wip: DAG encuentra24_empleos_daily

2026-07-29 06:49:44 +02:00
parent a6ac875820
commit 0f63449b0a
1 changed files with 71 additions and 0 deletions
@@ -0,0 +1,71 @@
+from airflow import DAG
+from airflow.operators.python import PythonOperator
+from datetime import datetime, timedelta
+import paramiko
+
+# -- Configuracion -------------------------------------------------------------
+SSH_HOST = '152.53.242.31'
+SSH_USER = 'root'
+SSH_PASS = 'XpvvdHnNr0Bq9qU'
+VENV     = '/home/jpuma/proyectos/bot-sunat/.venv/bin/python'
+SCRIPT   = '/opt/scrapers/encuentra24_empleos/scraper.py'
+
+default_args = {
+    'owner':             'jpuma',
+    'retries':           2,
+    'retry_delay':       timedelta(minutes=15),
+    'email':             ['jpuma@redneurocom.com'],
+    'email_on_failure':  True,
+    'email_on_retry':    False,
+    'execution_timeout': timedelta(hours=6),
+}
+
+
+def run_scraper(**ctx):
+    client = paramiko.SSHClient()
+    client.set_missing_host_key_policy(paramiko.AutoAddPolicy())
+    client.connect(
+        hostname=SSH_HOST,
+        port=22,
+        username=SSH_USER,
+        password=SSH_PASS,
+        timeout=30,
+    )
+    client.get_transport().set_keepalive(30)
+
+    cmd = f'{VENV} {SCRIPT}'
+    _, stdout, stderr = client.exec_command(cmd, timeout=21600)  # 6 h
+
+    output = stdout.read().decode('utf-8', errors='replace')
+    errors = stderr.read().decode('utf-8', errors='replace')
+    client.close()
+
+    print("=== STDOUT ===")
+    print(output)
+    if errors.strip():
+        print("=== STDERR ===")
+        print(errors)
+
+    if 'Finalizado:' not in output:
+        raise RuntimeError(
+            f"El scraper no completo correctamente. "
+            f"'Finalizado:' ausente en stdout. "
+            f"STDERR: {errors[:500]}"
+        )
+
+
+with DAG(
+    dag_id='encuentra24_empleos_daily',
+    default_args=default_args,
+    description='Scraper diario - Encuentra24 Empleos Centroamerica (6 paises)',
+    schedule_interval='30 5 * * *',
+    start_date=datetime(2024, 1, 1),
+    catchup=False,
+    max_active_runs=1,
+    tags=['scraping', 'encuentra24', 'ca', 'empleos'],
+) as dag:
+
+    ejecutar = PythonOperator(
+        task_id='ejecutar_scraper_encuentra24',
+        python_callable=run_scraper,
+    )