# agent_v0/server_v1/worker_stream.py
"""
Worker de Streaming Temps Réel — délègue au StreamProcessor (core pipeline).

Surveille les sessions live, analyse screenshots et crops via ScreenAnalyzer + CLIP,
et met à jour le graphe d'intention en temps réel.

Tous les calculs GPU tournent sur le serveur (RTX 5070).
"""

import logging
import threading
import time
from pathlib import Path
from typing import Set

from .stream_processor import StreamProcessor

logger = logging.getLogger("worker_stream")


class StreamWorker:
    """
    Worker qui surveille les sessions live et délègue au StreamProcessor.

    Deux modes de fonctionnement :
    - Polling (start) : boucle qui surveille le dossier live_sessions
    - Direct (process_*) : appelé directement par l'API pour traitement immédiat
    """

    def __init__(self, live_dir: str = "data/training/live_sessions", processor: StreamProcessor = None):
        self.live_dir = Path(live_dir)
        self.live_dir.mkdir(parents=True, exist_ok=True)
        self.running = False
        self.processed_files: Set[str] = set()

        # StreamProcessor partagé (créé si non fourni)
        self.processor = processor or StreamProcessor(data_dir=str(self.live_dir))

        self._thread: threading.Thread = None

    def start(self, blocking: bool = True):
        """Démarrer le worker en mode polling."""
        self.running = True
        logger.info("StreamWorker démarré — surveillance des sessions live.")

        if blocking:
            self._poll_loop()
        else:
            self._thread = threading.Thread(target=self._poll_loop, daemon=True)
            self._thread.start()

    def stop(self):
        """Arrêter proprement le worker."""
        self.running = False
        if self._thread and self._thread.is_alive():
            self._thread.join(timeout=5)
        logger.info("StreamWorker arrêté.")

    def _poll_loop(self):
        """Boucle de polling pour les sessions live sur disque."""
        while self.running:
            try:
                self._check_live_sessions()
                time.sleep(0.5)
            except KeyboardInterrupt:
                self.stop()
            except Exception as e:
                logger.error(f"Erreur worker loop: {e}")

    def _check_live_sessions(self):
        """Parcourir les sessions en cours pour trouver du travail."""
        if not self.live_dir.exists():
            return
        for session_path in self.live_dir.iterdir():
            if session_path.is_dir():
                self._process_session_incremental(session_path)

    def _process_session_incremental(self, session_path: Path):
        """Analyser les nouveaux éléments d'une session active."""
        session_id = session_path.name
        shots_dir = session_path / "shots"

        # Enregistrer la session si pas encore fait
        self.processor.session_manager.get_or_create(session_id)

        # Traiter les nouveaux screenshots full
        for shot_file in sorted(shots_dir.glob("*.png")) if shots_dir.exists() else []:
            file_key = str(shot_file)
            if file_key in self.processed_files:
                continue

            shot_id = shot_file.stem
            if "_crop" in shot_id:
                result = self.processor.process_crop(session_id, shot_id, str(shot_file))
                logger.debug(f"Crop traité: {shot_id}")
            elif shot_id.startswith("heartbeat_") or shot_id.startswith("focus_") or shot_id.startswith("res_shot_"):
                # Pas d'analyse GPU pour les heartbeats, focus et res_shot
                self.processor.session_manager.add_screenshot(session_id, shot_id, str(shot_file))
            elif shot_id.startswith("shot_") and "_full" in shot_id:
                result = self.processor.process_screenshot(session_id, shot_id, str(shot_file))
                logger.info(
                    f"Screenshot analysé: {shot_id} | "
                    f"{result.get('ui_elements_count', 0)} UI, "
                    f"{result.get('text_detected', 0)} textes"
                )
            else:
                # Autres screenshots non reconnus : stocker sans analyser
                self.processor.session_manager.add_screenshot(session_id, shot_id, str(shot_file))

            self.processed_files.add(file_key)

        # Traiter les événements
        event_file = session_path / "live_events.jsonl"
        if event_file.exists():
            self._ingest_events(session_id, event_file)

    def _ingest_events(self, session_id: str, event_file: Path):
        """Lire et ingérer les événements depuis un fichier JSONL."""
        import json

        event_key = f"{session_id}:events:{event_file.stat().st_size}"
        if event_key in self.processed_files:
            return

        try:
            with open(event_file, "r", encoding="utf-8") as f:
                for line in f:
                    line = line.strip()
                    if not line:
                        continue
                    try:
                        event_data = json.loads(line)
                        self.processor.process_event(session_id, event_data)
                    except json.JSONDecodeError:
                        continue
            self.processed_files.add(event_key)
        except Exception as e:
            logger.error(f"Erreur lecture événements {event_file}: {e}")

    # =========================================================================
    # API directe (appelé par api_stream.py)
    # =========================================================================

    def process_screenshot_direct(self, session_id: str, shot_id: str, file_path: str):
        """Traitement direct d'un screenshot (appelé par l'API)."""
        return self.processor.process_screenshot(session_id, shot_id, file_path)

    def process_crop_direct(self, session_id: str, shot_id: str, file_path: str):
        """Traitement direct d'un crop (appelé par l'API)."""
        return self.processor.process_crop(session_id, shot_id, file_path)

    def process_event_direct(self, session_id: str, event_data: dict):
        """Traitement direct d'un événement (appelé par l'API)."""
        return self.processor.process_event(session_id, event_data)

    def finalize_session(self, session_id: str):
        """Finaliser une session et construire le workflow."""
        return self.processor.finalize_session(session_id)

    @property
    def stats(self):
        return self.processor.stats


if __name__ == "__main__":
    logging.basicConfig(
        level=logging.INFO,
        format="%(asctime)s [STREAM-WORKER] %(message)s",
    )
    worker = StreamWorker()
    worker.start()