BSN-Chatsystem/queue_handler.py

#!/usr/bin/env python3
"""
Queue-Handler — SQLite-basierte Submission-Queue mit Idempotenz.
===============================================================

Architektur:
  Webhook → Signatur-Prüfung → Queue (sofort 200 OK) → Worker → GPU-Verarbeitung

Eigenschaften:
  - Persistente Queue in SQLite (kein Redis nötig)
  - Idempotenz via message_id UNIQUE (WhatsApp message ID / Telegram update_id)
  - Webhook-Signaturprüfung (WhatsApp X-Hub-Signature-256, Telegram optional)
  - Status-Tracking (pending → processing → completed/failed)

Nutzung:
  from queue_handler import QueueHandler
  qh = QueueHandler(db)
  qh.enqueue(channel="whatsapp", message_id="wamid.xxx", payload=body)
"""

import hashlib
import hmac
import json
import logging
import sqlite3
import time
from datetime import datetime, timezone
from typing import Optional

logger = logging.getLogger(__name__)


# ── Webhook-Signaturprüfung ──────────────────────────────────────────

def verify_whatsapp_signature(
    body: bytes,
    signature_header: str,
    app_secret: str,
) -> bool:
    """
    Prüft die WhatsApp X-Hub-Signature-256.

    Args:
        body: Der rohe Request-Body (bytes).
        signature_header: Der Wert des X-Hub-Signature-256 Headers.
        app_secret: Der Meta App Secret (aus Environment).

    Returns:
        True wenn die Signatur gültig ist, sonst False.
    """
    if not signature_header or not app_secret:
        logger.warning("WhatsApp-Signaturprüfung übersprungen — kein Secret/Header")
        return True  # Im Test-Modus erlauben

    # Format: "sha256=<hex-hash>"
    if not signature_header.startswith("sha256="):
        logger.warning("Unbekanntes Signatur-Format: %s", signature_header[:20])
        return False

    expected = signature_header[7:]  # Nach "sha256="
    actual = hmac.new(
        key=app_secret.encode("utf-8"),
        msg=body,
        digestmod=hashlib.sha256,
    ).hexdigest()

    # Constant-time comparison (gegen Timing-Angriffe)
    if not hmac.compare_digest(actual, expected):
        logger.warning("WhatsApp-Signatur UNGÜLTIG — echo/replay erkannt")
        return False

    return True


# ── Queue-Implementierung ────────────────────────────────────────────

class QueueHandler:
    """
    SQLite-basierte Submission-Queue mit Idempotenz.

    Verwendet die Tabelle `submission_queue` in der bestehenden SQLite-DB.
    """

    def __init__(self, db: sqlite3.Connection):
        self.db = db
        self._ensure_table()

    def _ensure_table(self) -> None:
        """Erstellt die Queue-Tabelle, falls nicht vorhanden."""
        self.db.execute("""
            CREATE TABLE IF NOT EXISTS submission_queue (
                id INTEGER PRIMARY KEY AUTOINCREMENT,
                message_id TEXT NOT NULL UNIQUE,
                channel TEXT NOT NULL CHECK(channel IN ('whatsapp', 'telegram')),
                payload TEXT NOT NULL,
                raw_body TEXT,
                status TEXT NOT NULL DEFAULT 'pending'
                    CHECK(status IN ('pending', 'processing', 'completed', 'failed')),
                attempt_count INTEGER NOT NULL DEFAULT 0,
                max_attempts INTEGER NOT NULL DEFAULT 3,
                last_error TEXT,
                created_at TEXT NOT NULL DEFAULT (datetime('now')),
                started_at TEXT,
                completed_at TEXT
            )
        """)
        self.db.execute("""
            CREATE INDEX IF NOT EXISTS idx_queue_status
            ON submission_queue(status, created_at)
        """)
        self.db.commit()

    def enqueue(
        self,
        channel: str,
        message_id: str,
        payload: dict,
        raw_body: Optional[bytes] = None,
    ) -> tuple[bool, str]:
        """
        Reiht eine Submission in die Queue ein.

        Args:
            channel: 'whatsapp' oder 'telegram'
            message_id: Eindeutige Nachrichten-ID (WhatsApp: wamid.xxx, Telegram: update_id)
            payload: Das geparste JSON der Nachricht.
            raw_body: Optionaler roher Body (für Debugging).

