Better handling of adaptation and algorithm parameters #24
Conversation
Taking the additional parameters out of the adaptation object and making them a member of sampler. Referring to the Sampler object directly to get the algorithm parameters in the examples.
There was a problem hiding this comment.
This all looks good to me and seems to resolve a problem I was having elsewhere, so thank you...
There should be an entry in NEWS for a change of this magnitude, though...
(If you find yourself needing to make a lot of changes to whitespace it can be convenient to commit those separately so that it's apparent that none of the changes do anything of substance...)
|
Second Adam's comment re NEWS, lovely that you keep ChangeLog current with good detail, a higher level change entry for NEWS.Rd is always good too. |
|
Okay, thanks. I had been writing the entries after the pull requests so I can reference the pull request, but I guess I know what number it will be beforehand anyway. I'll update that in the next pull request. |
Following on from our discussion in #23, I'm now passing the algorithm parameters to all of the necessary functions.
The first commit separates the algorithm parameters and adaptation object. The algorithm parameters are passed to the adaptation functions and accessed elsewhere through Sampler->GetAlgParams(). The second commit goes further to pass the algorithm parameters to the moveset functions as well.
Sorry about all of the white space changes hiking up the addition/deletion numbers. I'm not even sure how I would have gone about changing a whole section of indentation like that, but I had the ignore white space option in Gitkraken turned on so I didn't notice (I've turn it off now).