Function perform_clustering

_cluster_utils.py:244–283

(
    embeddings: np.ndarray, dim: int, threshold: float, verbose: bool = False,cluster_size: int = 20
)

def perform_clustering(
    embeddings: np.ndarray, dim: int, threshold: float, verbose: bool = False,cluster_size: int = 20
) -> List[np.ndarray]:
    reduced_embeddings_global = global_cluster_embeddings(embeddings, min(dim, len(embeddings) -2))
    global_clusters, n_global_clusters = GMM_cluster(     # (num, 2)
        reduced_embeddings_global, threshold,cluster_size=cluster_size
    )
    if len(global_clusters) != len(embeddings):
        print('debug')
    # if verbose:
    #     logging.info(f"Global Clusters: {n_global_clusters}")

    # all_clusters = [[] for _ in range(len(embeddings))]
    # embedding_to_index = {tuple(embedding): idx for idx, embedding in enumerate(embeddings)}
    # for i in tqdm(range(n_global_clusters)):
    #     global_cluster_embeddings_ = embeddings[
    #         np.array([i in gc for gc in global_clusters])
    #     ]  #找到当前簇的embedding
    #     if verbose:
    #         logging.info(
    #             f"Nodes in Global Cluster {i}: {len(global_cluster_embeddings_)}"
    #         )
    #     if len(global_cluster_embeddings_) == 0:
    #         continue

    #     # embedding indices #反向取idx
    #     indices = [
    #         embedding_to_index[tuple(embedding)]
    #         for embedding in global_cluster_embeddings_
    #     ]

    #     # update
    #     for idx in indices:
    #         all_clusters[idx].append(i)

    # all_clusters = [np.array(cluster) for cluster in all_clusters]

    # if verbose:
    #     logging.info(f"Total Clusters: {len(n_global_clusters)}")
    return global_clusters

def enclose_string_with_quotes(content: Any) -> str:
    """Enclose a string with quotes"""