conv.py	69	CODE
LOW	torch_geometric/nn/conv/x_conv.py	157	CODE
LOW	torch_geometric/nn/conv/point_transformer_conv.py	138	CODE
LOW	torch_geometric/nn/conv/supergat_conv.py	128	CODE
LOW	torch_geometric/nn/conv/cg_conv.py	91	CODE
LOW	torch_geometric/nn/conv/gat_conv.py	130	CODE
LOW	torch_geometric/nn/conv/dna_conv.py	66	CODE
LOW	torch_geometric/nn/conv/dna_conv.py	161	CODE
LOW	torch_geometric/nn/conv/transformer_conv.py	283	CODE
LOW	torch_geometric/nn/conv/pan_conv.py	130	CODE
LOW	torch_geometric/nn/conv/pna_conv.py	91	CODE
LOW	torch_geometric/nn/conv/rgat_conv.py	174	CODE
LOW	torch_geometric/nn/conv/cluster_gcn_conv.py	124	CODE
LOW	torch_geometric/nn/conv/gps_conv.py	62	CODE
LOW	torch_geometric/nn/conv/gen_conv.py	120	CODE
LOW	torch_geometric/nn/aggr/basic.py	139	CODE
LOW	…rch_geometric/distributed/dist_link_neighbor_loader.py	42	CODE
LOW	torch_geometric/distributed/dist_neighbor_loader.py	42	CODE
LOW	torch_geometric/distributed/dist_neighbor_sampler.py	70	CODE
LOW	torch_geometric/datasets/graphland.py	227	CODE
LOW	torch_geometric/datasets/fake.py	43	CODE
LOW	torch_geometric/datasets/fake.py	143	CODE
LOW	torch_geometric/datasets/opf.py	69	CODE
LOW	torch_geometric/datasets/protein_mpnn_dataset.py	75	CODE
LOW	torch_geometric/io/obj.py	42	CODE
21 more matches not shown…

Excessive Try-Catch Wrapping94 hits · 77 pts

Severity	File	Line	Snippet	Context
LOW	test/nn/dense/test_linear.py	274	except Exception:	CODE
LOW	torch_geometric/device.py	12	except Exception:	CODE
LOW	torch_geometric/config_store.py	18	except Exception:	CODE
LOW	torch_geometric/config_store.py	24	except Exception:	CODE
LOW	torch_geometric/_onnx.py	77	except Exception as e:	STRING
LOW	torch_geometric/_onnx.py	130	except Exception:	STRING
LOW	torch_geometric/_onnx.py	148	except Exception:	STRING
LOW	torch_geometric/_onnx.py	164	except Exception:	STRING
LOW	torch_geometric/_onnx.py	186	except Exception:	STRING
LOW⚡	torch_geometric/typing.py	82	except Exception:	CODE
LOW⚡	torch_geometric/typing.py	87	except Exception:	CODE
LOW⚡	torch_geometric/typing.py	89	except Exception as e:	CODE
LOW	torch_geometric/typing.py	141	except Exception as e:	CODE
LOW	torch_geometric/typing.py	169	except Exception as e:	CODE
LOW	torch_geometric/typing.py	302	except Exception:	CODE
LOW	torch_geometric/typing.py	313	except Exception:	CODE
LOW	torch_geometric/hash_tensor.py	46	except Exception:	CODE
LOW	torch_geometric/metrics/link_pred.py	13	except Exception:	CODE
LOW⚡	torch_geometric/llm/models/qa_gen.py	1154	except Exception as e:	CODE
LOW⚡	torch_geometric/llm/models/qa_gen.py	1163	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	613	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	635	except Exception:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	780	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	821	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	834	except Exception:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	852	except Exception as e2:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	1109	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	1344	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	1501	except Exception:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	1551	except Exception:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	1702	except Exception:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	2753	except Exception:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	2905	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	3022	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	3061	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	3323	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	3492	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	3515	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	3635	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	3854	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	3881	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	4201	except Exception as e:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	4376	except Exception as e:	CODE
LOW⚡	torch_geometric/llm/models/qa_gen.py	3586	except Exception as e:	STRING
LOW⚡	torch_geometric/llm/models/qa_gen.py	3596	except Exception as e1:	STRING
LOW	torch_geometric/llm/models/qa_gen.py	2231	except Exception:	STRING
LOW	torch_geometric/llm/models/qa_gen.py	2369	except Exception:	STRING
LOW	torch_geometric/llm/models/qa_gen.py	3750	except Exception as e:	STRING
LOW	torch_geometric/llm/models/txt2kg.py	195	except Exception as e:	CODE
LOW	torch_geometric/llm/models/txt2kg.py	317	except Exception:	CODE
LOW	torch_geometric/llm/models/txt2kg.py	334	except Exception:	CODE
LOW	…rch_geometric/explain/algorithm/graphmask_explainer.py	410	except Exception:	STRING
LOW	…rch_geometric/explain/algorithm/graphmask_explainer.py	416	except Exception:	STRING
LOW	torch_geometric/nn/sequential.py	265	except Exception as e: # pragma: no cover	CODE
LOW	torch_geometric/nn/conv/message_passing.py	961	except Exception as e: # pragma: no cover	STRING
LOW	torch_geometric/nn/conv/message_passing.py	994	except Exception as e: # pragma: no cover	STRING
LOW	torch_geometric/distributed/event_loop.py	22	except Exception as e:	CODE
LOW	torch_geometric/distributed/event_loop.py	84	except Exception as e:	STRING
MEDIUM	torch_geometric/distributed/event_loop.py	19	def on_done(*_):	CODE
LOW	torch_geometric/distributed/dist_neighbor_sampler.py	392	except Exception:	STRING
34 more matches not shown…

