Functions — github.com/danielpatrickhug/GitModel

hub / github.com/danielpatrickhug/GitModel / functions

Functions80 in github.com/danielpatrickhug/GitModel

⨍ Functions 80 ◇ Types & classes 13 ↳ Endpoints 1

↓ 4 callersFunctiongaussian_kernel_torch

Computes the Gaussian kernel matrix between two sets of embeddings using PyTorch. :param embs_a: Tensor of shape (batch_size_a, embedding_dim

src/ml_models/graph_networks/kernels.py:8

↓ 3 callersMethodcompose_inference

(self, text_block, messages)

src/ml_models/transformers/semantic_graph_context_generator.py:49

↓ 3 callersFunctioncos_sim

Computes the cosine similarity cos_sim(a[i], b[j]) for all i and j. :return: Matrix with res[i][j] = cos_sim(a[i], b[j])

src/ml_models/graph_networks/kernels.py:40

↓ 3 callersMethodprocess_transcript

(self, segments, file_name, git_repo_path, output_file_path, system_prompt, task, code_type)

src/ml_models/transformers/semantic_graph_context_generator.py:58

↓ 3 callersMethodrun

Run the pipeline.

src/pipeline.py:66

↓ 2 callersFunctioncompose_inference

(text_block, messages)

src/ml_models/transformers/category_decomposition.py:27

↓ 2 callersFunctioncompute_kernel_by_type

(embs, threshold=0.65, kernel_type="cosine", sigma=1.0)

src/ml_models/graph_networks/kernels.py:121

↓ 2 callersFunctioncreate_prompt_message_template

(text, role="user")

src/ml_models/transformers/category_decomposition.py:21

↓ 2 callersMethodcreate_prompt_message_template

(self, text, role="user")

src/ml_models/transformers/semantic_graph_context_generator.py:44

↓ 2 callersMethodfrom_yaml

(cls, yaml_path)

src/config.py:12

↓ 2 callersFunctionk_hop_message_passing_sparse

Compute the k-hop adjacency matrix and aggregated features using message passing. Parameters: A (numpy array or scipy sparse matrix): Th

src/ml_models/graph_networks/message_passing.py:32

↓ 1 callersMethod__init__

(self, in_features, out_features, dropout=0.0)

src/ml_models/graph_networks/gnn_head.py:176

↓ 1 callersFunctionchat_gpt_inference

(messages: list)

src/ml_models/transformers/category_decomposition.py:11

↓ 1 callersMethodchat_gpt_inference

(self, messages: list)

src/ml_models/transformers/semantic_graph_context_generator.py:35

↓ 1 callersFunctionclone_and_create_context_folder

(repo_name: str, id_name: str)

src/fetch_repo.py:6

↓ 1 callersFunctioncombine_graphs_with_gat

Combine a list of graphs using the Graph Attention Network. Parameters: graphs (List of Tuple): List of tuples where each tuple contains

src/ml_models/graph_networks/gnn_head.py:195

↓ 1 callersFunctioncommunity_detection

(A)

src/community_detection/extract_connected_components.py:5

↓ 1 callersMethodcompute_hierarchical_topic_tree

(self, topic_model, data, key="query")

src/ml_models/topic_modeling/topic_model.py:66

↓ 1 callersMethodcompute_kernel_matrix

Compute the kernel matrix using the specified kernel function. Parameters: node_features (numpy array): The feature matrix o

src/ml_models/graph_networks/gnn_head.py:24

↓ 1 callersMethoddecompose_repo

(self, contents, git_repo_path, name_id, topic_tree, out_path)

src/pipeline.py:118

↓ 1 callersFunctiondegree_matrix

compute degree matrix using adjacency distance matrix from pairwise distances :A: nxn size matrix embedding minmaxed using mu sigma and pairw

src/ml_models/graph_networks/kernels.py:152

↓ 1 callersFunctiondfs

Depth-first search algorithm.

src/community_detection/extract_connected_components.py:46

↓ 1 callersMethodembed_data

Embed the sentences/text using the MiniLM language model (which uses mean pooling)

src/ml_models/graph_networks/gnn_head.py:49

↓ 1 callersMethodencode

(self)

src/ml_models/graph_networks/gnn_head.py:134

↓ 1 callersFunctionextract_questions

(prompt_output)

src/system_prompts/format_system_prompts.py:20

↓ 1 callersMethodfind_files_with_substring

(self, root_dir, substring)

src/pipeline.py:44

↓ 1 callersFunctionformat_system_prompts

(git_repo_path, file_name)

src/system_prompts/format_system_prompts.py:4

↓ 1 callersFunctionformat_system_prompts_with_tree

(git_repo_path, file_name, topic_tree)

src/system_prompts/format_system_prompts.py:29

↓ 1 callersMethodgenerate_graph

Generate a graph using the kernel function and message passing. Parameters: node_features (numpy array): The feature matrix

src/ml_models/graph_networks/gnn_head.py:93

↓ 1 callersFunctionget_methods

(class_or_str)

src/ast_parsers/python_ast_parser.py:35

↓ 1 callersMethodget_repo_contents

(self, git_repo_path)

src/ml_models/transformers/semantic_graph_context_generator.py:21

↓ 1 callersMethodget_representative_docs

(self, topic_model)

src/ml_models/topic_modeling/topic_model.py:43

↓ 1 callersMethodget_topic_info

(self, topic_model)

src/ml_models/topic_modeling/topic_model.py:36

↓ 1 callersFunctiongraph_laplacian

compute graph laplacian using degree and adjacency matrix from pairwise distances :A: nxn size matrix embedding minmaxed using mu sigma and p

