import logging
from FlagEmbedding import FlagReranker
from ..config import get_settings

logger = logging.getLogger(__name__)

_reranker: FlagReranker | None = None


def _get_reranker() -> FlagReranker:
    global _reranker
    if _reranker is None:
        settings = get_settings()
        logger.info(f"[Reranker] Loading model: {settings.reranker_model}")
        _reranker = FlagReranker(settings.reranker_model, use_fp16=False)
        logger.info(f"[Reranker] Model loaded successfully: {settings.reranker_model}")
    return _reranker


def rerank(query: str, passages: list[str]) -> list[float]:
    if not passages:
        return []
    reranker = _get_reranker()
    pairs = [[query, p] for p in passages]
    # batch_size=8 limits peak RAM usage — critical for cloud CPU-only environments
    scores = reranker.compute_score(pairs, normalize=True, batch_size=8)
    if isinstance(scores, float):
        scores = [scores]
    return [float(s) for s in scores]