Functions — github.com/EternityJune25/ComoRAG

hub / github.com/EternityJune25/ComoRAG / functions

Functions257 in github.com/EternityJune25/ComoRAG

⨍ Functions 257◇ Types & classes 38

↓ 64 callersMethodget

Retrieve the embedding if cached.

src/comorag/embedding_model/base.py:238

↓ 17 callersMethodget_all_ids

(self)

src/comorag/embedding_store.py:141

↓ 15 callersMethodencode

(self, texts: List[str])

src/comorag/embedding_model/OpenAI.py:77

↓ 12 callersMethodrender

Render a template with the provided variables. Args: name (str): The name of the template. kwargs: Placehold

src/comorag/prompts/prompt_template_manager.py:102

↓ 11 callersFunctioncompute_mdhash_id

Compute the MD5 hash of the given content string and optionally prepend a prefix. Args: content (str): The input string to be hashed

src/comorag/utils/misc_utils.py:152

↓ 9 callersFunctionget_logger

Get a logger with a specific name and optional file logging. Args: name (str): Logger name, typically the module's `__name__`.

src/comorag/utils/logging_utils.py:12

↓ 8 callersMethodget_embeddings

(self, hash_ids, dtype=np.float32)

src/comorag/embedding_store.py:150

↓ 7 callersMethodget_temp_nodes_by_type

Get all nodes of specified type from temporary memory pool

src/comorag/utils/memory_utils.py:105

↓ 7 callersMethodinsert_strings

(self, texts: List[str])

src/comorag/embedding_store.py:63

↓ 6 callersMethodbatch_encode

(self, texts: List[str], **kwargs)

src/comorag/embedding_model/base.py:208

↓ 6 callersMethodget_level_embedding_store

Get EmbeddingStore instance for specified level Args: level: Level index output_dir: Output director

src/comorag/utils/timeline_utils.py:374

↓ 5 callersFunctionget_logger

(name)

script/eval_qa.py:9

↓ 5 callersMethodget_nodes_by_type

Get all nodes of specified type

src/comorag/utils/memory_utils.py:120

↓ 5 callersMethodget_row

(self, hash_id)

src/comorag/embedding_store.py:130

↓ 5 callersMethodget_text_for_all_rows

(self)

src/comorag/embedding_store.py:144

↓ 5 callersMethodupdate_hashes

Update hash values for all content using node_type as namespace

src/comorag/utils/memory_utils.py:37

↓ 4 callersMethod_count_tokens

Count the number of tokens in text Args: text: Input text Returns: int: Number

src/comorag/utils/timeline_utils.py:61

↓ 4 callersMethod_ensure_manager

(cls)

src/comorag/embedding_model/base.py:230

↓ 4 callersMethod_summarize_window

Summarize text within a window Args: texts: List of texts within the window is_final_summary: Whethe

src/comorag/utils/timeline_utils.py:229

↓ 4 callersMethodadd_to_temp_pool

Add node to temporary memory pool

src/comorag/utils/memory_utils.py:89

↓ 4 callersFunctionget_query_instruction

(linking_method)

src/comorag/prompts/linking.py:1

↓ 4 callersMethodinfer

Perform synchronous inference using the LLM. Args: chat (List[TextChatMessage]): Input chat history for the LLM.

src/comorag/llm/base.py:169

↓ 4 callersFunctionnormalize_answer

(s)

script/eval_qa.py:15

↓ 3 callersMethod_call_llm

(self, messages, max_completion_tokens=500, stop_sequence=None, temperature=0, top_p=1)

src/comorag/utils/agents.py:32

↓ 3 callersMethodadd_member

Add member to cluster

src/comorag/utils/cluster_utils.py:28

↓ 3 callersMethodbatch_upsert

Update existing attributes or add new ones from the given dictionary.

