Functions — github.com/AIcling/agentic

hub / github.com/AIcling/agentic_geo / functions

Functions107 in github.com/AIcling/agentic_geo

⨍ Functions 107◇ Types & classes 5

↓ 14 callersFunctionformat_number_7decimals

Format number to 7 decimal places.

src/run_geo.py:24

↓ 14 callersFunctionformat_scores_7decimals

Format scores (can be nested lists/arrays) to 7 decimal places.

src/run_geo.py:33

↓ 10 callersFunctioncall_gpt

Synchronous wrapper that calls async or sync version based on config. Note: Caching has been removed to prevent cache corruption from concurr

src/geo_functions.py:131

↓ 8 callersFunctionload_config

Load configuration from an INI file. Uses Python's standard configparser module. Args: config_file: Path to the configuratio

src/config_loader.py:15

↓ 7 callersFunctionget_local_llm_model

Get local LLM model name from config.

src/config_loader.py:131

↓ 7 callersFunctionimpression_subjective_impression

(sentences, query, n = 5, normalize = True, idx = 0, metric = 'subjective_impression')

src/utils.py:149

↓ 7 callersFunctionuse_local_llm

Check if local LLM should be used.

src/config_loader.py:121

↓ 6 callersFunctionextract_citations_new

(text)

src/utils.py:36

↓ 4 callersFunctionget_answer

Get answer for a query using cached sources and real-time LLM generation. Args: query: The query string summaries: Optional l

src/utils.py:301

↓ 4 callersFunctionget_effective_api_config

Get the effective API configuration based on USE_LOCAL_LLM setting.

src/config_loader.py:141

↓ 4 callersFunctionsearch_handler

(req, source_count = 8)

src/search_try.py:172

↓ 3 callersFunctionget_model_name

Get the current LLM model name for file naming.

src/run_geo.py:181

↓ 3 callersFunctionget_num_words

(line)

src/utils.py:33

↓ 3 callersFunctionuse_async

Check if async mode should be used.

src/config_loader.py:136

↓ 2 callersFunction_safe_write_cache

Safely write cache JSON to disk using atomic write (temp file + rename). If writing fails, does not interrupt main process. Note: This no

src/utils.py:255

↓ 2 callersFunctioncall_gpt_async

Async version of call_gpt using AsyncOpenAI client. Note: Caching has been removed to prevent cache corruption from concurrent writes. Al

src/geo_functions.py:62

↓ 2 callersFunctionclean_source_gpt35_async

Async version of clean_source_gpt35 using AsyncOpenAI client.

src/search_try.py:31

↓ 2 callersFunctiongenerate_answer_async

Async version of generate_answer using AsyncOpenAI client.

src/generative_le.py:32

↓ 2 callersFunctionget_summary

Extract summary from LLM response, removing code blocks if present.

src/geo_functions.py:33

↓ 2 callersFunctionload_results

Load JSON result file.

src/merge_and_analyze_results.py:9

↓ 2 callersFunctionprocess_single_query_async

Async process a single query with error handling and immediate result saving.

src/run_geo.py:1081

↓ 2 callersMethodrank_strategies

Rank all strategies for given query and context. Args: query: User query context: Content to optimize str

src/surrogate_model_evolved.py:393

↓ 2 callersFunctionreturnable_score_from_scores

(scores)

src/utils.py:152

↓ 2 callersFunctionsanitize_string

Sanitize string to prevent JSON encoding issues. Args: s: String to sanitize Returns: Sanitized string safe for JSON encoding

src/run_geo.py:367

↓ 2 callersMethodselect

(self, indices)

src/run_geo.py:1429

↓ 2 callersMethodshuffle

(self, seed=None)

src/run_geo.py:1425

↓ 1 callersFunction_extract_rewritten_content

Best-effort extractor to return ONLY rewritten content. Some evolved strategies include an output format like: - "First output your plan, then the

src/run_geo.py:46

↓ 1 callersMethod_init_weights

(self, init_bias: float)

src/surrogate_model_evolved.py:101

↓ 1 callersMethod_load_backbone_checkpoint

Load pretrained backbone weights from checkpoint file. Args: ckpt_path: Path to pytorch_model.bin or similar checkpoint file

src/surrogate_model_evolved.py:272

↓ 1 callersFunctionapply_evolved_strategy

Apply an evolved strategy to rewrite content using LLM. This function implements the Oracle Rewrite logic for evolved strategies. The strategy_pro

src/run_geo.py:98

↓ 1 callersFunctionbuild_critic_prompt

Build Critic input prompt (aligned with training).

