Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/NJU-LINK/DR3-Eval
/ functions
Functions
786 in github.com/NJU-LINK/DR3-Eval
⨍
Functions
786
◇
Types & classes
73
↓ 2 callers
Method
get_statistics_summary
Generate and return the statistics summary string.
src/core/orchestrator.py:1515
↓ 2 callers
Method
get_stats
Get statistics about loaded documents and chunks.
libs/miroflow-tools/src/miroflow_tools/tools/rag_tool.py:1054
↓ 2 callers
Method
get_summary
Get a summary of folder contents.
src/io/folder_processor.py:155
↓ 2 callers
Method
get_trace_id
Get the trace ID for this error.
src/llm/exceptions.py:117
↓ 2 callers
Method
handle_max_turns_reached_summary_prompt
Handle max turns reached summary prompt
src/llm/providers/openai_client.py:522
↓ 2 callers
Function
i
(t)
docs/static/js/bulma-slider.min.js:1
↓ 2 callers
Function
isIFrame
(unknown)
docs/static/js/bulma-carousel.js:258
↓ 2 callers
Function
isNode
(unknown)
docs/static/js/bulma-carousel.js:249
↓ 2 callers
Function
is_float
(element: any)
benchmarks/evaluators/eval_utils.py:270
↓ 2 callers
Method
load_long_context
(self, path: Path)
evaluators/utils/document_loader.py:163
↓ 2 callers
Method
load_source_folder
(self, source_folder: Path)
evaluators/utils/document_loader.py:176
↓ 2 callers
Function
normalize_number_str
(number_str: str)
benchmarks/evaluators/eval_utils.py:226
↓ 2 callers
Function
o
(t)
docs/static/js/bulma-carousel.min.js:1
↓ 2 callers
Function
parse_llm_response_for_tool_calls
Parse tool_calls or <use_mcp_tool> tags from LLM response text. Returns a list containing tool call information.
src/utils/parsing_utils.py:89
↓ 2 callers
Function
parse_rag_from_execution_log
Parse RAG retrieval information directly from the execution log. This extracts RAG tool calls and their results from the main execution
src/runners/batch_tasks.py:473
↓ 2 callers
Method
post_process_tool_call_result
Process tool call results
src/core/orchestrator.py:261
↓ 2 callers
Function
preview_folder
Preview folder contents without running a task. Args: folder_path: Path to the folder to preview recursive: Whether to s
src/runners/folder_task.py:504
↓ 2 callers
Function
process_folder_for_task
Process all files in a folder and prepare task description for LLM. This function: 1. Scans the folder and categorizes all files
src/io/folder_processor.py:429
↓ 2 callers
Function
ql
(c)
docs/static/js/fontawesome.all.min.js:5
↓ 2 callers
Function
remove_ellipsis
(s: str)
evaluators/utils/document_loader.py:69
↓ 2 callers
Function
resolve_query_from_jsonl
Resolve query from a jsonl file if the query is a file path. If query ends with .jsonl, it reads the file and finds the matching query
src/runners/folder_task.py:62
↓ 2 callers
Function
run_batch_evaluation
Run batch evaluation across multiple cases.
eval.py:151
↓ 2 callers
Function
run_folder_task
Run a task on a folder containing multiple files. Args: cfg: Hydra configuration object folder_path: Path to the folder
src/runners/folder_task.py:377
↓ 2 callers
Function
run_folder_task_simple
Simplified interface to run a folder task with default config. Args: folder_path: Path to the folder to process query: T
src/runners/folder_task.py:462
↓ 2 callers
Method
run_sub_agent
Run sub agent Note: APIConnectionSkipError, APIRateLimitError, and APITimeoutError are propagated to the caller for proper ha
src/core/orchestrator.py:415
↓ 2 callers
Function
safe_json_loads
Safely parse a JSON string with multiple fallbacks: 1. Try standard json.loads() 2. If it fails, try json_repair 3. If it still fails
src/utils/parsing_utils.py:24
↓ 2 callers
Function
save_report_comparison
Save both original and final reports to a comparison file. Args: results_dir: Directory to save results folder_name: Nam
src/runners/folder_task.py:309
↓ 2 callers
Function
save_report_to_md
Save the final report to a markdown file with timestamp. Args: results_dir: Directory to save results folder_name: Name
src/runners/folder_task.py:267
↓ 2 callers
Function
sc
(l,h)
docs/static/js/fontawesome.all.min.js:5
↓ 2 callers
Function
scan_folder
Scan a folder and categorize all files. Args: folder_path: Path to the folder to scan recursive: Whether to scan subdire
src/io/folder_processor.py:194
↓ 2 callers
Method
set_task_log
Set the task logger for structured logging.
