Custom distance functions no longer supported?

[rabdill]

Hello! bigmds has been very helpful in my research, and I was enthusiastic to start using the new version. However, it looks like the dist_fn parameter has been removed from all the places where it used to be an option. This makes my current analysis much more complicated, as it appears the only option now is Euclidean distance. I could be wrong about this, but the change isn't noted in the news file or described in the commit where it was changed, so would you mind clarifying the situation for those who need to use other distance measures to generate their MDS plots? Is it safe for me to use the previous version of the package that still had this option, or does the change address a bug I should try to avoid? Thanks for your help, and for putting together such a helpful package.

[pachoning]

Hi @rabdill , first of all, sorry for not making this huge change clearer. We had to get rid of the dist_fn parameter because a major review in our scientific paper forced us to do it. Said so, let me clarify some things:

• The previous version works perfectly with this parameter.
• As soon as I can, I will try to incorporate this parameter back in our code.

I am gonna keep this issue open so that I can address it as soon as possible.

[rabdill]

Thanks for such a quick response, and I'm sorry you're dealing with a frustrating situation. There is no need to rush on my account—it's good to know the v2.0.1 code is still safe to use in the meantime. Best of luck in the review process!

[dylanbeaudette]

Hi, thanks for this amazing package. I too would like to request that the dist_fn option be added back to the functions. I deal with specialized distance metrics far more frequently than Euclidean distance.

[pachoning]

Hi @dylanbeaudette and @rabdill , I know having back thedist_fn would be great. I am about to finish the PhD. As soon as I get some free time, I will add it back for sure.

[dylanbeaudette]

Good luck with the Ph.D!