import logging
from datetime import datetime, timedelta

logger = logging.getLogger("order_sync")

class OrderSyncJob:
    """增量同步订单到数仓，容忍上游乱序与重复投递。"""

    def __init__(self, source, sink, window_minutes=30):
        self.source = source
        self.sink = sink
        self.window = timedelta(minutes=window_minutes)

    def run(self, since=None):
        since = since or datetime.utcnow() - self.window
        batch = self.source.fetch_updated(since=since)
        logger.info("fetched %s rows since %s", len(batch), since)
        deduped = self._dedupe(batch)
        self.sink.upsert(deduped)
        return len(deduped)

    def _dedupe(self, rows):
        seen, out = set(), []
        for r in rows:
            if r.order_id in seen:
                continue
            seen.add(r.order_id)
            out.append(r)
        return out