        Returns:
            (success, message) — True = eingereiht, False = Duplikat/Fehler.
        """
        payload_json = json.dumps(payload, ensure_ascii=False)
        raw_bytes = raw_body.decode("utf-8", errors="replace") if raw_body else None

        try:
            self.db.execute(
                """
                INSERT INTO submission_queue (message_id, channel, payload, raw_body)
                VALUES (?, ?, ?, ?)
                """,
                (message_id, channel, payload_json, raw_bytes),
            )
            self.db.commit()
            logger.info(
                "Submission in Queue: channel=%s message_id=%s",
                channel,
                message_id,
            )
            return True, "enqueued"

        except sqlite3.IntegrityError:
            # UNIQUE constraint auf message_id → Duplikat erkannt
            logger.info(
                "Duplikat erkannt und verworfen: channel=%s message_id=%s",
                channel,
                message_id,
            )
            # Prüfen, ob die Nachricht bereits verarbeitet wurde
            row = self.db.execute(
                "SELECT status FROM submission_queue WHERE message_id = ?",
                (message_id,),
            ).fetchone()
            if row and row[0] == "completed":
                return False, "already_processed"
            return False, "duplicate_pending"

        except Exception as e:
            logger.exception("Fehler beim Enqueue: %s", e)
            return False, f"error: {e}"

    def dequeue(self, limit: int = 5) -> list[dict]:
        """
        Holt pending Submissions aus der Queue und markiert sie als 'processing'.

        Returns:
            Liste von Submission-Dicts mit id, message_id, channel, payload.
        """
        self.db.execute("BEGIN IMMEDIATE")
        try:
            rows = self.db.execute(
                """
                SELECT id, message_id, channel, payload, attempt_count
                FROM submission_queue
                WHERE status = 'pending'
                ORDER BY created_at ASC
                LIMIT ?
                """,
                (limit,),
            ).fetchall()

            if not rows:
                self.db.commit()
                return []

            ids = [row[0] for row in rows]
            placeholders = ",".join("?" for _ in ids)
            self.db.execute(
                f"""
                UPDATE submission_queue
                SET status = 'processing',
                    started_at = datetime('now'),
                    attempt_count = attempt_count + 1
                WHERE id IN ({placeholders})
                """,
                ids,
            )
            self.db.commit()

            result = []
            for row in rows:
                result.append({
                    "id": row[0],
                    "message_id": row[1],
                    "channel": row[2],
                    "payload": json.loads(row[3]),
                    "attempt_count": row[4],
                })

            if result:
                logger.info("Dequeued %d submissions", len(result))
            return result

        except Exception:
            self.db.rollback()
            raise

    def mark_completed(self, submission_id: int) -> None:
        """Markiert eine Submission als erfolgreich verarbeitet."""
        self.db.execute(
            """
            UPDATE submission_queue
            SET status = 'completed', completed_at = datetime('now')
            WHERE id = ?
            """,
            (submission_id,),
        )
        self.db.commit()

    def mark_failed(self, submission_id: int, error: str, requeue: bool = True) -> None:
        """
        Markiert eine Submission als fehlgeschlagen.
        Bei requeue=True und attempt_count < max_attempts: zurück auf 'pending'.
        """
        row = self.db.execute(
            "SELECT attempt_count, max_attempts FROM submission_queue WHERE id = ?",
            (submission_id,),
        ).fetchone()

        if not row:
            return

        attempt_count, max_attempts = row[0], row[1]

        if requeue and attempt_count < max_attempts:
            # Zurück in die Queue für Retry
            self.db.execute(
                """
                UPDATE submission_queue
                SET status = 'pending', last_error = ?, started_at = NULL
                WHERE id = ?
                """,
                (error, submission_id),
            )
            logger.info(
                "Submission %d fehlgeschlagen (Versuch %d/%d) — requeued",
                submission_id,
                attempt_count,
                max_attempts,
            )
        else:
            # Endgültig fehlgeschlagen
            self.db.execute(
                """
                UPDATE submission_queue
                SET status = 'failed', last_error = ?, completed_at = datetime('now')
                WHERE id = ?
                """,
                (error, submission_id),
            )
            logger.error(
                "Submission %d ENDGÜLTIG fehlgeschlagen (Versuch %d/%d): %s",
                submission_id,
                attempt_count,
                max_attempts,
                error,
            )
        self.db.commit()

    def get_queue_stats(self) -> dict:
        """Gibt Statistiken über die Queue zurück."""
        stats = {}
        for status in ("pending", "processing", "completed", "failed"):
            count = self.db.execute(
                "SELECT COUNT(*) FROM submission_queue WHERE status = ?",
                (status,),
            ).fetchone()[0]
            stats[status] = count
        return stats

    def dedup_check(self, message_id: str) -> Optional[str]:
        """
        Prüft, ob eine message_id bereits existiert.

        Returns:
            Den Status ('pending', 'processing', 'completed', 'failed') oder None.
        """
        row = self.db.execute(
            "SELECT status FROM submission_queue WHERE message_id = ?",
            (message_id,),
        ).fetchone()
        return row[0] if row else None