serialization and deserialization of the index

[galop1n]

It would be quite useful to be able to save/load an index. I do not see a way to do so in the current interface

[marcus-pousette]

Not at the moment. I will do some work on this lib in the near future (weeks, month time) to improve this library, but have not considered this yet since you can very quickly rebuild the index if you got the original documents at hand (basically re-insert them). Would be interesting to know whether you have any particular use case where rebuilding the index is not applicable?

[BafS]

When you have live and heavy data this could be important. It would also make it possible to share the index across clients, like that you don't have to rebuild in every client. It's typically not possible to build the index on the backend if you have encrypted data.