From mishkind at gmail.com  Sun Jul  1 14:56:44 2012
From: mishkind at gmail.com (Mishkin Derakhshan)
Date: Sun, 1 Jul 2012 14:56:44 -0400
Subject: [MINC-users] Official Installation Instructions
Message-ID: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>

Hi,

So I needed to install minc on a fresh ubuntu 12.04 32-bit machine at
home and I thought I would approach the issue as any other potential
minc user would by googling "minc".

There are a few hits that look right, but going with this is the first
sensible one:

http://www.bic.mni.mcgill.ca/ServicesSoftware/MINC/
which has some links to debian packages, but sadly, not the one I am
looking for. I also see links to packages.bic.mni.mcgill.ca/tgz/ to
download some tarballs. Maybe that is what I should do?

If (by chance) I click on the menus on the left I can stumble upon this:
http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit

This also looks promising, but again I don't see the exact version I'm
looking for, and 32-bit support sounds iffy. I remember seeing a bunch
of other links on the previous webpage (and google) so I check out:

http://en.wikibooks.org/wiki/MINC
which has some nice details about what is supported and gives me two more links.
http://en.wikibooks.org/wiki/MINC/Tutorials/BinaryInstalls
which for me is no good because it only goes up to ubuntu 10.10 and in
64 bit, so I look at this:
http://en.wikibooks.org/wiki/MINC/Tutorials/SourceInstall
which again has some nicely detailed instructions, and it looks like
this is what I want:
https://github.com/vfonov/EZminc/raw/master/build_quarantine/Makefile

I've used this in the past, and it has worked so I am inclined to just
use it. BUT, I remember this google hit for MINC:
https://github.com/BIC-MNI/

And I think to myself, well maybe I should use the EZminc from BIC-MNI
instead of vfonov. Seems more official.
Looking on that website though, it seems that what I really want is
this (there was a link to it on this page too
http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit)

https://github.com/BIC-MNI/minc-toolkit

The README displayed at the bottom of the page looks great, though it
lists dependencies for only 11.10, and there is a bad link (I'm
guessing, this should be BIC-MNI instead of vfonov):
git clone git://github.com/vfonov/minc-toolkit.git minc-toolkit

I know that many of these pages exist to satisfy the wide range of
systems that we are trying to support, but it would be useful to have
a "blessed" set of instructions. For example, if you google FSL or
Freesurfer, within 1 or 2 very easily guided clicks you know what to
do.

My suggestion, is to keep one of the pages (wikibooks, github, BIC
webpage) up to date, and have the other ones all have big bold links
that tell you to go to the "official" installation instructions page.

Sorry for the long rant,

mishkin

From lconcha at gmail.com  Sun Jul  1 19:37:01 2012
From: lconcha at gmail.com (Luis Concha)
Date: Sun, 1 Jul 2012 18:37:01 -0500
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
Message-ID: <CAOq9ZoHoNw+V+GsvRmc5-5hy5P+UGY2S+dJNjJkQmHSO4KJA0A@mail.gmail.com>

As an end-user, I could not agree more. More than half of the posts to this
mailing list are about compilation/installation issues, rather than usage
questions. Minc tools are great, but widespread adoption can only happen if
they are accessible, and wide spread adoption means quicker advancement and
evolution towards more greatness.
On Jul 1, 2012 2:00 PM, "Mishkin Derakhshan" <mishkind at gmail.com> wrote:

> Hi,
>
> So I needed to install minc on a fresh ubuntu 12.04 32-bit machine at
> home and I thought I would approach the issue as any other potential
> minc user would by googling "minc".
>
> There are a few hits that look right, but going with this is the first
> sensible one:
>
> http://www.bic.mni.mcgill.ca/ServicesSoftware/MINC/
> which has some links to debian packages, but sadly, not the one I am
> looking for. I also see links to packages.bic.mni.mcgill.ca/tgz/ to
> download some tarballs. Maybe that is what I should do?
>
> If (by chance) I click on the menus on the left I can stumble upon this:
> http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit
>
> This also looks promising, but again I don't see the exact version I'm
> looking for, and 32-bit support sounds iffy. I remember seeing a bunch
> of other links on the previous webpage (and google) so I check out:
>
> http://en.wikibooks.org/wiki/MINC
> which has some nice details about what is supported and gives me two more
> links.
> http://en.wikibooks.org/wiki/MINC/Tutorials/BinaryInstalls
> which for me is no good because it only goes up to ubuntu 10.10 and in
> 64 bit, so I look at this:
> http://en.wikibooks.org/wiki/MINC/Tutorials/SourceInstall
> which again has some nicely detailed instructions, and it looks like
> this is what I want:
> https://github.com/vfonov/EZminc/raw/master/build_quarantine/Makefile
>
> I've used this in the past, and it has worked so I am inclined to just
> use it. BUT, I remember this google hit for MINC:
> https://github.com/BIC-MNI/
>
> And I think to myself, well maybe I should use the EZminc from BIC-MNI
> instead of vfonov. Seems more official.
> Looking on that website though, it seems that what I really want is
> this (there was a link to it on this page too
> http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit)
>
> https://github.com/BIC-MNI/minc-toolkit
>
> The README displayed at the bottom of the page looks great, though it
> lists dependencies for only 11.10, and there is a bad link (I'm
> guessing, this should be BIC-MNI instead of vfonov):
> git clone git://github.com/vfonov/minc-toolkit.git minc-toolkit
>
> I know that many of these pages exist to satisfy the wide range of
> systems that we are trying to support, but it would be useful to have
> a "blessed" set of instructions. For example, if you google FSL or
> Freesurfer, within 1 or 2 very easily guided clicks you know what to
> do.
>
> My suggestion, is to keep one of the pages (wikibooks, github, BIC
> webpage) up to date, and have the other ones all have big bold links
> that tell you to go to the "official" installation instructions page.
>
> Sorry for the long rant,
>
> mishkin
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From vladimir.fonov at gmail.com  Mon Jul  2 08:46:17 2012
From: vladimir.fonov at gmail.com (Vladimir S. Fonov)
Date: Mon, 02 Jul 2012 08:46:17 -0400
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
Message-ID: <4FF19819.2070607@gmail.com>

Hello,

On 12-07-01 02:56 PM, Mishkin Derakhshan wrote:
> If (by chance) I click on the menus on the left I can stumble upon this:
> http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit
>
> This also looks promising, but again I don't see the exact version I'm
> looking for, and 32-bit support sounds iffy. I remember seeing a bunch
> of other links on the previous webpage (and google) so I check out:

So, did you try to install 32 bit version of the package or not? I mean, 
apart from your curiosity - was there any other reason for you to 
explore the source compilation route?