Modern Structural Boilerplate66 hits · 60 pts

Severity	File	Line	Snippet	Context
LOW	benchmark/citation/__init__.py	4	__all__ = [	CODE
LOW	benchmark/runtime/__init__.py	3	__all__ = [	CODE
LOW	benchmark/utils/__init__.py	8	__all__ = [	CODE
LOW	benchmark/points/__init__.py	4	__all__ = [	CODE
LOW	benchmark/kernel/__init__.py	4	__all__ = [	CODE
LOW	graphgym/custom_graphgym/pooling/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/optimizer/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/loss/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/config/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/encoder/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/network/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/act/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/head/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/layer/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/loader/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/train/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/stage/__init__.py	5	__all__ = [	CODE
LOW	graphgym/custom_graphgym/transform/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/experimental.py	37	def set_experimental_mode_enabled(mode: bool, options: Options = None) -> None:	STRING
LOW	torch_geometric/home.py	23	def set_home_dir(path: str) -> None:	STRING
LOW	torch_geometric/__init__.py	36	__all__ = [	CODE
LOW	torch_geometric/debug.py	11	def set_debug_enabled(mode: bool) -> None:	CODE
LOW	torch_geometric/visualization/__init__.py	6	__all__ = [	CODE
LOW	torch_geometric/llm/utils/feature_store.py	31	def _set_from_config(self, config: Dict[str, Any], attr_name: str) -> None:	CODE
LOW	torch_geometric/llm/utils/graph_store.py	54	def _set_from_config(self, config: Dict[str, Any], attr_name: str) -> None:	CODE
LOW	torch_geometric/llm/models/qa_gen.py	24	logger = logging.getLogger(__name__)	CODE
LOW	torch_geometric/llm/models/qa_gen.py	329	def update_generation_model(self, model: str) -> None:	CODE
LOW	torch_geometric/explain/__init__.py	7	__all__ = [	CODE
LOW	torch_geometric/nn/__init__.py	22	__all__ = [	CODE
LOW	torch_geometric/nn/sequential.py	240	def _set_jittable_template(self, raise_on_error: bool = False) -> None:	CODE
LOW	torch_geometric/nn/unpool/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/nn/dense/__init__.py	17	__all__ = [	STRING
LOW	torch_geometric/nn/models/basic_gnn.py	692	__all__ = [	STRING
LOW	torch_geometric/nn/conv/__init__.py	68	__all__ = [	CODE
LOW	torch_geometric/nn/conv/message_passing.py	926	def _set_jittable_templates(self, raise_on_error: bool = False) -> None:	STRING
LOW	torch_geometric/nn/conv/cugraph/__init__.py	6	__all__ = [	CODE
LOW	torch_geometric/nn/conv/utils/__init__.py	14	__all__ = [	CODE
LOW	torch_geometric/nn/norm/__init__.py	13	__all__ = [	CODE
LOW	torch_geometric/nn/pool/__init__.py	378	__all__ = [	STRING
LOW	torch_geometric/nn/pool/select/__init__.py	10	__all__ = [	STRING
LOW	torch_geometric/nn/pool/connect/__init__.py	10	__all__ = [	STRING
LOW	torch_geometric/datasets/utils/__init__.py	3	__all__ = [	CODE
LOW	torch_geometric/io/__init__.py	10	__all__ = [	CODE
LOW	torch_geometric/utils/__init__.py	62	__all__ = [	CODE
LOW	torch_geometric/contrib/__init__.py	13	__all__ = []	CODE
LOW	torch_geometric/contrib/nn/__init__.py	4	__all__ = []	CODE
LOW	torch_geometric/contrib/nn/models/mgnan.py	9	__all__ = [	CODE
LOW	torch_geometric/contrib/utils/__init__.py	3	__all__ = ['from_relbench']	CODE
LOW	torch_geometric/graphgym/utils/__init__.py	8	__all__ = [	CODE
LOW	torch_geometric/graphgym/models/__init__.py	10	__all__ = [	CODE
LOW	torch_geometric/graphgym/contrib/pooling/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/graphgym/contrib/optimizer/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/graphgym/contrib/loss/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/graphgym/contrib/config/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/graphgym/contrib/encoder/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/graphgym/contrib/network/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/graphgym/contrib/act/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/graphgym/contrib/head/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/graphgym/contrib/layer/__init__.py	5	__all__ = [	CODE
LOW	torch_geometric/graphgym/contrib/loader/__init__.py	5	__all__ = [	CODE
6 more matches not shown…

