[Bioperl-l] SearchIO speed up

[Sendu Bala]

Chris Fields wrote:
> I suppose you could do it that way.  What I was thinking would be having
> hit() and query() methods directly in HSPI (no inheritance)

Hilmar already suggested HSPIs would be best left as SimilarityPairs, so 
inheritance is required.


> They would return Bio::SeqFeature::Similarity objects directly.  Though I
> don't really know if it's possible, and also don't really see the purpose
> here if the speedup is marginal.

It's the rate-limiting step after all my other speedups (not considered 
in the present discussion). And if we can have a 50% speedup for 'free', 
why would we say no?

My changes in experimental branch already do something similar: new() 
calls SUPER but just doesn't pass it any args. That's a little slower 
than this (calling Bio::Root::Root::new directly), and you miss out on 
passing -verbose et al.