-- 
Best regards,
  Vladimir S. Fonov ~ vladimir.fonov <at> gmail.com



From mishkind at gmail.com  Mon Jul  2 22:04:04 2012
From: mishkind at gmail.com (Mishkin Derakhshan)
Date: Mon, 2 Jul 2012 22:04:04 -0400
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <4FF19819.2070607@gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
	<4FF19819.2070607@gmail.com>
Message-ID: <CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>

I don't want to turn this into a "please help me with my installation"
thread, but here is the summary of my results. The good news is that 2
out of the 3 methods I tried worked! In true minc fashion there is
more than one way to get 'er done.
That being said, I still think it is frustrating for an end user to
figure out which of the three they should have used.

1. http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit
using the 32-bit version of the toolkit, and the testsuite and the models.
I did have to run sudo apt-get install -f afterwards to install the
missing library (libjpeg or something like that), but all appears to
have installed okay. mincbeast in the pipeline does fail, but
register, Display, mincstats etc. seem to work fine.
ERROR! Allocation failure. Probably out of memory in
/home/vfonov/src/minc-toolkit/BEaST/alloc.c:186.
Died at /opt/minc/pipeline/pipeline_mritotal.pl line 182.
(Odd that /home/vfonov is hardcoded into the error msg?)

2. http://en.wikibooks.org/wiki/MINC/Tutorials/SourceInstall
I followed the instruction from github (substituting BIC-MNI for vfonov)
ie. wget https://github.com/BIC-MNI/EZminc/raw/master/build_quarantine/Makefile
There was some error about mni_autoreg (sorry I didn't copy it), but
everything managed to install anyway so it must have been just a
warning.

3. https://github.com/BIC-MNI/minc-toolkit
I tried compiling from source from git-hub using cmake. I ran into
these problems:
(i) ssh-keys for github.
cd ../Register
  git submodule init
  git submodule update --recursive
fatal: The remote end hung up unexpectedly
Solution: https://help.github.com/articles/generating-ssh-keys

(ii) missing c++ compiler
thinkpad[~/minc-toolkit-build]$ cmake ../minc-toolkit
CMAKE_CXX_COMPILER-NOTFOUND
Solution: sudo apt-get install g++
This should probably be added to the list of dependencies listed in the readme.

(iii) After googling for a while I gave up. Looks like a possible
error in syntax in the code?
thinkpad[~/minc-toolkit-build]$ cmake ../minc-toolkit
CMake Error at EBTKS/CMakeLists.txt:191 (STRING):
  string sub-command REGEX, mode REPLACE needs at least 6 arguments total to
  command.
CMake Error at N3/CMakeLists.txt:89 (STRING):
  string sub-command REGEX, mode REPLACE needs at least 6 arguments total to
  command.


On Mon, Jul 2, 2012 at 8:46 AM, Vladimir S. Fonov
<vladimir.fonov at gmail.com> wrote:
> Hello,
>
>
> On 12-07-01 02:56 PM, Mishkin Derakhshan wrote:
>>
>> If (by chance) I click on the menus on the left I can stumble upon this:
>> http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit
>>
>> This also looks promising, but again I don't see the exact version I'm
>> looking for, and 32-bit support sounds iffy. I remember seeing a bunch
>> of other links on the previous webpage (and google) so I check out:
>
>
> So, did you try to install 32 bit version of the package or not? I mean,
> apart from your curiosity - was there any other reason for you to explore
> the source compilation route?
>
> --
> Best regards,
>  Vladimir S. Fonov ~ vladimir.fonov <at> gmail.com
>
>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users

From vladimir.fonov at gmail.com  Mon Jul  2 23:20:36 2012
From: vladimir.fonov at gmail.com (Vladimir S. Fonov)
Date: Mon, 02 Jul 2012 23:20:36 -0400
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
	<4FF19819.2070607@gmail.com>
	<CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
Message-ID: <4FF26504.9080409@gmail.com>

Hello,

On 12-07-02 10:04 PM, Mishkin Derakhshan wrote:
> I don't want to turn this into a "please help me with my installation"
> thread, but here is the summary of my results. The good news is that 2
> out of the 3 methods I tried worked! In true minc fashion there is
> more than one way to get 'er done.
> That being said, I still think it is frustrating for an end user to
> figure out which of the three they should have used.
>
> 1. http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit
> using the 32-bit version of the toolkit, and the testsuite and the models.
> I did have to run sudo apt-get install -f afterwards to install the
> missing library (libjpeg or something like that), but all appears to
> have installed okay. mincbeast in the pipeline does fail, but
> register, Display, mincstats etc. seem to work fine.
> ERROR! Allocation failure. Probably out of memory in
> /home/vfonov/src/minc-toolkit/BEaST/alloc.c:186.
> Died at /opt/minc/pipeline/pipeline_mritotal.pl line 182.
> (Odd that /home/vfonov is hardcoded into the error msg?)

yes, this error is noted in the 
http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit

Caveat Emptor: Unfortunately, currently mincbeast is unable to run in 
32bit mode properly, so standard_pipelene.pl will fail (and the 
test-suite will also fail)


So, if you would like to eliminate it, you can try following:

solution a: optimizing mincbeast ( https://github.com/BIC-MNI/BEaST )
so it fits in a 2Gb memory limit of 32 bit OS

solution b: install 64 bit version of Ubuntu.


