jm-vis's picture
Upload GAIA agent implementation files for assessment
c922f8b
Raw History Blame Contribute Delete
5.33 kB
"""
Utilities module for GAIA implementation.
Contains helper functions for question parsing, answer formatting, logging, and CLI utilities.
"""
# Use the standard logging module
import logging as python_logging
from typing import Dict, Any, List, Optional, Union
# Import from the specialized directories
from src.gaia.utils.parsing.parsing import (
extract_key_information,
identify_required_tools,
determine_question_complexity,
parse_question,
COMPLEXITY_LEVELS
)
from src.gaia.utils.formatting.formatting import (
extract_answer,
format_answer,
validate_answer_format,
process_answer,
FORMAT_TYPES
)
from src.gaia.utils.logging.logging_framework import (
initialize_logging,
log_info,
log_warning,
log_error,
log_api_request,
log_api_response,
log_tool_selection,
log_tool_execution,
log_workflow_step,
log_memory_operation,
TimingContext,
get_trace_id,
set_trace_id,
generate_trace_id
)
# Create logger using the standard logging module
logger = python_logging.getLogger("gaia_agent.utils")
def clean_text(text: str) -> str:
"""
Clean text by removing extra whitespace and normalizing.
Args:
text: The text to clean
Returns:
Cleaned text
"""
if not text:
return ""
cleaned = " ".join(text.split())
return cleaned.strip()
def truncate_text(text: str, max_length: int = 1000, add_ellipsis: bool = True) -> str:
"""
Truncate text to a maximum length.
Args:
text: The text to truncate
max_length: Maximum length in characters
add_ellipsis: Whether to add ellipsis if truncated
Returns:
Truncated text
"""
if not text or len(text) <= max_length:
return text
truncated = text[:max_length]
if add_ellipsis:
truncated += "..."
return truncated
def merge_dicts(dict1: Dict[str, Any], dict2: Dict[str, Any]) -> Dict[str, Any]:
"""
Merge two dictionaries, with dict2 values taking precedence.
Args:
dict1: First dictionary
dict2: Second dictionary (takes precedence)
Returns:
Merged dictionary
"""
result = dict1.copy()
result.update(dict2)
return result
def get_answer_format(question: str) -> str:
"""
Determine the expected answer format based on the question.
Args:
question: The question to analyze
Returns:
Expected answer format type
"""
question_lower = question.lower()
if any(term in question_lower for term in ["how many", "how much", "count", "sum", "total", "calculate"]):
return FORMAT_TYPES["NUMBER"]
if any(term in question_lower for term in ["when", "what date", "what time", "what year"]):
return FORMAT_TYPES["DATE"]
if any(term in question_lower for term in ["is it", "are there", "does it", "can it", "will it", "should it"]):
if question_lower.startswith(("is ", "are ", "does ", "do ", "can ", "will ", "should ")):
return FORMAT_TYPES["BOOLEAN"]
if any(term in question_lower for term in ["list", "enumerate", "what are", "examples of", "types of"]):
return FORMAT_TYPES["LIST"]
if any(term in question_lower for term in ["who", "person", "which company", "which organization"]):
return FORMAT_TYPES["ENTITY"]
return FORMAT_TYPES["TEXT"]
def analyze_question(question: str) -> Dict[str, Any]:
"""
Comprehensive analysis of a question.
Args:
question: The question to analyze
Returns:
Dictionary with analysis results
"""
parsed = parse_question(question)
expected_format = get_answer_format(question)
return {
"question": question,
"parsed_info": parsed,
"expected_format": expected_format,
"required_tools": parsed["required_tools"],
"complexity": parsed["complexity"]["complexity_level"]
}
def format_response(response: str, expected_format: Optional[str] = None) -> Dict[str, Any]:
"""
Format a response according to the expected format.
Args:
response: The response to format
expected_format: Optional expected format type
Returns:
Dictionary with formatted response
"""
return process_answer(response, expected_format)
# Export additional utility functions and classes
__all__ = [
# Parsing utilities
'extract_key_information',
'identify_required_tools',
'determine_question_complexity',
'parse_question',
'COMPLEXITY_LEVELS',
# Formatting utilities
'extract_answer',
'format_answer',
'validate_answer_format',
'process_answer',
'FORMAT_TYPES',
# Logging utilities
'initialize_logging',
'log_info',
'log_warning',
'log_error',
'log_api_request',
'log_api_response',
'log_tool_selection',
'log_tool_execution',
'log_workflow_step',
'log_memory_operation',
'TimingContext',
'get_trace_id',
'set_trace_id',
'generate_trace_id',
# General utilities
'clean_text',
'truncate_text',
'merge_dicts',
'get_answer_format',
'analyze_question',
'format_response'
]