ASE/src/elab_orchestrator.py

#!.venv/bin/python
"""
Orchestratore dei worker che lanciano le elaborazioni
"""

# Import necessary libraries
import asyncio
import logging

# Import custom modules for configuration and database connection
from utils.config import loader_matlab_elab as setting
from utils.connect.send_email import send_error_email
from utils.csv.loaders import get_next_csv_atomic
from utils.database import WorkflowFlags
from utils.database.action_query import check_flag_elab, get_tool_info
from utils.database.loader_action import unlock, update_status
from utils.general import read_error_lines_from_logs
from utils.orchestrator_utils import run_orchestrator, shutdown_event, worker_context

# Initialize the logger for this module
logger = logging.getLogger()

# Delay tra un processamento CSV e il successivo (in secondi)
ELAB_PROCESSING_DELAY = 0.2
# Tempo di attesa se non ci sono record da elaborare
NO_RECORD_SLEEP = 60


async def worker(worker_id: int, cfg: object, pool: object) -> None:
    """Esegue il ciclo di lavoro per l'elaborazione dei dati caricati.

    Il worker preleva un record dal database che indica dati pronti per
    l'elaborazione, esegue un comando Matlab associato e attende
    prima di iniziare un nuovo ciclo.

    Supporta graceful shutdown controllando il shutdown_event tra le iterazioni.

    Args:
        worker_id (int): L'ID univoco del worker.
        cfg (object): L'oggetto di configurazione.
        pool (object): Il pool di connessioni al database.
    """
    # Imposta il context per questo worker
    worker_context.set(f"W{worker_id:02d}")

    debug_mode = logging.getLogger().getEffectiveLevel() == logging.DEBUG
    logger.info("Avviato")

    try:
        while not shutdown_event.is_set():
            try:
                logger.info("Inizio elaborazione")
                if not await check_flag_elab(pool):
                    record = await get_next_csv_atomic(pool, cfg.dbrectable, WorkflowFlags.DATA_LOADED, WorkflowFlags.DATA_ELABORATED)
                    if record:
                        rec_id, _, tool_type, unit_name, tool_name = [x.lower().replace(" ", "_") if isinstance(x, str) else x for x in record]
                        if tool_type.lower() != "gd":  # i tool GD non devono essere elaborati ???
                            tool_elab_info = await get_tool_info(WorkflowFlags.DATA_ELABORATED, unit_name.upper(), tool_name.upper(), pool)
                            if tool_elab_info:
                                if tool_elab_info["statustools"].lower() in cfg.elab_status:
                                    logger.info("Elaborazione ID %s per %s %s", rec_id, unit_name, tool_name)
                                    await update_status(cfg, rec_id, WorkflowFlags.START_ELAB, pool)
                                    matlab_cmd = f"timeout {cfg.matlab_timeout} ./run_{tool_elab_info['matcall']}.sh \
                                        {cfg.matlab_runtime} {unit_name.upper()} {tool_name.upper()}"
                                    proc = await asyncio.create_subprocess_shell(
                                        matlab_cmd, cwd=cfg.matlab_func_path, stdout=asyncio.subprocess.PIPE, stderr=asyncio.subprocess.PIPE
                                    )

                                    stdout, stderr = await proc.communicate()

                                    if proc.returncode != 0:
                                        logger.error("Errore durante l'elaborazione")
                                        logger.error(stderr.decode().strip())

                                        if proc.returncode == 124:
                                            error_type = f"Matlab elab excessive duration: killed after {cfg.matlab_timeout} seconds."
                                        else:
                                            error_type = f"Matlab elab failed: {proc.returncode}."

                                        # da verificare i log dove prenderli
                                        # with open(f"{cfg.matlab_error_path}{unit_name}{tool_name}_output_error.txt", "w") as f:
                                        #    f.write(stderr.decode().strip())
                                        # errors = [line for line in stderr.decode().strip() if line.startswith("Error")]
                                        # warnings = [line for line in stderr.decode().strip() if not line.startswith("Error")]

                                        errors, warnings = await read_error_lines_from_logs(
                                            cfg.matlab_error_path, f"_{unit_name}_{tool_name}*_*_output_error.txt"
                                        )
                                        await send_error_email(
                                            unit_name.upper(), tool_name.upper(), tool_elab_info["matcall"], error_type, errors, warnings
                                        )

                                    else:
                                        logger.info(stdout.decode().strip())
                                        await update_status(cfg, rec_id, WorkflowFlags.DATA_ELABORATED, pool)
                                    await unlock(cfg, rec_id, pool)
                                    await asyncio.sleep(ELAB_PROCESSING_DELAY)
                                else:
                                    logger.info(
                                        "ID %s %s - %s %s: MatLab calc by-passed.", rec_id, unit_name, tool_name, tool_elab_info["statustools"]
                                    )
                                    await update_status(cfg, rec_id, WorkflowFlags.DATA_ELABORATED, pool)
                                    await update_status(cfg, rec_id, WorkflowFlags.DUMMY_ELABORATED, pool)
                                    await unlock(cfg, rec_id, pool)
                        else:
                            await update_status(cfg, rec_id, WorkflowFlags.DATA_ELABORATED, pool)
                            await update_status(cfg, rec_id, WorkflowFlags.DUMMY_ELABORATED, pool)
                            await unlock(cfg, rec_id, pool)

                    else:
                        logger.info("Nessun record disponibile")
                        await asyncio.sleep(NO_RECORD_SLEEP)
                else:
                    logger.info("Flag fermo elaborazione attivato")
                    await asyncio.sleep(NO_RECORD_SLEEP)

            except asyncio.CancelledError:
                logger.info("Worker cancellato. Uscita in corso...")
                raise

            except Exception as e:  # pylint: disable=broad-except
                logger.error("Errore durante l'esecuzione: %s", e, exc_info=debug_mode)
                await asyncio.sleep(1)

    except asyncio.CancelledError:
        logger.info("Worker terminato per shutdown graceful")
    finally:
        logger.info("Worker terminato")


async def main():
    """Funzione principale che avvia l'elab_orchestrator."""
    await run_orchestrator(setting.Config, worker)


if __name__ == "__main__":
    asyncio.run(main())