> 3. https://github.com/BIC-MNI/minc-toolkit
> I tried compiling from source from git-hub using cmake. I ran into
> these problems:
> (i) ssh-keys for github.
> cd ../Register
>    git submodule init
>    git submodule update --recursive
> fatal: The remote end hung up unexpectedly
> Solution: https://help.github.com/articles/generating-ssh-keys

- looks like a broken url in the git submodule, so something didn't 
check out from git-hub prooperly, hence you can't compile some of the 
parts (and the rest of error messages).

-- 
Best regards,
  Vladimir S. Fonov ~ vladimir.fonov <at> gmail.com



From steve at sumost.ca  Mon Jul  2 23:24:27 2012
From: steve at sumost.ca (Steve M. Robbins)
Date: Mon, 2 Jul 2012 22:24:27 -0500
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
	<4FF19819.2070607@gmail.com>
	<CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
Message-ID: <20120703032427.GA1081@sumost.ca>

On Mon, Jul 01, 2012 at 10:04:04PM -0400, Mishkin Derakhshan wrote:

> So I needed to install minc on a fresh ubuntu 12.04 32-bit machine at
> home and I thought I would approach the issue as any other potential
> minc user would by googling "minc".

Since you are working with ubuntu, I'm curious why you did not
simply "apt-get install minc-tools"?

Thanks,
-Steve

From ayman.oweida at mail.mcgill.ca  Tue Jul  3 10:49:02 2012
From: ayman.oweida at mail.mcgill.ca (Ayman Oweida)
Date: Tue, 3 Jul 2012 14:49:02 +0000
Subject: [MINC-users] installing mni_autoreg
In-Reply-To: <CAG0CB22_NxeAj0seuk1Nw9bAHRGNmm=yzSqmYJsEM3cy0-XSHg@mail.gmail.com>
References: <FC413AEB-A273-4B10-BBBE-C86314410EFD@phenogenomics.ca>
	<36FAE925-5BFD-48B6-8173-636FE5A71124@phenogenomics.ca>
	<790582915299864F9DC262BDDC538DF51626E4@EXMBX2010-2.campus.MCGILL.CA>
	<CAG0CB20X+VapLq-gccr=QdGxO65gcjz2mkJYfOavZYbyJPXx8g@mail.gmail.com>
	<790582915299864F9DC262BDDC538DF51699B5@EXMBX2010-1.campus.MCGILL.CA>
	<CAG0CB22_NxeAj0seuk1Nw9bAHRGNmm=yzSqmYJsEM3cy0-XSHg@mail.gmail.com>
Message-ID: <790582915299864F9DC262BDDC538DF51741FA@EXMBX2010-1.campus.MCGILL.CA>


Just for the record:
cpan getopt::tabular did the job.  

I didn't have to install getopt::Tabular on my own.

Thanks for the help.

-Ayman

On 2012-06-24, at 6:52 AM, Andrew Janke wrote:

>> So, after running the commands below, I get the following:
>> 
>> Ayman-Oweidas-MacBook-Pro:mni_perllib-0.08 Ayman$ sudo make install
>> Password:
>> Manifying blib/man3/MNI::DataDir.3pm
>> Not a HASH reference at /System/Library/Perl/5.10.0/ExtUtils/Install.pm line 557.
>> make: *** [pure_site_install] Error 2
> 
> I am guessing that by this stage everything up until this has
> completed OK? I am guessing you are using the inbuilt perl in OSX? On
> older versions of OSX I found this problematic and used the macports
> or fink version but am not sure if this is still the case.
> 
> There are others on here who use OSX and MINC so I am hoping they will
> chime in here.
> 
> 
> a
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users



From zijdenbos at gmail.com  Wed Jul  4 12:19:53 2012
From: zijdenbos at gmail.com (Alex Zijdenbos)
Date: Wed, 4 Jul 2012 12:19:53 -0400
Subject: [MINC-users] xfmconcat -clobber bug
Message-ID: <CABEEk17VSptKUw-yNv5SUnuHpvrrU1e63VzKHuYOPopnf2DUwA@mail.gmail.com>

Hi all,

I just stumbled across a (actually two) problems with xfmconcat:

1) When run with -clobber, if the grid*.mnc files referenced in the
xfm are symlinks, it will try to clobber the files that are referenced
by the symlink, not the symlinks themselves, and

2) When it cannot overwrite the referent grid files (e.g., because of
write protection or ownership), it will throw a warning, but still
returns a 0 exit status.

So, when concat_grid_0.xfm is a symlink to a non-writeable file,
xfmconcat nominally succeeds:

$ ls -al concat_grid_0.mnc
0 lrwxrwxrwx 1 alex alex 13 2012-07-04 12:13 concat_grid_0.mnc -> /usr/bin/last*
$ xfmconcat -clobber a.xfm b.xfm concat.xfm
(from micreate): Unable to create file 'concat_grid_0.mnc'
Error: opening MINC file "concat_grid_0.mnc".
$ echo $?
0

Obviously, the resulting xfm (+grids) will not be what the user expected.

I think that at minimum, xfmconcat should fail when it can't write a
file; and I would also say that when clobbering, it should clobber a
symlink, not the referent file (but that latter point is up for debate
perhaps, which is actually why I raised it here).

-- A

From vladimir.fonov at gmail.com  Wed Jul  4 19:24:39 2012
From: vladimir.fonov at gmail.com (Vladimir S. FONOV)
Date: Wed, 04 Jul 2012 19:24:39 -0400
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
	<4FF19819.2070607@gmail.com>
	<CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
Message-ID: <4FF4D0B7.5030701@gmail.com>

Hello Everybody,

I have updated version of the binary minc-tools packages on 
http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit

now I build packages for 64 and 32 bit for Ubuntu 10.04 and 12.04, 64 
bit package for Debian 6 and 64 bit package for MacOS X 10.7.
Unfortunately, bic pipeline will still fail in 32 bit mode due to memory 
usage of mincbeast.

