56 lines
2.3 KiB
Python
56 lines
2.3 KiB
Python
# services/invoice_processor_service.py
|
|
import logging
|
|
from typing import Dict, List, Any
|
|
|
|
# Importamos nuestro nuevo cliente GCP de forma local y limpia
|
|
from .gcp_document_ai_client import process_document_gcp
|
|
|
|
# (Opcional, si tienes utilidades) from .utils import data_cleaner
|
|
|
|
# Importamos la configuración centralizada
|
|
from core.config import settings
|
|
|
|
# --- Lógica de negocio extraída del antiguo processing.py ---
|
|
def _get_confidence_threshold_for_field(field_type: str) -> float:
|
|
return settings.CONFIDENCE_THRESHOLDS.get(field_type, settings.CONFIDENCE_THRESHOLDS["__default__"])
|
|
|
|
def _extract_specific_fields(entities: List[Any]) -> Dict[str, str]:
|
|
extracted_data = {field: "Not found or low confidence" for field in settings.REQUIRED_FIELDS}
|
|
|
|
for entity in entities:
|
|
entity_type = entity.type_
|
|
threshold = _get_confidence_threshold_for_field(entity_type)
|
|
|
|
if entity_type in settings.REQUIRED_FIELDS and entity.confidence >= threshold:
|
|
value = entity.mention_text.replace('\n', ' ').strip()
|
|
# if entity_type == 'invoice_date':
|
|
# value = data_cleaner.normalize_date(value) or f"Unparseable Date: '{value}'"
|
|
extracted_data[entity_type] = value
|
|
return extracted_data
|
|
|
|
# --- Función principal del servicio ---
|
|
def process_invoice_from_bytes(file_bytes: bytes, mime_type: str) -> Dict[str, str]:
|
|
"""
|
|
Orquesta el proceso completo: llama a Document AI, extrae y limpia los datos.
|
|
"""
|
|
try:
|
|
# 1. Llamar a la API de Google a través de nuestro cliente dedicado
|
|
document = process_document_gcp(
|
|
project_id=settings.GCP_PROJECT_ID,
|
|
location=settings.GCP_LOCATION,
|
|
processor_id=settings.DOCAI_PROCESSOR_ID,
|
|
file_bytes=file_bytes,
|
|
mime_type=mime_type,
|
|
)
|
|
|
|
# 2. Aplicar la lógica de negocio para extraer y validar los campos
|
|
validated_data = _extract_specific_fields(document.entities)
|
|
|
|
logging.info("Documento procesado con éxito y datos validados.")
|
|
return validated_data
|
|
|
|
except Exception as e:
|
|
logging.error(f"Error en el flujo de procesamiento de factura: {e}", exc_info=True)
|
|
# Re-lanzamos la excepción para que el endpoint de la API la capture y devuelva un 500
|
|
raise
|