Hallucination Indicators4 hits · 40 pts

Severity	File	Line	Snippet	Context
CRITICAL	test/test_config_store.py	130	assert (cfg.dataset.transform.NormalizeFeatures._target_.split('.')[-1] ==	CODE
CRITICAL	test/test_config_store.py	135	assert (cfg.dataset.transform.AddSelfLoops._target_.split('.')[-1] ==	CODE
CRITICAL	test/data/lightning/test_datamodule.py	49	self.trainer.datamodule.train_dataset._data.x.add_(1)	CODE
CRITICAL	test/data/lightning/test_datamodule.py	176	self.trainer.datamodule.data.x.add_(1)	CODE

AI Slop Vocabulary19 hits · 34 pts

Severity	File	Line	Snippet	Context
MEDIUM	test/test_onnx.py	67	# Use comprehensive mocking to avoid any real ONNX calls	COMMENT
MEDIUM	test/test_onnx.py	146	# Use multiple patch targets to ensure comprehensive mocking	COMMENT
MEDIUM	test/test_onnx.py	178	# Use comprehensive mocking to prevent real ONNX calls	COMMENT
MEDIUM	torch_geometric/backend.py	8	# If set to `None`, will automatically decide whether to utilize	COMMENT
LOW	torch_geometric/llm/models/qa_gen.py	2816	# No hop contexts, just use top-level as-is	COMMENT
MEDIUM	torch_geometric/nn/sequential.py	31	r"""An extension of the :class:`torch.nn.Sequential` container in order to	CODE
LOW	torch_geometric/datasets/karate.py	76	# (we just use the first one).	STRING
MEDIUM	torch_geometric/datasets/medshapenet.py	75	r"""The absolute filepaths that must be present in order to skip	STRING
MEDIUM	torch_geometric/contrib/nn/models/rbcd_attack.py	602	<https://www.cs.cit.tum.de/daml/are-gnn-defenses-robust>`_ for details.	CODE
MEDIUM	torch_geometric/profile/profile.py	49	r"""A decorator to facilitate profiling a function, e.g., obtaining	CODE
MEDIUM	torch_geometric/profile/profile.py	145	r"""A context decorator to facilitate timing a function, e.g., obtaining	STRING
MEDIUM	torch_geometric/loader/node_loader.py	274	# f'best practices for PyG [{link}])')	STRING
MEDIUM	torch_geometric/loader/utils.py	233	# NOTE Here, we utilize `feature_store.multi_get` to give the feature store	STRING
MEDIUM	torch_geometric/loader/utils.py	267	# NOTE Here, we utilize `feature_store.multi_get` to give the feature store	STRING
LOW	torch_geometric/data/collate.py	274	# Other-wise, just return the list of values as it is.	COMMENT
MEDIUM	torch_geometric/data/dataset.py	199	r"""The absolute filepaths that must be present in order to skip	CODE
MEDIUM	torch_geometric/data/dataset.py	211	r"""The absolute filepaths that must be present in order to skip	STRING
MEDIUM	examples/llm/relbench_gretriever.py	223	# Single-graph paradigm: all nodes belong to batch index 0.	COMMENT
LOW	examples/llm/txt2qa.py	510	# If file_path is not relative to input_dir, just use the stem	COMMENT