src/comorag/llm/base.py:80

↓ 3 callersMethoddense_passage_retrieval

(self, query: str, need_cluster: bool = False)

src/comorag/ComoRAG.py:950

↓ 3 callersMethodget_all_probes

Get probe content from all nodes and remove duplicates Returns: List[str]: List of deduplicated probe content

src/comorag/utils/memory_utils.py:141

↓ 3 callersMethodget_summary_by_level

Get summaries for specified level Args: level: Level index Returns: List[str]:

src/comorag/utils/timeline_utils.py:350

↓ 3 callersMethodindex

(self, docs: List[str])

src/comorag/ComoRAG.py:198

↓ 3 callersFunctionmin_max_normalize

(x)

src/comorag/utils/misc_utils.py:141

↓ 2 callersMethod__init__

Initialize base Agent class Args: model: Model name, defaults to "gpt-4o-mini" llm_base_url: Base URL for LLM

src/comorag/utils/agents.py:15

↓ 2 callersMethod_get_optimal_clusters

Determine optimal number of clusters

src/comorag/utils/cluster_utils.py:175

↓ 2 callersMethod_reduce_dimensions

Use UMAP for dimension reduction

src/comorag/utils/cluster_utils.py:191

↓ 2 callersMethodbatch_encode

Encode a batch of texts into embeddings. Args: texts: List of texts to encode **kwargs: Additional p

src/comorag/embedding_model/BGEEmbedding.py:131

↓ 2 callersMethodbatch_infer

Perform batched synchronous inference using the LLM. Args: batch_chat (List[List[TextChatMessage]]): Input chat

src/comorag/llm/base.py:183

↓ 2 callersMethodbatch_openie

Conduct batch OpenIE synchronously using multi-threading which includes NER and triple extraction. Args: chunks (Dict[st

src/comorag/information_extraction/openie_openai.py:197

↓ 2 callersMethodbatch_upsert

Update existing attributes or add new ones from the given dictionary.

src/comorag/embedding_model/base.py:76

↓ 2 callersFunctionconvert_text_chat_messages_to_input_ids

(messages: List[TextChatMessage], tokenizer: PreTrainedTokenizer, add_assistant_header=True)

src/comorag/llm/vllm_offline.py:14

↓ 2 callersFunctionfilter_invalid_triples

Filters out invalid and duplicate triples from a list of triples. A valid triple meets the following criteria: 1. It contains exactly th

src/comorag/utils/llm_utils.py:218

↓ 2 callersFunctionfix_broken_generated_json

Fixes a malformed JSON string by: - Removing the last comma and any trailing content. - Iterating over the JSON string once to determine

src/comorag/utils/llm_utils.py:146

↓ 2 callersMethodfrom_dict

Create an LLMConfig instance from a dictionary.

src/comorag/embedding_model/base.py:89

↓ 2 callersFunctionget_batches

(vecs, batch_size)

src/comorag/utils/embed_utils.py:35

↓ 2 callersMethodget_cluster_by_id

Get cluster by specified ID

src/comorag/utils/cluster_utils.py:371

↓ 2 callersMethodget_cluster_membership

Get all clusters and membership scores for the specified chunk Args: hash_id: Chunk's hash_id

src/comorag/utils/cluster_utils.py:359

↓ 2 callersMethodget_content_hashes

Get hash values for all content

src/comorag/utils/memory_utils.py:55

↓ 2 callersFunctionget_gold_answers

(samples)

src/comorag/utils/misc_utils.py:207

↓ 2 callersMethodget_hash_id_to_order

Returns a dictionary mapping each hash value to its sequential position in the original text list. Returns: Dict

src/comorag/embedding_store.py:159

↓ 2 callersMethodget_rows

(self, hash_ids, dtype=np.float32)

src/comorag/embedding_store.py:133

↓ 2 callersMethodget_template

Retrieve a template by name. Args: name (str): The name of the template. Returns: Union[Template, L