Also, I have replaced N3 with N4 (from ITK) so the results of the 
pipeline run will be different, also test suite will most likely produce 
different results, so it will fail.

It would be great if people could make test run on different machines 
and submit results.

-- 
Best regards,

Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com



From pierre.bellec at criugm.qc.ca  Wed Jul  4 20:39:33 2012
From: pierre.bellec at criugm.qc.ca (Pierre Bellec)
Date: Wed, 4 Jul 2012 20:39:33 -0400
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <4FF4D0B7.5030701@gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
	<4FF19819.2070607@gmail.com>
	<CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
	<4FF4D0B7.5030701@gmail.com>
Message-ID: <CADPHwxAVc4MnCkyVJVUvJC0qbBsyN5ifNV58Jx_gcsQ3Z-pVBw@mail.gmail.com>

Fantastic work Vladimir. I will definitely try running analysis with your
new packages and keep you posted. Thanks for all of your efforts to make
MINC easier to install. I will update the instructions on the MINC wiki
book.

Best,

Pierre Bellec, PhD
Research Centre of the Montreal Geriatric Institute
& Department of Computer Science and Operations Research
University of Montreal, Qu?bec, Canada
http://simexp-lab.org/brainwiki/doku.php?id=pierrebellec




2012/7/4 Vladimir S. FONOV <vladimir.fonov at gmail.com>

> Hello Everybody,
>
> I have updated version of the binary minc-tools packages on
> http://www.bic.mni.mcgill.ca/**ServicesSoftware/**
> ServicesSoftwareMincToolKit<http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit>
>
> now I build packages for 64 and 32 bit for Ubuntu 10.04 and 12.04, 64 bit
> package for Debian 6 and 64 bit package for MacOS X 10.7.
> Unfortunately, bic pipeline will still fail in 32 bit mode due to memory
> usage of mincbeast.
>
> Also, I have replaced N3 with N4 (from ITK) so the results of the pipeline
> run will be different, also test suite will most likely produce different
> results, so it will fail.
>
> It would be great if people could make test run on different machines and
> submit results.
>
> --
> Best regards,
>
> Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
>
>
> ______________________________**_________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
>

From a.janke at gmail.com  Thu Jul  5 07:56:22 2012
From: a.janke at gmail.com (Andrew Janke)
Date: Thu, 5 Jul 2012 21:56:22 +1000
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <CADPHwxAVc4MnCkyVJVUvJC0qbBsyN5ifNV58Jx_gcsQ3Z-pVBw@mail.gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
	<4FF19819.2070607@gmail.com>
	<CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
	<4FF4D0B7.5030701@gmail.com>
	<CADPHwxAVc4MnCkyVJVUvJC0qbBsyN5ifNV58Jx_gcsQ3Z-pVBw@mail.gmail.com>
Message-ID: <CAG0CB22M4tiDbN_3R3kyWTOkCav5uD-U5Mt5M0B5ytHPLFxJUA@mail.gmail.com>

Hi all,

First let me say I have nothing but thanks for the work that Vlad is
putting into the binary builds. I am only too aware of the amount of
work that this takes.

So, that said. I think we need to set some sort of process/rules on
how such releases are made. Especially if newer releases of this
depart from old behaviour, I for one am very much against hidden
changes in analysis behaviour (N3 -> N4) that will change peoples
results.

So, the decision needs to be made on which and where the official
packages are. I do have mincbundle (the old style of ubuntu/debian)
packages but have been loathe to release them if only to reduce
confusion. In the past the packages I have built have always been
focused entirely on preserving behaviour (from a data analysis point
of view).

Still I see this going one of a few ways, currently the situation is like this:

1) The latest MINC development is always on github

2) Stable releases are on packages.bic.mni.mcgill.ca/tgz

3) Vlad has done an enormous amount of work building us all nice packages here:

   http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit

I'm open to suggestions as to how we manages this but for now I'll
continue to apply patches to github. We are due for the next release
of MINC (2.2) but I still have a few more manual pages to write before
this happens. At this point there will be a new (source) release on
packages.bic.mni.mcgill.ca/tgz.

So, Pierre by all means update the MINC wikibook but perhaps we need
to make a distinction between what people can expect from binary
builds and "official" source packages.

Thanks all!


a


On 5 July 2012 10:39, Pierre Bellec <pierre.bellec at criugm.qc.ca> wrote:
> Fantastic work Vladimir. I will definitely try running analysis with your
> new packages and keep you posted. Thanks for all of your efforts to make
> MINC easier to install. I will update the instructions on the MINC wiki
> book.

From vladimir.fonov at gmail.com  Thu Jul  5 08:14:46 2012
From: vladimir.fonov at gmail.com (Vladimir S. FONOV)
Date: Thu, 05 Jul 2012 08:14:46 -0400
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <CAG0CB22M4tiDbN_3R3kyWTOkCav5uD-U5Mt5M0B5ytHPLFxJUA@mail.gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
	<4FF19819.2070607@gmail.com>
	<CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
	<4FF4D0B7.5030701@gmail.com>
	<CADPHwxAVc4MnCkyVJVUvJC0qbBsyN5ifNV58Jx_gcsQ3Z-pVBw@mail.gmail.com>
	<CAG0CB22M4tiDbN_3R3kyWTOkCav5uD-U5Mt5M0B5ytHPLFxJUA@mail.gmail.com>
Message-ID: <4FF58536.8030400@gmail.com>

Hello,


On 7/5/2012 7:56 AM, Andrew Janke wrote:
> So, that said. I think we need to set some sort of process/rules on
> how such releases are made. Especially if newer releases of this
> depart from old behaviour, I for one am very much against hidden
> changes in analysis behaviour (N3 ->  N4) that will change peoples
> results.

