feat: Integrate CSV cleaning functionality into the application, adding a new GUI tab for user interaction, enabling CSV loading, analysis, and rule application for data cleaning. Implement backend tracking for batch processing and enhance the BatchProcessor with tracking capabilities.

2026-04-22 00:04:15 -05:00
parent 7a38b0843b
commit 6b3eccdc19
30 changed files with 3590 additions and 19 deletions
@@ -0,0 +1,107 @@
+import requests
+import logging
+import os
+import socket
+import getpass
+from typing import Optional, Dict, Any
+from datetime import datetime
+
+logger = logging.getLogger(__name__)
+
+class TrackingManager:
+    """Gestiona la comunicación con el servidor de seguimiento web."""
+
+    def __init__(self, base_url: str = "http://localhost:8000"):
+        self.base_url = base_url
+        self.current_batch_id: Optional[str] = None
+        self.is_enabled = True
+        self._session = requests.Session() # Usar sesión para reutilizar conexiones
+
+    def start_batch(self, root_path: str, total_radicados: int, config: Optional[Dict] = None,
+                    csv_output_path: Optional[str] = None):
+        """Notifica el inicio de un nuevo lote al servidor."""
+        if not self.is_enabled:
+            return
+
+        try:
+            payload = {
+                "root_path": root_path,
+                "total_radicados": total_radicados,
+                "csv_output_path": csv_output_path,
+                "hostname": socket.gethostname(),
+                "username": getpass.getuser(),
+                "config": config or {}
+            }
+            response = self._session.post(f"{self.base_url}/batches", json=payload, timeout=5)
+            if response.status_code == 200:
+                self.current_batch_id = response.json().get("id")
+                logger.info(f"Tracking iniciado: Batch ID {self.current_batch_id}")
+            else:
+                logger.warning(f"Error al iniciar tracking: {response.status_code}")
+        except Exception as e:
+            logger.error(f"Error de conexión con el servidor de tracking: {e}")
+
+    def update_progress(self, current_index: int, success_count: int, error_count: int, 
+                        total_tokens: int, total_cost_usd: float, status: str = "processing"):
+        """Actualiza el progreso del lote actual."""
+        if not self.is_enabled or not self.current_batch_id:
+            return
+
+        try:
+            payload = {
+                "current_radicado_index": current_index,
+                "success_count": success_count,
+                "error_count": error_count,
+                "total_tokens": total_tokens,
+                "total_cost_usd": total_cost_usd,
+                "status": status
+            }
+            self._session.patch(f"{self.base_url}/batches/{self.current_batch_id}", json=payload, timeout=1)
+        except Exception as e:
+            logger.debug(f"Error al actualizar tracking: {e}")
+
+    def log(self, message: str, level: str = "info", radicado: str = None):
+        """Envía una entrada de log al servidor de tracking."""
+        if not self.is_enabled or not self.current_batch_id:
+            return
+        try:
+            payload = {"message": message, "level": level, "radicado": radicado}
+            self._session.post(
+                f"{self.base_url}/batches/{self.current_batch_id}/logs",
+                json=payload, timeout=1
+            )
+        except Exception as e:
+            logger.debug(f"Error al enviar log a tracking: {e}")
+
+    def report_phase(self, phase: str, radicado: str = None):
+        """Reporta la fase actual de procesamiento (extract, normalize, csv, completed)."""
+        if not self.is_enabled or not self.current_batch_id:
+            return
+        try:
+            payload = {"current_phase": phase, "current_radicado": radicado}
+            self._session.patch(f"{self.base_url}/batches/{self.current_batch_id}", json=payload, timeout=1)
+        except Exception as e:
+            logger.debug(f"Error al reportar fase a tracking: {e}")
+
+    def report_error(self, error_message: str):
+        """Reporta un error crítico al servidor."""
+        if not self.is_enabled or not self.current_batch_id:
+            return
+
+        try:
+            payload = {"last_error": error_message}
+            self._session.patch(f"{self.base_url}/batches/{self.current_batch_id}", json=payload, timeout=1)
+        except Exception as e:
+            logger.debug(f"Error al reportar error de tracking: {e}")
+
+    def complete_batch(self, success_count: int, error_count: int, total_tokens: int, total_cost_usd: float):
+        """Notifica la finalización exitosa del lote."""
+        self.update_progress(
+            current_index=-1, # Indica finalización si se desea
+            success_count=success_count,
+            error_count=error_count,
+            total_tokens=total_tokens,
+            total_cost_usd=total_cost_usd,
+            status="completed"
+        )
+        self.current_batch_id = None