src/comorag/prompts/prompt_template_manager.py:153

↓ 2 callersMethodload_all_summaries

Load summary content for all levels Returns: Dict[int, List[str]]: Dictionary with level as key and summary list

src/comorag/utils/timeline_utils.py:433

↓ 2 callersMethodload_existing_openie

(self, chunk_keys: List[str])

src/comorag/ComoRAG.py:714

↓ 2 callersMethodmem_encode

(self, query: str, docs: Dict, memory_pool: MemoryPool, probe: str = None)

src/comorag/ComoRAG.py:557

↓ 2 callersMethodmerge_openie_results

(self, all_openie_info: List[dict], chunks_to_save:

src/comorag/ComoRAG.py:740

↓ 2 callersMethodprocess

(self, query, chunks=None, semantic_summaries=None, timeline_summaries=None, content=None, task='fusion', max_

src/comorag/utils/agents.py:60

↓ 2 callersMethodsave_openie_results

(self, all_openie_info: List[dict])

src/comorag/ComoRAG.py:755

↓ 2 callersMethodsummarize

(self, context, max_tokens=150)

src/comorag/utils/summarization_utils.py:13

↓ 2 callersMethodtri_retrieve

(self, query: str, memory_pool: MemoryPool, ver_top_k: int = None, sem_top_k: int = None, epi_top_k: int = Non

src/comorag/ComoRAG.py:456

↓ 2 callersMethodtry_answer

(self, queries: List[str], num_to_retrieve: int = None)

src/comorag/ComoRAG.py:432

↓ 1 callersMethod_create_summary_prompt

Create summary prompt Args: texts: List of texts to be summarized is_final_summary: Whether this is

src/comorag/utils/timeline_utils.py:192

↓ 1 callersFunction_extract_ner_from_response

(real_response)

src/comorag/information_extraction/openie_openai.py:30

↓ 1 callersMethod_generate_and_save_summaries

Generate new summaries and save them Args: output_dir: Summary storage directory

src/comorag/utils/timeline_utils.py:493

↓ 1 callersFunction_get_embedding_model_class

(embedding_model_name: str = "None")

src/comorag/embedding_model/__init__.py:10

↓ 1 callersFunction_get_llm_class

(config: BaseConfig)

src/comorag/llm/__init__.py:12

↓ 1 callersMethod_init_embedding_config

Extract embedding model-specific parameters to init the EmbeddingConfig. Returns: None

src/comorag/embedding_model/BGEEmbedding.py:63

↓ 1 callersMethod_init_embedding_config

Extract embedding model-specific parameters to init the EmbeddingConfig. Returns: None

src/comorag/embedding_model/OpenAI.py:46

↓ 1 callersMethod_init_llm_config

(self)

src/comorag/llm/openai_gpt.py:153

↓ 1 callersMethod_load_clustering_results

Load existing clustering results

src/comorag/utils/cluster_utils.py:123

↓ 1 callersMethod_load_data

(self)

src/comorag/embedding_store.py:92

↓ 1 callersMethod_load_templates

Load all templates from Python scripts in the templates directory.

src/comorag/prompts/prompt_template_manager.py:46

↓ 1 callersMethod_recursive_clustering

(self, texts, max_iterations=5, current_iteration=0)

src/comorag/ComoRAG.py:1107

↓ 1 callersMethod_save_data

(self)

src/comorag/embedding_store.py:109

↓ 1 callersMethod_upsert

(self, hash_ids, texts, embeddings)

src/comorag/embedding_store.py:122

↓ 1 callersMethodadd_fact_edges

(self, chunk_ids: List[str], chunk_triples: List[Tuple])

src/comorag/ComoRAG.py:627

↓ 1 callersMethodadd_fused_node

Add fused content as new node to memory pool Args: probe: Probe used for fusion fused_content: Fused content

src/comorag/utils/memory_utils.py:278

↓ 1 callersMethodadd_new_edges

(self)

