teste / src /utils /logging_decorators.py
torxyton's picture
feat: Implementar sistema de logging avançado com SQLite3
b94122a
Raw History Blame Contribute Delete
14 kB
import functools
import time
import traceback
import inspect
from typing import Any, Callable, Optional
from ..core.database_logger import get_logger, LogLevel, LogCategory
def log_execution(category: LogCategory = LogCategory.SYSTEM,
log_args: bool = False,
log_result: bool = False,
log_performance: bool = True):
"""Decorador para logging automático de execução de funções"""
def decorator(func: Callable) -> Callable:
@functools.wraps(func)
def wrapper(*args, **kwargs):
logger = get_logger()
# Informações da função
module_name = func.__module__
function_name = func.__name__
# Obter número da linha
try:
line_number = inspect.getsourcelines(func)[1]
except:
line_number = 0
# Preparar metadados
metadata = {
'function_signature': str(inspect.signature(func))
}
if log_args:
metadata['args'] = str(args)
metadata['kwargs'] = str(kwargs)
start_time = time.time()
try:
# Log de início da execução
logger.log(
level=LogLevel.DEBUG,
category=category,
message=f"Iniciando execução da função {function_name}",
module=module_name,
function=function_name,
line_number=line_number,
metadata=metadata
)
# Executar função
result = func(*args, **kwargs)
execution_time = time.time() - start_time
# Preparar metadados do resultado
result_metadata = metadata.copy()
if log_result:
result_metadata['result'] = str(result)[:1000] # Limitar tamanho
# Log de sucesso
logger.log(
level=LogLevel.INFO,
category=category,
message=f"Função {function_name} executada com sucesso",
module=module_name,
function=function_name,
line_number=line_number,
metadata=result_metadata,
execution_time=execution_time
)
# Log de performance se habilitado
if log_performance:
logger.log_performance_metric(
metric_name=f"{module_name}.{function_name}_execution_time",
metric_value=execution_time,
unit="seconds",
category=category.value,
metadata={'function': function_name, 'module': module_name}
)
return result
except Exception as e:
execution_time = time.time() - start_time
# Log de erro
logger.log(
level=LogLevel.ERROR,
category=category,
message=f"Erro na execução da função {function_name}: {str(e)}",
module=module_name,
function=function_name,
line_number=line_number,
metadata=metadata,
stack_trace=traceback.format_exc(),
execution_time=execution_time
)
raise
return wrapper
return decorator
def log_api_call(endpoint: str = None):
"""Decorador específico para logging de chamadas de API"""
def decorator(func: Callable) -> Callable:
@functools.wraps(func)
def wrapper(*args, **kwargs):
logger = get_logger()
module_name = func.__module__
function_name = func.__name__
try:
line_number = inspect.getsourcelines(func)[1]
except:
line_number = 0
# Extrair informações da requisição se disponível
request_info = {}
if 'request' in kwargs:
request = kwargs['request']
request_info = {
'method': getattr(request, 'method', 'UNKNOWN'),
'url': getattr(request, 'url', 'UNKNOWN'),
'user_agent': getattr(request, 'headers', {}).get('user-agent', 'UNKNOWN')
}
metadata = {
'endpoint': endpoint or function_name,
'request_info': request_info
}
start_time = time.time()
try:
# Log início da chamada API
logger.log(
level=LogLevel.INFO,
category=LogCategory.API,
message=f"Chamada API iniciada: {endpoint or function_name}",
module=module_name,
function=function_name,
line_number=line_number,
metadata=metadata
)
result = func(*args, **kwargs)
execution_time = time.time() - start_time
# Log sucesso da API
logger.log(
level=LogLevel.INFO,
category=LogCategory.API,
message=f"Chamada API concluída com sucesso: {endpoint or function_name}",
module=module_name,
function=function_name,
line_number=line_number,
metadata=metadata,
execution_time=execution_time
)
# Métrica de performance da API
logger.log_performance_metric(
metric_name=f"api_{endpoint or function_name}_response_time",
metric_value=execution_time,
unit="seconds",
category="API",
metadata={'endpoint': endpoint or function_name}
)
return result
except Exception as e:
execution_time = time.time() - start_time
logger.log(
level=LogLevel.ERROR,
category=LogCategory.API,
message=f"Erro na chamada API {endpoint or function_name}: {str(e)}",
module=module_name,
function=function_name,
line_number=line_number,
metadata=metadata,
stack_trace=traceback.format_exc(),
execution_time=execution_time
)
raise
return wrapper