Modern AI Meta-Vocabulary12 hits · 25 pts

Severity	File	Line	Snippet	Context
MEDIUM	test/llm/test_rag_loader.py	237	# Create RAG loader	COMMENT
MEDIUM	torch_geometric/llm/models/qa_gen.py	3079	# Truncate texts to fit embedding model's max length (512 tokens)	COMMENT
MEDIUM	torch_geometric/llm/models/qa_gen.py	3140	# Truncate answer to fit embedding model's max length (512 tokens)	COMMENT
MEDIUM	torch_geometric/llm/models/qa_gen.py	3218	# Truncate questions to fit embedding model's max length (512 tokens)	COMMENT
MEDIUM	torch_geometric/llm/models/qa_gen.py	243	# Handle embedding model GPU memory	STRING
MEDIUM	examples/rect.py	12	# RECT focuses on the zero-shot, i.e. completely-imbalanced label setting:	COMMENT
MEDIUM	examples/rect.py	14	# RECT (or more specifically its supervised part RECT-L) model in the zero-shot	COMMENT
MEDIUM	examples/llm/txt2kg_rag.py	273	# Read corpus data to create the KG and for document retrieval (RAG).	STRING
MEDIUM	examples/llm/txt2kg_rag.py	381	# creating the embedding model	STRING
MEDIUM	examples/llm/txt2kg_rag.py	453	# creating the embedding model	STRING
MEDIUM	examples/llm/txt2kg_rag.py	521	# VectorDB retrieval just vanilla vector RAG	STRING
MEDIUM	examples/llm/txt2qa_config/text_config_vllm.yaml	31	embedding: 0.3 # GPU memory for embedding model (typically smaller)	CODE

Structural Annotation Overuse18 hits · 24 pts

Severity	File	Line	Snippet	Context
LOW⚡	torch_geometric/llm/models/qa_gen.py	1390	# Step 1: Clean markdown code blocks	COMMENT
LOW⚡	torch_geometric/llm/models/qa_gen.py	1395	# Step 2: Try direct parsing first (fastest path)	COMMENT
LOW	torch_geometric/llm/models/qa_gen.py	1406	# Step 3: Find JSON array/object with bracket counting	COMMENT
LOW	torch_geometric/llm/models/qa_gen.py	1470	# Step 4: Last resort - old regex fallback	COMMENT
LOW⚡	torch_geometric/llm/models/qa_gen.py	2009	# Step 1: Calculate exact (floating-point) values for each category	STRING
LOW⚡	torch_geometric/llm/models/qa_gen.py	2016	# Step 2: Sort by fractional part descending (0.9, 0.9, 0.2)	STRING
LOW⚡	torch_geometric/llm/models/qa_gen.py	2021	# Step 3: Assign floor values first (everyone gets their integer part)	STRING
LOW⚡	torch_geometric/llm/models/qa_gen.py	2027	# Step 4: Distribute remaining items based on fractional parts	STRING
LOW⚡	torch_geometric/llm/models/qa_gen.py	2035	# Step 5: Edge case - if still items left (e.g., rounding errors)	STRING
LOW	examples/distributed/graphlearn_for_pytorch/README.md	22	#### Step 1: Prepare and partition the data	COMMENT
LOW	examples/distributed/graphlearn_for_pytorch/README.md	30	#### Step 2: Run the example in each training node	COMMENT
LOW	examples/distributed/graphlearn_for_pytorch/README.md	55	#### Step 1: Setup a distributed file system	COMMENT
LOW	examples/distributed/graphlearn_for_pytorch/README.md	62	#### Step 2: Prepare and partition the data	COMMENT
LOW	examples/distributed/graphlearn_for_pytorch/README.md	85	#### Step 3: Set up the configure file	COMMENT
LOW	examples/distributed/graphlearn_for_pytorch/README.md	89	#### Step 4: Launch the distributed training	COMMENT
LOW⚡	examples/quiver/single_gpu_quiver.py	21	# Step 1: Using Quiver's sampler	COMMENT
LOW	examples/quiver/single_gpu_quiver.py	99	# Step 2: Using Quiver's Feature	COMMENT
LOW	examples/quiver/single_gpu_quiver.py	117	# Step 3: Training the PyG Model with Quiver	COMMENT

