reduce_outliers result has not been updated.

Nobody has claimed this yet.

Hello,
Because the quantity of -1s comprises approximately 50% of all the quantities, it is desired to reduce the outliers. The code is as follows:

from bertopic import BERTopic
import pandas as pd
import nltk
from nltk.corpus import stopwords

# 加载数据集
df1 = pd.read_csv("/content/drive/MyDrive/VR_origin_0115_remove_update_notime.csv")
text_column = 0
# 创建BERTopic模型
model = BERTopic(nr_topics="auto")
# 将文本转换为向量，并聚类
topics, probs = model.fit_transform(df1.iloc[:, text_column])
new_topics = model.reduce_outliers(df1.iloc[:, text_column],topics)
# 获取主题的关键词

model.get_topic_info()

After running the code, I found that the result remains the same as the original model and does not reflect the reduction of outliers. I would like to inquire about where the problem lies and how to resolve it.

Note:
1.The goal is to ensure that the final result obtained from get_topic_info reflects the reduction of outliers.
2.I have tried using topic_model.update_topics(docs, topics=new_topics), and while the result reflects the updated topics, the information corresponding to -1 is deleted. However, I still want to retain the outliers; I just wish to reduce the quantity of -1 occurrences due to their significant proportion.

Reproduce the supplied BERTopic example and compare reduce_outliers, get_topic_info, and update_topics using the same topics and documents. Start by tracing how reduce_outliers updates topic assignments and how get_topic_info reads them; done means the reported -1 count is reduced while an outlier topic remains visible.

Written by the indexing model from the issue text.