nsh87 / receptormarker

Source for 'receptormarker' package for R: antibody receptor and phenotypic marker analysis
http://receptormarker.com
BSD 2-Clause "Simplified" License
4 stars 9 forks source link

Add PCA legend #65

Closed nsh87 closed 8 years ago

nsh87 commented 8 years ago

This is a rebase based on PR #58.

catterbu commented 8 years ago

@nsh87 looks like the one commit in this PR is the only one that is needed. As I recall, when i tried it, the colors did not always match up as desired in the legend. I may have been incorrect in this observation, but you may check it on your local before merging it into dev.

nsh87 commented 8 years ago

@catterbu: going to merge everything to dev and check it out on dev and local before pushing to production. had a bit of a hiccup today and accidentally put dev on production for a bit and there were definitely issues cluster, just a heads up. it might be worth it for you to get the site running on your local so you can test things out with the GUI. potentially much faster. a bit of a complicated set up a the moment, though.

catterbu commented 8 years ago

@nsh87 Are the directions for getting the site running locally on the ReceptorMarkerFrontEnd readme? Any idea what the particular issues are right now?

nsh87 commented 8 years ago

@catterbu: ya they are, but i made a bunch of crazy changes recently and there are some manual migrations that must be done to take your existing installation and upgrade it to current. the alternative is for ReceptorMarkerFrontend and ReceptorMarkerBacken repos to just delete what you have, clone the current masters and start things up from there. That should work, but haven't tried it yet.

catterbu commented 8 years ago

@nsh87 okay. I have not tried running it before on Linux anyway. If you are able to tell me about any of the errors, I can start exploring them.

nsh87 commented 8 years ago

@catterbu: I just sent you a G Drive folder with a couple files. I'd run multi_clustering() on the first two files after removing the unneeded columns. The second file is boolean. Then try to generate all the plots with them. The site uses krange=2:15, so use that, too. If there aren't any errors it should be all fine. Then try the file in the test folder, too and see what happens. I have a bunch of files on my local that I've played around with over the past few months. I'll get some more over to you tomorrow.

nsh87 commented 8 years ago

@catterbu: I started some instructions to get Vagrant and VirtualBox working on Ubuntu here: https://github.com/nsh87/mimic-iii-vm/tree/dev. I need to test them on an actual Ubuntu machine, so might as well extract the relevant portion from the code block in the Requirements section there.