Redundant / Tautological Comments15 hits · 20 pts

Severity	File	Line	Snippet	Context
LOW	test/explain/algorithm/test_pg_explainer.py	222	# Check if the explanation is valid	COMMENT
LOW	test/explain/algorithm/test_pg_explainer.py	371	# Check if the explanation is valid	COMMENT
LOW	test/explain/algorithm/test_pg_explainer.py	467	# Check if the explanation is valid	COMMENT
LOW	test/loader/test_link_neighbor_loader.py	284	# Check if each seed edge has a different batch:	COMMENT
LOW	torch_geometric/llm/utils/backend_utils.py	383	) # Check if pandas is installed	CODE
LOW	torch_geometric/llm/models/qa_gen.py	3847	# Check if segment overlaps with QA time range	COMMENT
LOW	torch_geometric/llm/models/qa_gen.py	4174	# Check if we have context_segments from processing	COMMENT
LOW	…rch_geometric/explain/algorithm/attention_explainer.py	128	# Check if all components appear in the module name in	COMMENT
LOW	…rch_geometric/explain/algorithm/attention_explainer.py	180	# Check if we collected any attention coefficients.	COMMENT
LOW	torch_geometric/datasets/tag_dataset.py	368	# Check if the .pt files already exist	STRING
LOW	torch_geometric/io/fs.py	167	# Open file has protocol stripped.	STRING
LOW	torch_geometric/utils/embedding.py	97	# Check if the outputs is a dictionary mapping node types to embeddings	COMMENT
LOW	torch_geometric/contrib/nn/models/rbcd_attack.py	497	# Check if middle point is root	COMMENT
LOW	examples/llm/txt2qa.py	394	# Check if file exists before removing	COMMENT
LOW	examples/llm/txt2qa.py	531	# Check if output file already exists (resume functionality)	COMMENT

Decorative Section Separators8 hits · 20 pts

Severity	File	Line	Snippet	Context
MEDIUM	test/utils/test_scatter.py	144	# ================	COMMENT
MEDIUM	test/utils/test_scatter.py	153	# ================	COMMENT
MEDIUM	test/utils/test_segment.py	58	# ================	COMMENT
MEDIUM	test/utils/test_segment.py	63	# ================	COMMENT
MEDIUM	torch_geometric/graphgym/config.py	321	# ===================================	STRING
MEDIUM	torch_geometric/graphgym/config.py	329	# ===================================	STRING
MEDIUM	examples/explain/mgnan_graph_mutagenicity.py	246	# --------------------------------------------------------------	COMMENT
MEDIUM	examples/explain/mgnan_graph_mutagenicity.py	248	# --------------------------------------------------------------	COMMENT