The N3 is still part of minc-toolkit package, but now you also have N4 
as part of Conver3D binary and a tool from ANTS. What I have really 
changed is the standard_pipeline.pl script used in the test suite.

The reason for this change is that N3 does not produce consistent 
results on different platforms (i.e Debian 6 vs Ubuntu 11.04), so 
analysis behaviour is actually not the same if you use N3 compiled with 
different compilers.

-- 
Best regards,

  Vladimir.FONOV <at> gmail <dot> com

From francois.hebert007 at gmail.com  Thu Jul  5 15:27:29 2012
From: francois.hebert007 at gmail.com (francois hebert)
Date: Thu, 5 Jul 2012 15:27:29 -0400
Subject: [MINC-users] Maximum intensity projection
In-Reply-To: <CANcjydbofwHBoh-GTNJbC7-S-cNABvdiggWB5Fww8qCA1BkK-Q@mail.gmail.com>
References: <CACGPbDzNPo0EOwNSKGYt7D0BtuF_cPOMG8eNFC03FNK-Ap_M5Q@mail.gmail.com>
	<CAG0CB21QK15eeR_bejhjWPOKmAM7VUaiCz_QCsbqf1sV=ta=Tw@mail.gmail.com>
	<CANcjydbofwHBoh-GTNJbC7-S-cNABvdiggWB5Fww8qCA1BkK-Q@mail.gmail.com>
Message-ID: <CACGPbDzJGKnJC9AYyzdTyZTdzoKTeAYKGPJbzj=cywGyvnf+wA@mail.gmail.com>

Thanks a lot for all these tricks.

Francois Hebert



2012/6/15 Peter Neelin <peter.neelin at gmail.com>

> On Jun 13, 2012 10:02 PM, "Andrew Janke" <a.janke at gmail.com> wrote:
>
> > On 9 June 2012 01:36, francois hebert <francois.hebert007 at gmail.com>
> wrote:
> > > I would like to do a Maximum Intensity Projection in each orthogonal
> axis.
> > > Is there such tool in the minc library?
> >
> > I know I'm late to party but for the dinosaurs amongst us who don't
> > always play with the new toys you could do this using the basic MINC
> > commands like this (via a perl script):
> > [fancy perl script deleted]
> > &do_cmd('mincmath', '-clobber', '-max', @files, @ARGV[1]);
> >
> > I will admit that doing this in pyminc is whole lot easier!
>
> Wouldn't a -dimension option to mincmath do the same thing but in one line?
> I haven't tested it, so I'm not sure if it will allow accumulating the
> maximum over image dimensions (the original question was for all 3
> orthogonal projections). If it does not, then a mincreshape to re-order the
> dimensions and a mincmath -dimension should do the trick for the other 2
> dimensions.
>
> Peter
> --
> Peter Neelin
> peter.neelin at gmail.com
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From francois.hebert007 at gmail.com  Thu Jul  5 15:38:12 2012
From: francois.hebert007 at gmail.com (francois hebert)
Date: Thu, 5 Jul 2012 15:38:12 -0400
Subject: [MINC-users] Getting max and min in a sphere centered at each voxel
Message-ID: <CACGPbDwmDGtg=_WOtuiCV7ko6nVF8D0VL_YB=eUc82PZbLkqJw@mail.gmail.com>

Hi All,

I would like to know if there is a minc tool that allow to get the max and
min inside a sphere of a predefined diameter at each voxel.


Regards,

Francois Hebert

From vladimir.fonov at gmail.com  Thu Jul  5 15:47:00 2012
From: vladimir.fonov at gmail.com (Vladimir S. FONOV)
Date: Thu, 05 Jul 2012 15:47:00 -0400
Subject: [MINC-users] Getting max and min in a sphere centered at each
 voxel
In-Reply-To: <CACGPbDwmDGtg=_WOtuiCV7ko6nVF8D0VL_YB=eUc82PZbLkqJw@mail.gmail.com>
References: <CACGPbDwmDGtg=_WOtuiCV7ko6nVF8D0VL_YB=eUc82PZbLkqJw@mail.gmail.com>
Message-ID: <4FF5EF34.7040203@gmail.com>

Hello,

On 12-07-05 03:38 PM, francois hebert wrote:
> I would like to know if there is a minc tool that allow to get the max and
> min inside a sphere of a predefined diameter at each voxel.

i would think make_phantom to create a spherical mask followed by 
mincstats -min -max

-- 
Best regards,

Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com



From lconcha at gmail.com  Thu Jul  5 15:56:26 2012
From: lconcha at gmail.com (Luis Concha)
Date: Thu, 5 Jul 2012 14:56:26 -0500
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <4FF58536.8030400@gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
	<4FF19819.2070607@gmail.com>
	<CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
	<4FF4D0B7.5030701@gmail.com>
	<CADPHwxAVc4MnCkyVJVUvJC0qbBsyN5ifNV58Jx_gcsQ3Z-pVBw@mail.gmail.com>
	<CAG0CB22M4tiDbN_3R3kyWTOkCav5uD-U5Mt5M0B5ytHPLFxJUA@mail.gmail.com>
	<4FF58536.8030400@gmail.com>
Message-ID: <CAOq9ZoFyKMJG+5Q8SOwbJ6-Qy7Bg6wZ5-Jp3yp3t6FTj+Bc8tQ@mail.gmail.com>

Thank you, Vlad and Andrew, for updating the binaries. I will also test
them and come back with results.

I second the opinion of specifying which binaries and/or sources are
stable, unstable, or official releases, as I have myself been confused by
that in the past.

Cheers.

