Open matthewhorridge opened 5 years ago
@fanavarro I've transferred this issue from the mailing list to here.
While I don't currently have any tips to improve the loading performance of ontologies, I do have a question for you: In your production environment, are all the ontologies very large, like NCBI Taxon or are the likely to be smaller? Much smaller?
Hi @matthewhorridge.
A great part of our ontologies are smaller. However, we have a couple of them that are larger than NCBI Taxon.
@fanavarro the reason for me asking is that I was wondering whether some more fine-grained dormant project time could be necessary. Small ontologies could have short default purge times and larger ontologies could have specific purge times set for them.
I've been doing some profiling and there's some improvements that I can make. Load time would be 2-3 times quicker, not an order of magnitude quicker, but the improvements are worth making. They require a fair bit of work though, so not something that could be released in the next few days.
Thanks for your answer @matthewhorridge. What you said is very interesting, it makes sense that purge time should be dependent on load time. I will check our settings.
Having a load time 2-3 times quicker would be fantastic, but I know this kind of improvement is difficult to achieve.
Thanks for your answer again!
From the mailing list: