Pandas warning in fairness metrics

[mmosc]

Hi!
When computing the KL divergence, I get the following warnings from pandas:

/home/marta/miniconda3/envs/deap/lib/python3.8/site-packages/recsyslearn/fairness/utils.py:92: FutureWarning: In a future version, the Index constructor will not infer numeric dtypes when passed object-dtype sequences (matching Series behavior)
  top_n = top_n.merge(rel_matrix, on=["user", "item", "group"], how="outer")

I think we should try to conform to pandas' suggestion (or maybe silence the warning... 😅 )

Ciao!
Marta

[giuliowaitforitdavide]

Ok, we should take care of that. As alrrady told you, we should investigate if it's something due to some legacy code in Pandas library or from our side because I'm not sure about that. Would you give it a try? I will look after that too

[mmosc]

Okay! I'll look into that.