Function compute_bm25_tf

tests/test_hybrid.py:248–251

(
    raw_tf: int, doc_len: int, avg_doc_len: float, k: float = K1, b: float = B
)

def compute_bm25_tf(
    raw_tf: int, doc_len: int, avg_doc_len: float, k: float = K1, b: float = B
) -> float:
    return raw_tf * (k + 1) / (raw_tf + k * (1 - b + b * doc_len / avg_doc_len))


def merge_cqa_dupstack(data_path: str, verbose: bool = False):