WRScpp for other platforms other than x86_64-apple-darwin13.1.0 #3
Comments
|
Hi Jeremy, Sorry for the late reply. I am not familiar with your platform. Is your platform a PC? If so, can you try this: https://github.com/mrxiaohe/WRScppWin |
|
Hi Xiao He, Sorry for the mistake on the platform. Meanwhile, may I ask if this package also support installation in a Linux platform as well ? |
|
Hi Jeremy, I don't have a Linux machine to compile the package. Sorry! Best, |
|
@mrxiaohe Thanks for providing the raw files and doing the hard work on this. 👍 I was able to compile it on a linux machine (Linux Mint 17 Qiana (GNU/Linux 3.13.0-24-generic x86_64)) using the raw code, but I had to make some adjustments to the robustmethods_CPP.cpp file and the robustmethods_CPP.R files. I'm not sure if I did it correctly as this is the first time I have built a package from source by hand in R and I am relatively new to R and C++ in general. I'll be happy to share my experience, but I can't confirm that I've done everything correctly or that I even remember exactly what I did. There is a lot in there that I don't fully understand. edit: (I just noticed you made an update to the raw code 23 hours ago. I was using the older version. Some of the adjustments that I made I think would still need to be made in the more recent code (e.g. tsp1reg_C returns only coef (which is unnamed, so temp1$coef is not valid) and not res). Again, I'm new to this, so I may just be using it incorrectly. I may try with this new version this weekend and see if I can document what I have done to get it to work for me with the expected results.) |
|
Hi Joe, I didn't change the actual functions in the raw file. So it shouldn't affect the performance of any of the functions. Can you start a repository for the updated code that you are using? Perhaps you can also upload the compiled file. Also it would be great if you could note the changes you made. I will add links to your repository in my repositories, so that people who want a Linux version can get it from you. |
|
Hey Joe, I added your links in my WRSCPP Readme. Thank you! Best, |
|
Hi Xiao, Best, Joe |
|
Hi Joe, The strange thing is that after I restarted my computer, I was able to compile the code...I have no idea what was going on. That's why I didn't pursue it further. What warnings did you get? I got warnings when I compiled the code on my Mac. After I did some investigative Googling, it seemed that the warnings were OK. |
|
Here are all the warnings I got when I compile on Linux. They seem to be related to the errors I am getting on my Mac. I restarted my mac and the errors still persist. Joe |
|
Hi Joe, The problem has to do with the headers included. But the issue is that some of the functions included need some of these headers (I think the functions that compute robust location and scatter estimators need those. But also, did you get error messages or warning messages. Often times warning messages are not serious, and you can often still load and use the compiled code. |
|
Hi Joe, I think I solved the Fortran problem. I created a new branch in the repository for the raw code. You can try compile the code by using the following command in your terminal:
You need Let me know if it worked. |
|
Hi Xiao He, It seems to work OK. Still some warnings about undeclared/unused variables, but that's not a big deal. I was able to compile it from Linux, Windows, and Mac. There are still some issues with things like tsreg_c not returning the coefficients and residuals in the expected way (in robustfunctions_cpp.R) and with exporting some of the functions. My WRScpp_test repository has these fixes (I think) and it what I used to compile on all my machines using Best, Joe |
|
Re: tsreg_C: that is a bit strange because it returns coefs on my computer: |
|
It is strange. When I install your Mac Binary, I have no problem. However, when I use the As I understand (and I said before, I am no expert) temp1<-.Call("tsp1reg_C", X=x, Y=y, HD=as.integer(HD))
coef<-temp1$coef
res<-temp1$resto coef<-.Call("tsp1reg_C", X=x, Y=y, HD=as.integer(hd))
res<-y-coef[2]*x-coef[1]I had also had problems with the call to temp1<-.Call("tsreg_for", X=x, Y=y, IT=as.integer(iter), HD=as.integer(HD))
coef<-c(temp1$alpha,temp1$beta)
res<-temp1$resto temp1<-.Call("tsreg_C", X=x,Y=y, IT=as.integer(iter), HD=as.integer(hd))
coef<-temp1$coef
res<-temp1$resas I noticed that There were also some functions in This is kind of out of topic for the thread. If you want, I can start an issue document on your raw code. Best, Joe |
Hi,
After installing the package WRScpp, I tried to use the package by calling library(WRScpp).
However, the following error is received.
Error: package 'WRScpp' was built for x86_64-apple-darwin13.1.0
May I ask if this package will be usable in other platforms in the near future say
Platform: x86_64-w64-mingw32/x64 (64-bit)
Regards,
Jeremy
The text was updated successfully, but these errors were encountered: