Files
ASE/load_orchestrator.py
2025-06-03 19:37:11 +02:00

176 lines
6.0 KiB
Python
Executable File

#!.venv/bin/python
# Import necessary libraries
import logging
import importlib
import asyncio
import os
import aiomysql
import contextvars
# Import custom modules for configuration and database connection
from utils.config import loader_load_data as setting
from utils.database import CSV_RECEIVED
# Crea una context variable per identificare il worker
worker_context = contextvars.ContextVar('worker_id', default='00')
# Formatter personalizzato che include il worker_id
class WorkerFormatter(logging.Formatter):
def format(self, record):
record.worker_id = worker_context.get()
return super().format(record)
# Initialize the logger for this module
logger = logging.getLogger()
# Delay tra un processamento CSV e il successivo (in secondi)
CSV_PROCESSING_DELAY = 0.2
# Tempo di attesa se non ci sono record da elaborare
NO_RECORD_SLEEP = 60
async def get_next_csv_atomic(pool, table_name):
"""Preleva atomicamente il prossimo CSV da elaborare"""
async with pool.acquire() as conn:
async with conn.cursor() as cur:
# Usa SELECT FOR UPDATE per lock atomico
await cur.execute(f"""
SELECT id, unit_type, tool_type, unit_name, tool_name
FROM {table_name}
WHERE locked = 0 AND status = %s
ORDER BY id
LIMIT 1
FOR UPDATE SKIP LOCKED
""", (CSV_RECEIVED,))
result = await cur.fetchone()
if result:
await cur.execute(f"""
UPDATE {table_name}
SET locked = 1
WHERE id = %s
""", (result[0],))
await conn.commit()
return result
async def worker(worker_id: int, cfg: object, pool) -> None:
# Imposta il context per questo worker
worker_context.set(f"W{worker_id}")
debug_mode = (logging.getLogger().getEffectiveLevel() == logging.DEBUG)
logger.info("Avviato")
while True:
try:
logger.info("Inizio elaborazione")
record = await get_next_csv_atomic(pool, cfg.dbrectable)
if record:
success = await load_csv(record, cfg, pool)
if not success:
logger.error("Errore durante l'elaborazione")
await asyncio.sleep(CSV_PROCESSING_DELAY)
else:
await asyncio.sleep(NO_RECORD_SLEEP)
except Exception as e:
logger.error(f"Errore durante l'esecuzione: {e}", exc_info=debug_mode)
await asyncio.sleep(1)
async def load_csv(record: tuple, cfg: object, pool) -> bool:
debug_mode = (logging.getLogger().getEffectiveLevel() == logging.DEBUG)
logger.debug("Inizio ricerca nuovo CSV da elaborare")
id, unit_type, tool_type, unit_name, tool_name = record
logger.info(f"Trovato CSV da elaborare: ID={id}, Tipo={unit_type}_{tool_type}, Nome={unit_name}_{tool_name}")
# Costruisce il nome del modulo da caricare dinamicamente
module_names = [f'utils.parsers.by_name.{unit_name.lower()}_{tool_name.lower()}',
f'utils.parsers.by_name.{unit_name.lower()}_{tool_type.lower()}',
f'utils.parsers.by_name.{unit_name.lower()}_all',
f'utils.parsers.by_type.{unit_type.lower()}_{tool_type.lower()}']
modulo = None
for module_name in module_names:
try:
logger.debug(f"Caricamento dinamico del modulo: {module_name}")
modulo = importlib.import_module(module_name)
logger.debug(f"Funzione 'main_loader' caricata dal modulo {module_name}")
break
except (ImportError, AttributeError) as e:
logger.debug(f"Modulo {module_name} non presente o non valido. {e}", exc_info=debug_mode)
if not modulo:
logger.error(f"Nessun modulo trovato {module_names}")
return False
# Ottiene la funzione 'main_loader' dal modulo
funzione = getattr(modulo, "main_loader")
# Esegui la funzione
logger.info(f"Elaborazione con modulo {modulo} per ID={id}")
await funzione(cfg, id, pool)
logger.info(f"Elaborazione completata per ID={id}")
return True
async def main():
"""Main function: avvia i worker e gestisce il ciclo principale."""
logger.info("Avvio del sistema...")
cfg = setting.Config()
logger.info("Configurazione caricata correttamente")
try:
# Configura il logging globale
log_level = os.getenv("LOG_LEVEL", "INFO").upper()
debug_mode = (logging.getLogger().getEffectiveLevel() == logging.DEBUG)
# Configura il logging con il formatter personalizzato
handler = logging.FileHandler(cfg.logfilename)
formatter = WorkerFormatter(
"%(asctime)s - PID: %(process)d.Worker-%(worker_id)s.%(name)s.%(funcName)s.%(levelname)s: %(message)s"
)
handler.setFormatter(formatter)
# Rimuovi eventuali handler esistenti e aggiungi il nostro
logger.handlers.clear()
logger.addHandler(handler)
logger.setLevel(getattr(logging, log_level))
logger.info("Logging configurato correttamente")
# Numero massimo di worker concorrenti
logger.info(f"Avvio di {cfg.max_threads} worker concorrenti")
pool = await aiomysql.create_pool(
host=cfg.dbhost,
user=cfg.dbuser,
password=cfg.dbpass,
db=cfg.dbname,
minsize=1,
maxsize=cfg.max_threads*4
)
# Avvia i worker
workers = [
asyncio.create_task(worker(i, cfg, pool))
for i in range(cfg.max_threads)
]
logger.info("Sistema avviato correttamente. In attesa di nuovi task...")
try:
await asyncio.gather(*workers, return_exceptions=debug_mode)
finally:
pool.close()
await pool.wait_closed()
except KeyboardInterrupt:
logger.info("Info: Shutdown richiesto... chiusura in corso")
except Exception as e:
logger.error(f"Errore principale: {e}", exc_info=debug_mode)
if __name__ == "__main__":
asyncio.run(main())