libs/miroflow-tools/src/miroflow_tools/manager.py:73
↓ 2 callers
Function
split_string
( s: str, char_list: list[str] = [",", ";"], )
benchmarks/evaluators/eval_utils.py:237
↓ 2 callers
Function
tl
(c)
docs/static/js/fontawesome.all.min.js:5
↓ 2 callers
Method
to_json
(self)
evaluators/utils/run_all.py:59
↓ 2 callers
Function
tokenize
(s: str)
evaluators/utils/document_loader.py:87
↓ 2 callers
Function
ul
(c)
docs/static/js/fontawesome.all.min.js:5
↓ 2 callers
Method
update_case
(self, case_id: str = None, increment: int = 0)
eval.py:108
↓ 2 callers
Function
verify_answer_gaia
Use GAIA-style judge to verify if the predicted answer is correct.
benchmarks/evaluators/eval_utils.py:221
↓ 2 callers
Function
vl
()
docs/static/js/fontawesome.all.min.js:5
↓ 2 callers
Function
zl
(h,c)
docs/static/js/fontawesome.all.min.js:5
↓ 1 callers
Function
ClientFactory
Automatically select provider and create LLM client based on configuration
src/llm/factory.py:24
↓ 1 callers
Function
DocxConverter
(local_path: str)
src/io/input_handler.py:437
↓ 1 callers
Function
HtmlConverter
(local_path: str)
src/io/input_handler.py:430
↓ 1 callers
Function
Ml
(c)
docs/static/js/fontawesome.all.min.js:5
↓ 1 callers
Function
Nc
(c)
docs/static/js/fontawesome.all.min.js:5
↓ 1 callers
Function
PptxConverter
Converts PPTX files to Markdown. Supports headings, tables, images with alt text, and recursively extracts text from GROUP shapes. Args:
src/io/input_handler.py:687
↓ 1 callers
Function
XlsxConverter
Converts Excel files to plain Markdown tables using openpyxl. Does NOT preserve color/style formatting to keep output concise. Args:
src/io/input_handler.py:596
↓ 1 callers
Function
ZipConverter
Extracts ZIP files to a permanent local directory and returns a listing of extracted files.
src/io/input_handler.py:957
↓ 1 callers
Method
__init__
(self, config: Optional[EvalConfig] = None)
evaluators/utils/base.py:75
↓ 1 callers
Method
__init__
Initialize the benchmark run Args: cfg: Hydra configuration object
benchmarks/common_benchmark.py:691
↓ 1 callers
Method
__init__
(self, target_path: str, task_per_run: int, data_path: str)
benchmarks/check_progress/common.py:308
↓ 1 callers
Method
_apply_cache_control
Apply cache control to the last user message and system message (if applicable)
src/llm/providers/anthropic_client.py:396
↓ 1 callers
Method
_build_batch_summary_prompt
Build the batch summary prompt.
libs/miroflow-tools/src/miroflow_tools/tools/rag_rerank.py:334
↓ 1 callers
Method
_build_index
(self)
evaluators/utils/document_loader.py:168
↓ 1 callers
Method
_build_rerank_prompt
Build the rerank prompt.
libs/miroflow-tools/src/miroflow_tools/tools/rag_rerank.py:164
↓ 1 callers
Method
_calculate_total_score
(self, results: Dict[str, EvalResult])
evaluators/utils/run_all.py:272
↓ 1 callers
Method
_call_anthropic_vllm
Call Anthropic Claude Vision API.
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_vqa.py:293
↓ 1 callers
Method
_call_custom_video
Call custom video API (OpenAI-compatible)
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_video.py:622
↓ 1 callers
Method
_call_frames_video
Analyze video by extracting key frames and using VQA model. This approach: 1. Extracts N evenly-spaced frames from t
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_video.py:428
↓ 1 callers
Method
_call_gemini
Call Google Gemini API for video understanding
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_video.py:246
↓ 1 callers
Method
_call_gemini_audio
Call Gemini Audio API via OpenAI-compatible endpoint (supports audio_url format)
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_audio.py:387
↓ 1 callers
Method
_call_gpt4o_audio
Call GPT-4o-audio API (best for complex audio understanding)
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_audio.py:488
↓ 1 callers
Method
_call_gpt4o_video
Call GPT-4o-video API
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_video.py:365
↓ 1 callers
Method
_call_openai_compatible_vllm
Call OpenAI-compatible endpoint (Qwen, custom, etc.).
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_vqa.py:349
↓ 1 callers
Method
_call_openai_vllm
Call OpenAI-compatible VLLM (GPT-4V, GPT-4o, etc.).