From vladimir.fonov at gmail.com  Thu Jul  5 16:06:14 2012
From: vladimir.fonov at gmail.com (Vladimir S. FONOV)
Date: Thu, 05 Jul 2012 16:06:14 -0400
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <CAOq9ZoFyKMJG+5Q8SOwbJ6-Qy7Bg6wZ5-Jp3yp3t6FTj+Bc8tQ@mail.gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
	<4FF19819.2070607@gmail.com>
	<CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
	<4FF4D0B7.5030701@gmail.com>
	<CADPHwxAVc4MnCkyVJVUvJC0qbBsyN5ifNV58Jx_gcsQ3Z-pVBw@mail.gmail.com>
	<CAG0CB22M4tiDbN_3R3kyWTOkCav5uD-U5Mt5M0B5ytHPLFxJUA@mail.gmail.com>
	<4FF58536.8030400@gmail.com>
	<CAOq9ZoFyKMJG+5Q8SOwbJ6-Qy7Bg6wZ5-Jp3yp3t6FTj+Bc8tQ@mail.gmail.com>
Message-ID: <4FF5F3B6.3040501@gmail.com>

Hello,


I think, at the moment minc-toolkit is not quite stable.

On 12-07-05 03:56 PM, Luis Concha wrote:
> Thank you, Vlad and Andrew, for updating the binaries. I will also test
> them and come back with results.
>
> I second the opinion of specifying which binaries and/or sources are
> stable, unstable, or official releases, as I have myself been confused by
> that in the past.


-- 
Best regards,

Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com



From francois.hebert007 at gmail.com  Thu Jul  5 16:39:27 2012
From: francois.hebert007 at gmail.com (francois hebert)
Date: Thu, 5 Jul 2012 16:39:27 -0400
Subject: [MINC-users] Getting max and min in a sphere centered at each
	voxel
In-Reply-To: <4FF5EF34.7040203@gmail.com>
References: <CACGPbDwmDGtg=_WOtuiCV7ko6nVF8D0VL_YB=eUc82PZbLkqJw@mail.gmail.com>
	<4FF5EF34.7040203@gmail.com>
Message-ID: <CACGPbDwj8HwsxHGY+EYfXfm19uSmExHFYP5_EBZM2csacxbQVg@mail.gmail.com>

Hi,

you suggestion seems good for one voxel but I would like a filter that will
go through all the voxels.

Francois Hebert

2012/7/5 Vladimir S. FONOV <vladimir.fonov at gmail.com>

> Hello,
>
>
> On 12-07-05 03:38 PM, francois hebert wrote:
>
>> I would like to know if there is a minc tool that allow to get the max and
>> min inside a sphere of a predefined diameter at each voxel.
>>
>
> i would think make_phantom to create a spherical mask followed by
> mincstats -min -max
>
> --
> Best regards,
>
> Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
>
>
> ______________________________**_________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
>

From vladimir.fonov at gmail.com  Thu Jul  5 17:49:13 2012
From: vladimir.fonov at gmail.com (Vladimir S. FONOV)
Date: Thu, 05 Jul 2012 17:49:13 -0400
Subject: [MINC-users] Getting max and min in a sphere centered at each
 voxel
In-Reply-To: <CACGPbDwj8HwsxHGY+EYfXfm19uSmExHFYP5_EBZM2csacxbQVg@mail.gmail.com>
References: <CACGPbDwmDGtg=_WOtuiCV7ko6nVF8D0VL_YB=eUc82PZbLkqJw@mail.gmail.com>
	<4FF5EF34.7040203@gmail.com>
	<CACGPbDwj8HwsxHGY+EYfXfm19uSmExHFYP5_EBZM2csacxbQVg@mail.gmail.com>
Message-ID: <4FF60BD9.1040802@gmail.com>

Then you would need to use gray-scale morphology and perform dilation to 
get max and erosion for min. I have a tool based on ITK which does that .

On 12-07-05 04:39 PM, francois hebert wrote:
> Hi,
>
> you suggestion seems good for one voxel but I would like a filter that will
> go through all the voxels.
>
> Francois Hebert
>
> 2012/7/5 Vladimir S. FONOV <vladimir.fonov at gmail.com>
>
>> Hello,
>>
>>
>> On 12-07-05 03:38 PM, francois hebert wrote:
>>
>>> I would like to know if there is a minc tool that allow to get the max and
>>> min inside a sphere of a predefined diameter at each voxel.
>>>
>>
>> i would think make_phantom to create a spherical mask followed by
>> mincstats -min -max
>>
>> --
>> Best regards,
>>
>> Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
>>
>>
>> ______________________________**_________________
>> MINC-users at bic.mni.mcgill.ca
>> http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
>>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>


-- 
Best regards,

Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com



From a.janke at gmail.com  Thu Jul  5 17:53:36 2012
From: a.janke at gmail.com (Andrew Janke)
Date: Fri, 6 Jul 2012 07:53:36 +1000
Subject: [MINC-users] Getting max and min in a sphere centered at each
	voxel
In-Reply-To: <4FF60BD9.1040802@gmail.com>
References: <CACGPbDwmDGtg=_WOtuiCV7ko6nVF8D0VL_YB=eUc82PZbLkqJw@mail.gmail.com>
	<4FF5EF34.7040203@gmail.com>
	<CACGPbDwj8HwsxHGY+EYfXfm19uSmExHFYP5_EBZM2csacxbQVg@mail.gmail.com>
	<4FF60BD9.1040802@gmail.com>
Message-ID: <CAG0CB230cmieB3nAwLnrjLUFD_HxHaX7y3AC3bB56UngotTVMA@mail.gmail.com>

mincmorph will also do this but only behaves well for small
structuring elements (up to 5x5). Beyond that it's a tad slow.