Verbosity Indicators12 hits · 15 pts

Severity	File	Line	Snippet	Context
LOW⚡	torch_geometric/llm/models/qa_gen.py	1390	# Step 1: Clean markdown code blocks	COMMENT
LOW⚡	torch_geometric/llm/models/qa_gen.py	1395	# Step 2: Try direct parsing first (fastest path)	COMMENT
LOW	torch_geometric/llm/models/qa_gen.py	1406	# Step 3: Find JSON array/object with bracket counting	COMMENT
LOW	torch_geometric/llm/models/qa_gen.py	1470	# Step 4: Last resort - old regex fallback	COMMENT
LOW⚡	torch_geometric/llm/models/qa_gen.py	2009	# Step 1: Calculate exact (floating-point) values for each category	STRING
LOW⚡	torch_geometric/llm/models/qa_gen.py	2016	# Step 2: Sort by fractional part descending (0.9, 0.9, 0.2)	STRING
LOW⚡	torch_geometric/llm/models/qa_gen.py	2021	# Step 3: Assign floor values first (everyone gets their integer part)	STRING
LOW⚡	torch_geometric/llm/models/qa_gen.py	2027	# Step 4: Distribute remaining items based on fractional parts	STRING
LOW⚡	torch_geometric/llm/models/qa_gen.py	2035	# Step 5: Edge case - if still items left (e.g., rounding errors)	STRING
LOW⚡	examples/quiver/single_gpu_quiver.py	21	# Step 1: Using Quiver's sampler	COMMENT
LOW	examples/quiver/single_gpu_quiver.py	99	# Step 2: Using Quiver's Feature	COMMENT
LOW	examples/quiver/single_gpu_quiver.py	117	# Step 3: Training the PyG Model with Quiver	COMMENT

Fake / Example Data11 hits · 12 pts

Severity	File	Line	Snippet	Context
LOW⚡	test/transforms/test_pad.py	20	def fake_data() -> Data:	CODE
LOW⚡	test/transforms/test_pad.py	426	data = fake_data()	CODE
LOW⚡	test/transforms/test_pad.py	439	data = fake_data()	CODE
LOW⚡	test/transforms/test_pad.py	464	data = fake_data()	CODE
LOW	test/transforms/test_pad.py	327	@pytest.mark.parametrize('data', [fake_data(), fake_hetero_data()])	CODE
LOW	test/transforms/test_pad.py	341	data = fake_data()	CODE
LOW	test/transforms/test_pad.py	367	data = fake_data()	CODE
LOW	test/transforms/test_pad.py	404	@pytest.mark.parametrize('data', [fake_data(), fake_hetero_data()])	CODE
LOW	test/transforms/test_pad.py	479	data = fake_data()	CODE
LOW	torch_geometric/nn/fx.py	141	if node.op == 'placeholder':	STRING
LOW	torch_geometric/nn/fx.py	178	if node.op not in ['placeholder', 'output']:	STRING

Docstring Block Structure2 hits · 10 pts

Severity	File	Line	Snippet	Context
HIGH	torch_geometric/nn/models/gpse.py	1009	Calculates the average cosine similarity between predicted and true features on a batch of graphs. Args:	STRING
HIGH	torch_geometric/contrib/utils/_relbench.py	11	Converts a :class:`relbench.base.Database` object into a :class:`~torch_geometric.data.HeteroData` object. Each	STRING

AI Response Leakage1 hit · 8 pts

Severity	File	Line	Snippet	Context
HIGH	examples/datapipe.py	1	# In this example, you will find data loading implementations using PyTorch	COMMENT

Over-Commented Block6 hits · 6 pts

Severity	File	Line	Snippet	Context
LOW	test/utils/test_scatter.py	141		COMMENT
LOW	test/sampler/test_sampler_neighbor_sampler.py	521	expected_output = SamplerOutput(	COMMENT
LOW	torch_geometric/nn/conv/meshcnn_conv.py	401	# Option 2	COMMENT
LOW	torch_geometric/graphgym/config.py	301	# Auto match computational budget, match upper bound / lower bound	COMMENT
LOW	torch_geometric/graphgym/config.py	321	# ===================================	COMMENT
LOW	examples/tgn.py	1	# This code achieves a performance of around 96.60%. However, it is not	COMMENT

Cross-Language Confusion1 hit · 5 pts

Severity	File	Line	Snippet	Context
HIGH	torch_geometric/utils/geodesic.py	29	To install, run :obj:`pip install cython && pip install gdist`.	CODE

Overly Generic Function Names2 hits · 2 pts

Severity	File	Line	Snippet	Context
LOW	torch_geometric/distributed/event_loop.py	92	def run_task(self, coro):	STRING
LOW	examples/llm/txt2qa.py	516	async def process_data(self, data: str, data_id: str,	CODE

Slop Phrases1 hit · 0 pts

Severity	File	Line	Snippet	Context
LOW	torch_geometric/io/fs.py	145	# If the source path points to a directory, we need to make sure to	STRING

Analysis Overview

What These Metrics Mean

Score History

Severity Breakdown

Directory Score Breakdown

Pattern Findings