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_vqa.py:235
↓ 1 callers
Method
_call_openai_whisper
Call OpenAI Whisper API
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_audio.py:225
↓ 1 callers
Method
_call_qwen_audio
Call Qwen-Audio API (supports both transcription and audio QA)
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_audio.py:339
↓ 1 callers
Method
_call_qwen_video
Call Qwen-VL-Video API
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_video.py:385
↓ 1 callers
Method
_call_whisper_os
Call Open-Source Whisper API
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_audio.py:293
↓ 1 callers
Method
_check_cache_valid
Check if cache is valid for the current file.
libs/miroflow-tools/src/miroflow_tools/tools/rag_tool.py:218
↓ 1 callers
Method
_check_citations_with_extraction
Check which required titles are cited in the report.
evaluators/citation_coverage.py:375
↓ 1 callers
Method
_collect_stream_response
Collect streaming response chunks into a complete response object (async version)
src/llm/providers/openai_client.py:562
↓ 1 callers
Method
_collect_stream_response_sync
Collect streaming response chunks into a complete response object (sync version)
src/llm/providers/openai_client.py:644
↓ 1 callers
Method
_compress_item
Compress a single item. Args: idx: Index of item to compress compress_func: Async function to compre
src/core/progressive_memory.py:243
↓ 1 callers
Function
_defineProperty
(obj, key, value)
docs/static/js/bulma-carousel.js:475
↓ 1 callers
Method
_display_summary
Display summary statistics
benchmarks/check_progress/common.py:558
↓ 1 callers
Method
_display_summary
Display summary statistics
benchmarks/check_progress/common.py:910
↓ 1 callers
Method
_encode_image
Encode image to base64 or return URL.
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_vqa.py:160
↓ 1 callers
Method
_evaluate_checklist
(self, result_text: str, items: List[Dict])
evaluators/format_compliance.py:152
↓ 1 callers
Method
_extract_claims_single_call
(self, report_text: str)
evaluators/factual_accuracy.py:1168
↓ 1 callers
Method
_extract_explicit_citations
Extract all explicit citations from the report text. Supported formats: - [long_context: "..."] - [filename.md, page
evaluators/citation_coverage.py:248
↓ 1 callers
Function
_extract_file_content
Extract text content from a file using existing converters. Args: file_info: FileInfo object for the file max_content_le
src/io/folder_processor.py:252
↓ 1 callers
Method
_extract_run_number
Extract run number from directory path for sorting
benchmarks/check_progress/common.py:357
↓ 1 callers
Method
_extract_video_frames
Extract evenly-spaced frames from video and convert to base64. Args: video_path: Path to video file
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_video.py:555
↓ 1 callers
Method
_fallback_compress
Fallback compression using simple truncation
src/core/progressive_memory.py:404
↓ 1 callers
Method
_filter_message_history
Filter message history, keep specified number of tool results
src/llm/base_client.py:323
↓ 1 callers
Function
_findKey
(json: Any, key: str)
src/io/input_handler.py:494
↓ 1 callers
Method
_find_failed_claims
Find claims that failed verification in existing evaluation data.
evaluators/factual_accuracy.py:625
↓ 1 callers
Method
_format_citation_coverage_details
(self, result: EvalResult)
evaluators/utils/run_all.py:645
↓ 1 callers
Method
_format_factual_accuracy_details
(self, result: EvalResult)
evaluators/utils/run_all.py:561
↓ 1 callers
Method
_format_factuality_v2_details
(self, result: EvalResult)
evaluators/utils/run_all.py:319
↓ 1 callers
Method
_format_format_compliance_details
(self, result: EvalResult)
evaluators/utils/run_all.py:608
↓ 1 callers
Method
_format_information_recall_details
(self, result: EvalResult)
evaluators/utils/run_all.py:484
↓ 1 callers
Method
_format_overall_quality_details
(self, result: EvalResult)
evaluators/utils/run_all.py:677
↓ 1 callers
Method
_generate_follow_up_questions
Generate follow-up questions for multi-turn verification.
libs/miroflow-tools/src/miroflow_tools/tools/enhanced_vqa.py:477
↓ 1 callers
Method
_get_api_key
Get API key with priority: 1. Environment variable based on provider (e.g., OPENAI_API_KEY, ANTHROPIC_API_KEY) 2. Config file
src/llm/base_client.py:91
↓ 1 callers
Function
_get_audio_extension
Determine the appropriate audio file extension from URL or content type. Args: url: The URL of the audio file content_type:
libs/miroflow-tools/src/miroflow_tools/mcp_servers/audio_mcp_server_os.py:37
↓ 1 callers
Function
_get_audio_info
Get audio information string.
src/io/folder_processor.py:396
↓ 1 callers
Method
_get_base_url
Get base URL with priority: 1. Environment variable based on provider (e.g., OPENAI_BASE_URL, ANTHROPIC_BASE_URL) 2. Config f
src/llm/base_client.py:120
↓ 1 callers
Method
_get_db_path
Get SQLite database path for caching embeddings.
libs/miroflow-tools/src/miroflow_tools/tools/rag_tool.py:156
← previous
next →
201–300 of 786, ranked by callers