Suggested feature: antsSaveAll

[dorianps]

It would be extremely useful to have a function that can save the entire workspace in a single line. Right now is very difficult to keep track of what is a pointer and what is not, then save them separately, then write code again to load them on new session. A save-all function can take this burden away, create cleaner code with one-line saving, and offer to restart at different levels of a script. Of course, this should be complemented with a load function.

[bkandel]

That's a good idea. Shouldn't be too hard to do. I almost never save my
workspace, but this would be helpful to have around. I'll try to get to
this tomorrow.

2015-03-03 9:55 GMT-05:00 dorianps [email protected]:

It would be extremely useful to have a function that can save the entire
workspace in a single line. Right now is very difficult to keep track of
what is a pointer and what is not, then save them separately, then write
code again to load them on new session. A save-all function can take this
burden away, create cleaner code with one-line saving, and offer to restart
at different levels of a script.

—
Reply to this email directly or view it on GitHub
#16.

[dorianps]

There are multiple benefits:

1. You can restart at any point you left with days in between.
2. You can save in one window, and load another window to proceed with a different parameter (no need to run all the code until that point).
3. You can avoid having hours lost by unexpected events (blackouts, connection problems).
4. Will make the software more usable.

Last night there were multiple network interruptions, so Rstudio died. I spent 2 hours this morning just to reach the same state I had when I started the analysis last night.

[bkandel]

Just as a practical matter--do you mount the cluster drive? Mounting the
cluster and then running R/Rstudio on your local machine is much less
sensitive to network blips than running R on the cluster.

2015-03-03 10:18 GMT-05:00 dorianps [email protected]:

There are multiple benefits:

1. You can restart at any point you left with days in between.
2. You can save in one window, and load another window to proceed with a
   different parameter (no need to run all the code until that point).
3. You can avoid having hours lost by unexpected events (blackouts,
   connection problems).
4. Will make the software more usable.

Last night there were multiple network interruptions, so Rstudio died. I
spent 2 hours this morning just to reach the same state I had when I
started the analysis last night.

—
Reply to this email directly or view it on GitHub
#16 (comment).

[dorianps]

Hmmm, I only mount it as external drive. You mean I might be able to point my Windows Rstudio to the cluster R in linux and make it work? That would be great. Let me know if you do that.

[bkandel]

Hmm...no, that would probably be worse. I'm a bit surprised then that
Rstudio died when the cluster connection failed. I guess it tried to read
or write and couldn't and then hung?

2015-03-03 10:36 GMT-05:00 dorianps [email protected]:

Hmmm, I only mount it as external drive. You mean I might be able to point
my Windows Rstudio to the cluster R in linux and make it work? That would
be great. Let me know if you do that.

—
Reply to this email directly or view it on GitHub
#16 (comment).

[dorianps]

Yes, off topic, but the way I use it is by calling Rstudio from cluster in an X-server. If the connection to the cluster breaks, Rstudio can't push the display, so it dies. There is an rstudio-server supposed to be better, I am waiting for Michael to install it. Nevertheless, a save function for antsR will be necessary sooner or later.

[bkandel]

Ah, OK. So what I was suggesting was to run RStudio on your local machine,
and then just point it to the (mounted) cluster drive to read whenever it
has to load data. That way if the connection gets lost, Rstudio will be
fine and if need be you can save your work locally to continue later.

2015-03-03 11:07 GMT-05:00 dorianps [email protected]:

Yes, off topic, but the way I use it is by calling Rstudio from cluster in
an X-server. If the connection to the cluster breaks, Rstudio can't push
the display, so it dies. There is an rstudio-server supposed to be better,
I am waiting for Michael to install it. Nevertheless, a save function for
antsR will be necessary sooner or later.

—
Reply to this email directly or view it on GitHub
#16 (comment).

[dorianps]

Problem is there is no antsR in Windows, haven't figured out a way to install it.

[bkandel]

Just pushed save.ANTsR and load.ANTsR, which should solve this problem. It works for my test cases, but if you could make sure it works for you that would be helpful. Ended up being a little trickier than I thought it would be.

[stnava]

generated a new warning from R CMD check

• checking R code for possible problems ... NOTE

load.ANTsR: no visible binding for global variable ‘ANTsRImageData’

Found the following assignments to the global environment:

File ‘ANTsR/R/save.ANTsR.R’:

assign(as.character(ANTsRImageData$names[ii]),
antsImageRead(file.path(filename,

paste(as.character(ANTsRImageData$names[ii], ".nii.gz", sep = ""))),

ANTsRImageData$dims[ii], ANTsRImageData$pixeltypes[ii]),

envir = .GlobalEnv)

• checking Rd files ... OK

brian

On Thu, Mar 5, 2015 at 10:58 PM, bkandel [email protected] wrote:

Just pushed save.ANTsR and load.ANTsR, which should solve this problem. It
works for my test cases, but if you could make sure it works for you that
would be helpful. Ended up being a little trickier than I thought it would
be.

—
Reply to this email directly or view it on GitHub
#16 (comment).

[bkandel]

Shoot. Sorry about that. Should be fixed now.

2015-03-06 16:14 GMT-05:00 stnava [email protected]:

generated a new warning from R CMD check

• checking R code for possible problems ... NOTE

load.ANTsR: no visible binding for global variable ‘ANTsRImageData’

Found the following assignments to the global environment:

File ‘ANTsR/R/save.ANTsR.R’:

assign(as.character(ANTsRImageData$names[ii]),
antsImageRead(file.path(filename,

paste(as.character(ANTsRImageData$names[ii], ".nii.gz", sep = ""))),

ANTsRImageData$dims[ii], ANTsRImageData$pixeltypes[ii]),

envir = .GlobalEnv)

• checking Rd files ... OK

brian

On Thu, Mar 5, 2015 at 10:58 PM, bkandel [email protected] wrote:

Just pushed save.ANTsR and load.ANTsR, which should solve this problem.
It
works for my test cases, but if you could make sure it works for you that
would be helpful. Ended up being a little trickier than I thought it
would
be.

—
Reply to this email directly or view it on GitHub
#16 (comment).

—
Reply to this email directly or view it on GitHub
#16 (comment).

[stnava]

still seeing:

Found the following assignments to the global environment:
File ‘ANTsR/R/save.ANTsR.R’:
assign(as.character(ANTsRImageData$names[ii]), antsImageRead(file.path(filename,
paste(as.character(ANTsRImageData$names[ii], ".nii.gz", sep = ""))),
ANTsRImageData$dims[ii], as.character(ANTsRImageData$pixeltypes[ii])),
envir = .GlobalEnv)

[stnava]

i think you need something like this:

http://www.talkstats.com/showthread.php/38050-CRAN-submission-problem-help-brainstorming

[bkandel]

Ah, that's a good idea. Should've realized that would be considered
"anti-social" by CRAN.

2015-03-08 13:02 GMT-04:00 stnava [email protected]:

i think you need something like this:

http://www.talkstats.com/showthread.php/38050-CRAN-submission-problem-help-brainstorming

—
Reply to this email directly or view it on GitHub
#16 (comment).