Is there a preferred way to store drfps?

[cthoyt]

I've just spent some time to automate generating drfps from the Rhea database (see here) and would like to save a pre-cached version for later use. The word2vec world has a specific format for its embedding vectors - do you have a standard way for storing a set of drfps where each has a string key associated?

I suppose I could just store a pickle of my Pandas dataframe, or some other serialization format, but that seems a bit space inefficient

[daenuprobst]

Hey @cthoyt, sorry I completely missed this issue. I guess you have solved it but if someone has the same question issue, I usually go with gzip and pickle. Numpy's serialization might be an option as well.