Value of min_similarity attribute is overwritten in merge_models function

[Ceglowa]

In

BERTopic/bertopic/_bertopic.py

Line 3020 in bcb3ca2

def merge_models(cls, models, min_similarity: float = .7, embedding_model=None):

you are able to pass the variable min_similarity, which is 0.7 by default.
However, in

BERTopic/bertopic/_bertopic.py

Line 3089 in bcb3ca2

min_similarity = 0.7

the value is being overwritten with 0.7 value. Immediately after the assignment, the variable is used. So that means, that the value passed in the function doesn't influence the merging process

[MaartenGr]

That is correct and indeed an issue! Fortunately, this is already fixes in the upcoming PR which I hope to release in the upcoming weeks after some additional documentation changes.

[Ceglowa]

Hi @MaartenGr. I see that this issue is already fixed. Thanks for fixing that. Closing the ticket