In 0.0.6.8.6, document differing by formating (cap, extra spaces, comma, etc.) stay duplicated. The function that compare documents at import should make a tolower on all strings and remove space and punctuation before comparison. A python script from Quentin already exists for the clean-up : https://gitlab.iscpif.fr/davidchavalarias/gargantext_ressources/blob/master/script/removeDuplicates.py