src/surrogate_model_evolved.py:51

↓ 1 callersFunctioncalculate_method_statistics

Calculate statistics for each method.

src/merge_and_analyze_results.py:30

↓ 1 callersFunctionclean_source_gpt35

Synchronous wrapper that calls async or sync version based on config.

src/search_try.py:86

↓ 1 callersFunctionclean_source_text

(text: str)

src/search_try.py:154

↓ 1 callersFunctionconvert_to_number

(x, min_val = 1.0)

src/utils.py:198

↓ 1 callersFunctionecn

(sentence)

src/utils.py:37

↓ 1 callersFunctiongenerate_answer

Synchronous wrapper that calls async or sync version based on config.

src/generative_le.py:81

↓ 1 callersFunctionget_api_key

Get OpenAI API key from config.

src/config_loader.py:108

↓ 1 callersFunctionimpression_pos_count_simple

(sentences, n = 5, normalize=True)

src/utils.py:76

↓ 1 callersFunctionimpression_word_count_simple

(sentences, n = 5, normalize=True)

src/utils.py:64

↓ 1 callersFunctionimpression_wordpos_count_simple

(sentences, n = 5, normalize=True)

src/utils.py:51

↓ 1 callersFunctioninit_results_file

Initialize results file with model name.

src/run_geo.py:191

↓ 1 callersFunctionmain

()

src/merge_and_analyze_results.py:184

↓ 1 callersFunctionmain

()

src/merge_results.py:135

↓ 1 callersFunctionmain

()

scripts/archive_to_strategies.py:23

↓ 1 callersFunctionmain

()

scripts/prep_evolved_for_hf.py:19

↓ 1 callersFunctionmain

()

base_model/download_base_model.py:20

↓ 1 callersFunctionmain_async

Main async entry point.

src/run_geo.py:1234

↓ 1 callersFunctionmerge_results

Merge two result files.

src/merge_and_analyze_results.py:15

↓ 1 callersFunctionmerge_results

Merge all split result files.

src/merge_results.py:18

↓ 1 callersFunctionpreload_geobench_to_cache

Preload GEO-Bench sources into cache. Args: dataset_file: Path to GEO-Bench jsonl cache_file: Target cache file path max_

src/preload_cache_from_geobench.py:8

↓ 1 callersFunctionpreload_msdata_to_cache

Preload MSdata sources into cache. Args: dataset_file: Path to MSdata JSON file (JSON array format) cache_file: Target cache file

src/preload_cache_from_msdata.py:8

↓ 1 callersFunctionprint_statistics

Print statistics.

src/merge_and_analyze_results.py:116

↓ 1 callersFunctionprocess_with_semaphore

(task)

src/run_geo.py:1214

↓ 1 callersFunctionrun_concurrent_async

Concurrent execution using asyncio.

src/run_geo.py:1204

↓ 1 callersFunctionrun_sequential_async

Sequential execution using async (original method, but async).

src/run_geo.py:1194

↓ 1 callersFunctionsanitize_for_json

Convert object to JSON-serializable form. Args: obj: Object to sanitize Returns: JSON-serializable version of the object

src/run_geo.py:419

↓ 1 callersFunctionsave_merged_results

Save merged results.

src/merge_and_analyze_results.py:165

↓ 1 callersFunctionsave_statistics

Save statistics to JSON.

