Saving template networks which can be just duplicated?


#1

Pasted from github

there is finite number of possible networks in this system. can we save some server time by saving template networks which can be just duplicated when same request for a new network comes?
example: user requests a new network of German procurement from Agriculture field in year 2012-2014. if this network was already queried before, it is stored in DB and therefore its data can be just duplicated. is this worth trying?

  • is this feasible?
  • would it pose problems with saving clusters? If so, perhaps we could have an instance of ‘virgin’ and ‘clustered’ network?

(Georgiana B) #2

Indeed, we create a new network every time even though the query is the same because the network can be changed via clustering/declustering.

However, there is obviously room for improvement. I think it may be more efficient to have some smart caching/indexing than to keep an extra copy of the network.

I will look more into this when times comes for backend performance improvements.