return decorator
def log_ai_model_usage(model_name: str = None):
"""Decorador para logging de uso de modelos de IA"""
def decorator(func: Callable) -> Callable:
@functools.wraps(func)
def wrapper(*args, **kwargs):
logger = get_logger()
module_name = func.__module__
function_name = func.__name__
try:
line_number = inspect.getsourcelines(func)[1]
except:
line_number = 0
metadata = {
'model_name': model_name or function_name,
'input_size': len(str(args)) + len(str(kwargs))
}
start_time = time.time()
try:
logger.log(
level=LogLevel.INFO,
category=LogCategory.AI_MODEL,
message=f"Iniciando processamento do modelo {model_name or function_name}",
module=module_name,
function=function_name,
line_number=line_number,
metadata=metadata
)
result = func(*args, **kwargs)
execution_time = time.time() - start_time
# Atualizar metadados com informações do resultado
result_metadata = metadata.copy()
result_metadata['output_size'] = len(str(result))
logger.log(
level=LogLevel.INFO,
category=LogCategory.AI_MODEL,
message=f"Modelo {model_name or function_name} processado com sucesso",
module=module_name,
function=function_name,
line_number=line_number,
metadata=result_metadata,
execution_time=execution_time
)
# Métricas específicas de IA
logger.log_performance_metric(
metric_name=f"ai_model_{model_name or function_name}_inference_time",
metric_value=execution_time,
unit="seconds",
category="AI_MODEL",
metadata={
'model': model_name or function_name,
'input_size': metadata['input_size'],
'output_size': result_metadata['output_size']
}
)
return result
except Exception as e:
execution_time = time.time() - start_time
logger.log(
level=LogLevel.ERROR,
category=LogCategory.AI_MODEL,
message=f"Erro no modelo {model_name or function_name}: {str(e)}",
module=module_name,
function=function_name,
line_number=line_number,
metadata=metadata,
stack_trace=traceback.format_exc(),
execution_time=execution_time
)
raise
return wrapper
return decorator
class LoggingContext:
"""Context manager para logging de blocos de código"""
def __init__(self, operation_name: str, category: LogCategory = LogCategory.SYSTEM,
level: LogLevel = LogLevel.INFO, metadata: dict = None):
self.operation_name = operation_name
self.category = category
self.level = level
self.metadata = metadata or {}
self.logger = get_logger()
self.start_time = None
def __enter__(self):
self.start_time = time.time()
# Obter informações do caller
frame = inspect.currentframe().f_back
module_name = frame.f_globals.get('__name__', 'unknown')
function_name = frame.f_code.co_name
line_number = frame.f_lineno
self.logger.log(
level=self.level,
category=self.category,
message=f"Iniciando operação: {self.operation_name}",
module=module_name,
function=function_name,
line_number=line_number,
metadata=self.metadata
)
return self
def __exit__(self, exc_type, exc_val, exc_tb):
execution_time = time.time() - self.start_time
# Obter informações do caller
frame = inspect.currentframe().f_back
module_name = frame.f_globals.get('__name__', 'unknown')
function_name = frame.f_code.co_name
line_number = frame.f_lineno
if exc_type is None:
# Sucesso
self.logger.log(
level=self.level,
category=self.category,
message=f"Operação concluída com sucesso: {self.operation_name}",
module=module_name,
function=function_name,
line_number=line_number,
metadata=self.metadata,
execution_time=execution_time
)
else:
# Erro
self.logger.log(
level=LogLevel.ERROR,
category=self.category,
message=f"Erro na operação {self.operation_name}: {str(exc_val)}",
module=module_name,
function=function_name,
line_number=line_number,
metadata=self.metadata,
stack_trace=traceback.format_exc(),
execution_time=execution_time
)
# Log de performance
self.logger.log_performance_metric(
metric_name=f"operation_{self.operation_name.replace(' ', '_')}_time",
metric_value=execution_time,
unit="seconds",
category=self.category.value,
metadata={'operation': self.operation_name}
)
def quick_log(message: str, level: LogLevel = LogLevel.INFO,
category: LogCategory = LogCategory.SYSTEM, **kwargs):
"""Função utilitária para logging rápido"""
logger = get_logger()
# Obter informações do caller
frame = inspect.currentframe().f_back
module_name = frame.f_globals.get('__name__', 'unknown')
function_name = frame.f_code.co_name
line_number = frame.f_lineno
logger.log(
level=level,
category=category,
message=message,
module=module_name,
function=function_name,
line_number=line_number,
**kwargs
)