src/merge_and_analyze_results.py:177

↓ 1 callersMethodscore_strategy

Score a single strategy using SHORT_PROMPT. Args: query: User query context: Content to optimize strategy

src/surrogate_model_evolved.py:332

↓ 1 callersMethodselect_best_strategy

Select the best strategy based on Critic prediction. Critic SHORT_PROMPT strategy FULL_PROMPT Args: query: User query

src/surrogate_model_evolved.py:483

↓ 1 callersFunctionsummarize_text_identity

(source, query)

src/search_try.py:168

↓ 1 callersFunctiontruncate_content

Truncate content, Head+Tail mode supported. IMPORTANT: This must match the truncation used during training to ensure training-inference consi

src/surrogate_model_evolved.py:22

↓ 1 callersFunctionvalidate_json_serializable

Validate that an object can be JSON serialized. Args: obj: Object to validate path: Current path in object tree (for error reporting) Return

( self, hidden_size: int = 1536, intermediate_size: int = 1024, dropout: float

src/surrogate_model_evolved.py:84

Method__init__

Initialize the evolved Critic model. Args: base_model_path: Path to Qwen2.5-1.5B-Instruct base model (for structure)

src/surrogate_model_evolved.py:134

Method__init__

Initialize the evolved strategy selector. Args: base_model_path: Path to Qwen2.5-1.5B-Instruct base model (for structure)

src/surrogate_model_evolved.py:429

Functionauthoritative_optimization_mine

(summary)

src/geo_functions.py:234

Functioncheck_summaries_exist

(sources, summaries)

src/utils.py:294

Functionciting_credible_sources_mine

(summary)

src/geo_functions.py:270

Functionfluent_optimization_gpt

(summary)

src/geo_functions.py:220

Methodforward

Args: hidden_states: (batch_size, hidden_size) Returns: scores: (batch_size,)

src/surrogate_model_evolved.py:107

Functionget_api_base

Get OpenAI API base URL from config.

src/config_loader.py:116

Functionget_local_llm_base

Get local LLM base URL from config.

src/config_loader.py:126

Functionget_prompt

(source, query)

src/utils.py:19

Methodget_strategy_by_id

Get strategy by genotype_id.

src/surrogate_model_evolved.py:535

Functionget_strategy_prompts

Get strategy prompts (for compatibility with old interface). Note: This is kept for API compatibility, but evolved strategies are loaded from

src/surrogate_model_evolved.py:547

Functionidentity

(summary, source=None)

src/run_geo.py:139

Functionimpression_diversity_detailed

(sentences, query, n = 5, normalize = True, idx = 0)

src/utils.py:131

Functionimpression_follow_detailed

(sentences, query, n = 5, normalize = True, idx = 0)

src/utils.py:137

Functionimpression_influence_detailed

(sentences, query, n = 5, normalize = True, idx = 0)

src/utils.py:140

Functionimpression_para_based

(sentences, n = 5, normalize = True, alpha = 1.1, beta = 1.5, gamma2 = 1/math.e)

src/utils.py:89

Functionimpression_relevance_detailed

(sentences, query, n = 5, normalize = True, idx = 0)

src/utils.py:143

Functionimpression_subjcount_detailed

(sentences, query, n = 5, normalize = True, idx = 0)

src/utils.py:146

Functionimpression_subjpos_detailed

(sentences, query, n = 5, normalize = True, idx = 0)

src/utils.py:128

Functionimpression_uniqueness_detailed

(sentences, query, n = 5, normalize = True, idx = 0)

src/utils.py:134

Functionimprove

Process a query through GEO optimization pipeline. Note: static_cache now defaults to True to use pre-loaded cache from GEO-Bench. This prevents u

src/run_geo.py:945

Functionimprove_with_model_selection

Process a query using surrogate model for strategy selection. New workflow: 1. Use surrogate model to rank all strategies 2. Select the best stra

src/run_geo.py:628

Functionload_evolved_strategies

Load evolved strategies from JSON file. Args: strategies_path: Path to final_strategies_with_prompts.json Returns: List of s

src/surrogate_model_evolved.py:556

Functionmore_quotes_mine

(summary)

src/geo_functions.py:255

Functionreload_config

Force reload configuration from file and environment variables.

src/config_loader.py:160

Functionsave_query_result

Append a single query result to the results file with robust error handling. Args: file_path: Path to the results JSON file query_result: Dicti

src/run_geo.py:454

next →1–100 of 107, ranked by callers