Function _resize

paperflow/providers/embedding.py:30–48

(vector: List[float], target_dim: int)

def _resize(vector: List[float], target_dim: int) -> List[float]:
    if not vector:
        return [0.0] * target_dim
    if len(vector) == target_dim:
        return vector
    if len(vector) < target_dim:
        return vector + [0.0] * (target_dim - len(vector))

    bucket = len(vector) / target_dim
    out: List[float] = []
    for index in range(target_dim):
        start = int(index * bucket)
        end = int((index + 1) * bucket) or start + 1
        segment = vector[start:end]
        out.append(sum(segment) / len(segment))
    norm = sum(value * value for value in out) ** 0.5
    if norm > 0:
        out = [value / norm for value in out]
    return out


class OpenAIEmbedding: