@reicruz we'll now cleanup old versions of the classifier when version # in .csv file name is increased. The nlcManager.trainIfNeeded() call will initiate the training of the newer version.
I also added a prefix env var, in the off chance that multiple instances of the bot are running at different version levels and pointing to the same NLC instance. That way one could use this prefix to prevent one instance from deleting the classifier used by the other instance.
Coverage increased (+0.7%) to 81.25% when pulling c47936b540167273dfae44a64a060b7ebe94cf64 on issue/3 into 28b29968d8ae262d199bdbc1d6c5f47248d27bb9 on master.
@reicruz we'll now cleanup old versions of the classifier when version # in .csv file name is increased. The nlcManager.trainIfNeeded() call will initiate the training of the newer version.
I also added a prefix env var, in the off chance that multiple instances of the bot are running at different version levels and pointing to the same NLC instance. That way one could use this prefix to prevent one instance from deleting the classifier used by the other instance.