src/comorag/ComoRAG.py:803

↓ 1 callersMethodadd_new_nodes

(self)

src/comorag/ComoRAG.py:776

↓ 1 callersMethodadd_passage_edges

(self, chunk_ids: List[str], chunk_triple_entities: List[List[str]])

src/comorag/ComoRAG.py:650

↓ 1 callersMethodadd_synonymy_edges

(self)

src/comorag/ComoRAG.py:670

↓ 1 callersMethodaugment_graph

(self)

src/comorag/ComoRAG.py:768

↓ 1 callersMethodcalculate_metric_scores

Calculate exact match (EM) scores Args: gold_answers: List of standard answers, each element is a list of answers

script/eval_qa.py:41

↓ 1 callersMethodcalculate_metric_scores

Calculate F1 scores Args: gold_answers: List of standard answers, each element is a list of answers predicte

script/eval_qa.py:76

↓ 1 callersMethodclear_temp_pool

Clear temporary memory pool

src/comorag/utils/memory_utils.py:93

↓ 1 callersMethodcompute_probe_note_embeddings

Compute embedding values for probe + atomic notes for all nodes Args: force_recompute: Whether to force recomputation of

src/comorag/utils/memory_utils.py:149

↓ 1 callersFunctionconvert_format_to_template

Converts a .format() style string to a Template-style string. Args: original_string (str): The original string using .format() place

src/comorag/utils/llm_utils.py:32

↓ 1 callersMethodcreate_fusion_content

Fuse related nodes based on probe Args: probe: Probe used for retrieval top_k_percent: Use top percentage of

src/comorag/utils/memory_utils.py:238

↓ 1 callersFunctiondynamic_retry

Factory function to create a retry decorator with dynamic parameters. Args: experiment_config (BaseConfig): Configuration containing

src/comorag/utils/llm_utils.py:102

↓ 1 callersFunctionextract_answers_from_results

(results: List[Dict])

script/eval_qa.py:178

↓ 1 callersFunctionextract_entity_nodes

(chunk_triples: List[List[Triple]])

src/comorag/utils/misc_utils.py:121

↓ 1 callersFunctionextract_final_answer

(output_text: str)

script/eval_qa.py:155

↓ 1 callersFunctionfind_and_merge_results

Find all json files containing 'results' and merge them, returning the merged result list Args: root_path: Root directory path

script/eval_qa.py:121

↓ 1 callersMethodfind_probes

(self, query: str, context: str = None, previous_probes: str = None, max_completion_tokens: int = 500)

src/comorag/utils/agents.py:151

↓ 1 callersFunctionfind_unclosed

Identifies the unclosed braces and brackets in the JSON string. Args: json_str (str): The JSON string to analyze.

src/comorag/utils/llm_utils.py:162

↓ 1 callersFunctionflatten_facts

(chunk_triples: List[Triple])

src/comorag/utils/misc_utils.py:134

↓ 1 callersMethodfrom_dict

Create an LLMConfig instance from a dictionary.

src/comorag/llm/base.py:93

↓ 1 callersMethodfrom_experiment_config

(cls, global_config: BaseConfig)

src/comorag/llm/openai_gpt.py:116

↓ 1 callersMethodfuse_memory_nodes

(self, query: str, content: str, max_completion_tokens: int = 1000)

src/comorag/utils/agents.py:118

↓ 1 callersMethodfusion

(self, query, vers, sems, epis)

src/comorag/utils/agents.py:88

↓ 1 callersMethodgenerate_timeline_summary

Generate timeline summary, only generate one level of summary Returns: Dict[str, Any]: Dictionary containing sum

src/comorag/utils/timeline_utils.py:287

↓ 1 callersMethodget_all_hashes

Get hash values for all nodes, grouped by node type Returns: Dict[str, List[str]]: Keys are node types, values are lists

src/comorag/utils/memory_utils.py:128

next →1–100 of 257, ranked by callers