You should be able to get a rough hack at what you want with something
like this for a 3x3x3 structuring element to start with:

   mincmorph -3D26 -erode in.mnc max.mnc
   mincmorph -3D26 -dilate in.mnc min.mnc

This should then get what you want across the whole volume in min.mnc
and max.mnc.


a

On 6 July 2012 07:49, Vladimir S. FONOV <vladimir.fonov at gmail.com> wrote:
> Then you would need to use gray-scale morphology and perform dilation to get
> max and erosion for min. I have a tool based on ITK which does that .
>
>
> On 12-07-05 04:39 PM, francois hebert wrote:
>>
>> Hi,
>>
>> you suggestion seems good for one voxel but I would like a filter that
>> will
>> go through all the voxels.
>>
>> Francois Hebert
>>
>> 2012/7/5 Vladimir S. FONOV <vladimir.fonov at gmail.com>
>>
>>> Hello,
>>>
>>>
>>> On 12-07-05 03:38 PM, francois hebert wrote:
>>>
>>>> I would like to know if there is a minc tool that allow to get the max
>>>> and
>>>> min inside a sphere of a predefined diameter at each voxel.
>>>>
>>>
>>> i would think make_phantom to create a spherical mask followed by
>>> mincstats -min -max
>>>
>>> --
>>> Best regards,
>>>
>>> Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
>>>
>>>
>>> ______________________________**_________________
>>> MINC-users at bic.mni.mcgill.ca
>>>
>>> http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
>>>
>> _______________________________________________
>> MINC-users at bic.mni.mcgill.ca
>> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>>
>
>
> --
> Best regards,
>
> Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
>
>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users

From zijdenbos at gmail.com  Thu Jul  5 20:10:25 2012
From: zijdenbos at gmail.com (Alex Zijdenbos)
Date: Thu, 5 Jul 2012 20:10:25 -0400
Subject: [MINC-users] Getting max and min in a sphere centered at each
 voxel
In-Reply-To: <CAG0CB230cmieB3nAwLnrjLUFD_HxHaX7y3AC3bB56UngotTVMA@mail.gmail.com>
References: <CACGPbDwmDGtg=_WOtuiCV7ko6nVF8D0VL_YB=eUc82PZbLkqJw@mail.gmail.com>
	<4FF5EF34.7040203@gmail.com>
	<CACGPbDwj8HwsxHGY+EYfXfm19uSmExHFYP5_EBZM2csacxbQVg@mail.gmail.com>
	<4FF60BD9.1040802@gmail.com>
	<CAG0CB230cmieB3nAwLnrjLUFD_HxHaX7y3AC3bB56UngotTVMA@mail.gmail.com>
Message-ID: <5F68039A26D441D4A82C01FDED75029C@gmail.com>

Ah - so mincmorph actually does grey level morphology? Somehow I've only ever used it for binary morphology. Nice.  

However, I do think you meant:

mincmorph -3D26 -erosion in.mnc min.mnc
mincmorph -3D26 -dilation in.mnc max.mnc

Then with this perl script:

 ?make_mincmorph_kernel.pl (http://cl.ly/2420323b382t2z3p3E3N)

you can generate spherical kernels of arbitrary size; but yes it will get a bit slow for large ones. I have generated 10x10x10 kernels though and that is all still very doable (you couldn't finish your coffee in the time it takes to run).

-- A

On Thursday, 5 July, 2012 at 5:53 PM, Andrew Janke wrote:  
> mincmorph will also do this but only behaves well for small
> structuring elements (up to 5x5). Beyond that it's a tad slow.
>  
> You should be able to get a rough hack at what you want with something
> like this for a 3x3x3 structuring element to start with:
>  
> mincmorph -3D26 -erode in.mnc max.mnc
> mincmorph -3D26 -dilate in.mnc min.mnc
>  
> This should then get what you want across the whole volume in min.mnc
> and max.mnc.
>  
>  
> a
>  
> On 6 July 2012 07:49, Vladimir S. FONOV <vladimir.fonov at gmail.com (mailto:vladimir.fonov at gmail.com)> wrote:
> > Then you would need to use gray-scale morphology and perform dilation to get
> > max and erosion for min. I have a tool based on ITK which does that .
> >  
> >  
> > On 12-07-05 04:39 PM, francois hebert wrote:
> > >  
> > > Hi,
> > >  
> > > you suggestion seems good for one voxel but I would like a filter that
> > > will
> > > go through all the voxels.
> > >  
> > > Francois Hebert
> > >  
> > > 2012/7/5 Vladimir S. FONOV <vladimir.fonov at gmail.com (mailto:vladimir.fonov at gmail.com)>
> > >  
> > > > Hello,
> > > >  
> > > >  
> > > > On 12-07-05 03:38 PM, francois hebert wrote:
> > > >  
> > > > > I would like to know if there is a minc tool that allow to get the max
> > > > > and
> > > > > min inside a sphere of a predefined diameter at each voxel.
> > > > >  
> > > >  
> > > >  
> > > > i would think make_phantom to create a spherical mask followed by
> > > > mincstats -min -max
> > > >  
> > > > --
> > > > Best regards,
> > > >  
> > > > Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
> > > >  
> > > >  
> > > > ______________________________**_________________
> > > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > >  
> > > > http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
> > > _______________________________________________
> > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > >  
> >  
> >  
> >  
> > --
> > Best regards,
> >  
> > Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
> >  
> >  
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> >  
>  
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>  
>  



From a.janke at gmail.com  Fri Jul  6 02:43:47 2012
From: a.janke at gmail.com (Andrew Janke)
Date: Fri, 6 Jul 2012 16:43:47 +1000
Subject: [MINC-users] xfmconcat -clobber bug
In-Reply-To: <CABEEk17VSptKUw-yNv5SUnuHpvrrU1e63VzKHuYOPopnf2DUwA@mail.gmail.com>
References: <CABEEk17VSptKUw-yNv5SUnuHpvrrU1e63VzKHuYOPopnf2DUwA@mail.gmail.com>
Message-ID: <CAG0CB23jv1iAZp-gujOOA6hPz+y4utmQ98BsJBju-0aCqACjyQ@mail.gmail.com>

Hi Alex,

This really isn't a bug in xfmconcat but rather a bug in volume_io
somewhere around here:

   https://github.com/BIC-MNI/minc/blob/master/volume_io/MNI_formats/gen_xf_io.c#L187

The question is though given that it's going to be a pain to get the
value of the clobber flag into this function what is the desired
behaviour. My thoughts would be that if the code has got to writing a
transform it should simply write the transform irrespective of whether
the target is a link or a file.