src/ml_models/graph_networks/kernels.py:165

↓ 1 callersFunctioninstantiate_class_from_config

(config: Config)

src/dynamic_import.py:10

↓ 1 callersFunctionk_hop_message_passing

Compute the k-hop adjacency matrix and aggregated features using message passing. Parameters: A (numpy array): The adjacency matrix of t

src/ml_models/graph_networks/message_passing.py:6

↓ 1 callersFunctionlinearly_sum_gnn_heads

Linearly sum a list of adjacency matrices and normalize the result. Also computes the average of the node embeddings and checks that the numb

src/ml_models/graph_networks/gnn_head.py:225

↓ 1 callersMethodload_jsonl

(self, filepaths)

src/pipeline.py:50

↓ 1 callersFunctionparse_github_repo

(local_dir)

src/ast_parsers/python_ast_parser.py:47

↓ 1 callersFunctionparse_python_file

(file_path)

src/ast_parsers/python_ast_parser.py:7

↓ 1 callersMethodprepare_attention_input

(self, Wh)

src/ml_models/graph_networks/gnn_head.py:163

↓ 1 callersFunctionreportWebVitals

(onPerfEntry)

git-frontend/src/reportWebVitals.js:1

↓ 1 callersFunctionrun_migrations_offline

Run migrations in 'offline' mode. This configures the context with just a URL and not an Engine, though an Engine is acceptable here as w

src/alembic/env.py:29

↓ 1 callersFunctionrun_migrations_online

Run migrations in 'online' mode. In this scenario we need to create an Engine and associate a connection with the context.

src/alembic/env.py:53

Method__init__

(self, pipeline_settings, semantic_graph_context_generator, gnn_heads, topic_model)

src/config.py:5

Method__init__

(self, config: Config)

src/config.py:19

Method__init__

(self, config: Config)

src/pipeline.py:24

Method__init__

(self, config)

src/ml_models/topic_modeling/topic_model.py:26

Method__init__

(self, config: Config)

src/ml_models/transformers/semantic_graph_context_generator.py:15

Method__init__

(self, config: Config)

src/ml_models/graph_networks/gnn_head.py:18

Method__init__

(self, config: Config)

src/ml_models/graph_networks/gnn_head.py:123

Method__init__

(self, in_features, out_features, dropout=0.0)

src/ml_models/graph_networks/gnn_head.py:143

src/ml_models/topic_modeling/topic_model.py:29

Method__repr__

(self)

src/ml_models/transformers/semantic_graph_context_generator.py:18

Method__repr__

(self)

src/ml_models/graph_networks/gnn_head.py:21

Method__repr__

(self)

src/ml_models/graph_networks/gnn_head.py:126

Functionanalyze_repo

(input_data: AnalyzeRepoInput)

main.py:29

Functioncategory_generation

Decompose a category into its constituent parts. Args: category (str): The category to decompose. Returns: list: The list of

src/ml_models/transformers/category_decomposition.py:37

Functioncompute_kernel

(embs)

src/ml_models/graph_networks/kernels.py:145

Functioncompute_kernel_with_activation_fn

(embs, activation_fn=F.relu, kernel_type="cosine", sigma=1.0)

src/ml_models/graph_networks/kernels.py:134

Functionconnected_components

Find connected components in a graph represented by an adjacency matrix.

src/community_detection/extract_connected_components.py:55

Functioncos_sim_torch

Computes the cosine similarity cos_sim(a[i], b[j]) for all i and j. Using torch.nn.functional.cosine_similarity :return: Matrix with res[

src/ml_models/graph_networks/kernels.py:62

Methoddecompose_repo

(self, git_repo_path, name_id, out_path, skip_graph_generation=False)

src/ml_models/transformers/semantic_graph_context_generator.py:107

Functionestimate_pdf

estimate scores probability density function :scores: list of distance scores from topic features to topic centroid :return: distribution

src/ml_models/graph_networks/kernels.py:31

Methodfit_topic_model

(self, topic_model, data, embeddings, key="query")

src/ml_models/topic_modeling/topic_model.py:32

Methodforward

(self, h, adj)

src/ml_models/graph_networks/gnn_head.py:153

Methodforward

(self, embs, adj_matrices)

src/ml_models/graph_networks/gnn_head.py:188

Methodfrom_config

(cls, config: Config)

Methodget_repo_contents

(self, git_repo_path)

src/pipeline.py:114

FunctionhandleOpenaiSecretChange

(event)

git-frontend/src/App.js:25

FunctionhandleRepoUrlChange

(event)

git-frontend/src/App.js:21

FunctionhandleSubmit

()

git-frontend/src/App.js:29

Methodmean_pooling

(self, model_output, attention_mask)

src/ml_models/graph_networks/gnn_head.py:129

Functionprint_communities

(data, A)

src/community_detection/extract_connected_components.py:21

Functionprune_ref_docs

Drops unnecessary documents from the reference embeddings and updates the list of reference documents, and then recomputes the adjacency matr

src/ml_models/graph_networks/kernels.py:83

Methodreduce_outliers

(self, topic_model, data, topics, probs, key="query", strategy="c-tf-idf")

src/ml_models/topic_modeling/topic_model.py:46

Methodreduce_topics

(self, topic_model, data, nr_topics, key="query")

src/ml_models/topic_modeling/topic_model.py:39

Methodrun

Run the topic modeling.

src/ml_models/topic_modeling/topic_model.py:71