Function must_read_replace

scripts/drift_engine.py:986–1042

蹇呰浣滆€?鏈烘瀯婕傜Щ锛氭浛鎹㈤儴鍒嗗繀璇讳綔鑰呭拰鏈烘瀯

(
    profile: Dict[str, Any],
    params: Dict[str, Any],
    selected_papers: Optional[List[Dict]] = None,
    date: Optional[str] = None,
)

def must_read_replace(
    profile: Dict[str, Any],
    params: Dict[str, Any],
    selected_papers: Optional[List[Dict]] = None,
    date: Optional[str] = None,
) -> Tuple[Dict[str, Any], Dict[str, Any]]:
    """
    蹇呰浣滆€?鏈烘瀯婕傜Щ锛氭浛鎹㈤儴鍒嗗繀璇讳綔鑰呭拰鏈烘瀯
    """
    updated = copy.deepcopy(profile)
    must_read = updated.get("must_read", {})

    replace_ratio = params.get("replace_ratio", 0.5)

    # 鏇挎崲浣滆€?
    old_authors = list(must_read.get("authors", []))
    if old_authors and random.random() < replace_ratio:
        keep_count = max(0, int(len(old_authors) * (1 - replace_ratio)))
        new_authors_pool = params.get("new_authors_pool", [])

        # 淇濈暀閮ㄥ垎鏃т綔鑰?
        kept_authors = random.sample(old_authors, keep_count) if keep_count > 0 else []

        # 娣诲姞鏂颁綔鑰?
        new_authors_count = len(old_authors) - keep_count
        new_authors = random.sample(new_authors_pool, min(new_authors_count, len(new_authors_pool)))

        must_read["authors"] = kept_authors + new_authors

    # 鏇挎崲鏈烘瀯
    old_institutions = list(must_read.get("institutions", []))
    if old_institutions and random.random() < replace_ratio:
        keep_count = max(0, int(len(old_institutions) * (1 - replace_ratio)))
        new_institutions_pool = params.get("new_institutions_pool", [])

        kept_institutions = random.sample(old_institutions, keep_count) if keep_count > 0 else []
        new_institutions_count = len(old_institutions) - keep_count
        new_institutions = random.sample(
            new_institutions_pool, min(new_institutions_count, len(new_institutions_pool))
        )

        must_read["institutions"] = kept_institutions + new_institutions

    updated["must_read"] = must_read

    # 鏇存柊 drift_state
    updated["drift_state"] = _update_drift_state(profile, date, score_delta=0.2)

    drift_event = {
        "method": "must_read_replace",
        "authors_before": old_authors,
        "authors_after": list(must_read.get("authors", [])),
        "institutions_before": old_institutions,
        "institutions_after": list(must_read.get("institutions", [])),
    }

    return updated, drift_event