So, do we introduce a test to see if the output grid file exists
(already) and if it does, remove it? This should get around all
combinations of this that I can think of.


a

On 5 July 2012 02:19, Alex Zijdenbos <zijdenbos at gmail.com> wrote:
> Hi all,
>
> I just stumbled across a (actually two) problems with xfmconcat:
>
> 1) When run with -clobber, if the grid*.mnc files referenced in the
> xfm are symlinks, it will try to clobber the files that are referenced
> by the symlink, not the symlinks themselves, and
>
> 2) When it cannot overwrite the referent grid files (e.g., because of
> write protection or ownership), it will throw a warning, but still
> returns a 0 exit status.
>
> So, when concat_grid_0.xfm is a symlink to a non-writeable file,
> xfmconcat nominally succeeds:
>
> $ ls -al concat_grid_0.mnc
> 0 lrwxrwxrwx 1 alex alex 13 2012-07-04 12:13 concat_grid_0.mnc -> /usr/bin/last*
> $ xfmconcat -clobber a.xfm b.xfm concat.xfm
> (from micreate): Unable to create file 'concat_grid_0.mnc'
> Error: opening MINC file "concat_grid_0.mnc".
> $ echo $?
> 0
>
> Obviously, the resulting xfm (+grids) will not be what the user expected.
>
> I think that at minimum, xfmconcat should fail when it can't write a
> file; and I would also say that when clobbering, it should clobber a
> symlink, not the referent file (but that latter point is up for debate
> perhaps, which is actually why I raised it here).
>
> -- A
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users

From francois.hebert007 at gmail.com  Fri Jul  6 10:24:50 2012
From: francois.hebert007 at gmail.com (francois hebert)
Date: Fri, 6 Jul 2012 10:24:50 -0400
Subject: [MINC-users] Getting max and min in a sphere centered at each
	voxel
In-Reply-To: <5F68039A26D441D4A82C01FDED75029C@gmail.com>
References: <CACGPbDwmDGtg=_WOtuiCV7ko6nVF8D0VL_YB=eUc82PZbLkqJw@mail.gmail.com>
	<4FF5EF34.7040203@gmail.com>
	<CACGPbDwj8HwsxHGY+EYfXfm19uSmExHFYP5_EBZM2csacxbQVg@mail.gmail.com>
	<4FF60BD9.1040802@gmail.com>
	<CAG0CB230cmieB3nAwLnrjLUFD_HxHaX7y3AC3bB56UngotTVMA@mail.gmail.com>
	<5F68039A26D441D4A82C01FDED75029C@gmail.com>
Message-ID: <CACGPbDzHOa7UMGij6uwO2+Lz5c1NNCkHNtyxzgcy0r37=kXODA@mail.gmail.com>

Thanks all for your advice.

Francois Hebert



2012/7/5 Alex Zijdenbos <zijdenbos at gmail.com>

> Ah - so mincmorph actually does grey level morphology? Somehow I've only
> ever used it for binary morphology. Nice.
>
> However, I do think you meant:
>
> mincmorph -3D26 -erosion in.mnc min.mnc
> mincmorph -3D26 -dilation in.mnc max.mnc
>
> Then with this perl script:
>
>  make_mincmorph_kernel.pl (http://cl.ly/2420323b382t2z3p3E3N)
>
> you can generate spherical kernels of arbitrary size; but yes it will get
> a bit slow for large ones. I have generated 10x10x10 kernels though and
> that is all still very doable (you couldn't finish your coffee in the time
> it takes to run).
>
> -- A
>
> On Thursday, 5 July, 2012 at 5:53 PM, Andrew Janke wrote:
> > mincmorph will also do this but only behaves well for small
> > structuring elements (up to 5x5). Beyond that it's a tad slow.
> >
> > You should be able to get a rough hack at what you want with something
> > like this for a 3x3x3 structuring element to start with:
> >
> > mincmorph -3D26 -erode in.mnc max.mnc
> > mincmorph -3D26 -dilate in.mnc min.mnc
> >
> > This should then get what you want across the whole volume in min.mnc
> > and max.mnc.
> >
> >
> > a
> >
> > On 6 July 2012 07:49, Vladimir S. FONOV <vladimir.fonov at gmail.com(mailto:
> vladimir.fonov at gmail.com)> wrote:
> > > Then you would need to use gray-scale morphology and perform dilation
> to get
> > > max and erosion for min. I have a tool based on ITK which does that .
> > >
> > >
> > > On 12-07-05 04:39 PM, francois hebert wrote:
> > > >
> > > > Hi,
> > > >
> > > > you suggestion seems good for one voxel but I would like a filter
> that
> > > > will
> > > > go through all the voxels.
> > > >
> > > > Francois Hebert
> > > >
> > > > 2012/7/5 Vladimir S. FONOV <vladimir.fonov at gmail.com (mailto:
> vladimir.fonov at gmail.com)>
> > > >
> > > > > Hello,
> > > > >
> > > > >
> > > > > On 12-07-05 03:38 PM, francois hebert wrote:
> > > > >
> > > > > > I would like to know if there is a minc tool that allow to get
> the max
> > > > > > and
> > > > > > min inside a sphere of a predefined diameter at each voxel.
> > > > > >
> > > > >
> > > > >
> > > > > i would think make_phantom to create a spherical mask followed by
> > > > > mincstats -min -max
> > > > >
> > > > > --
> > > > > Best regards,
> > > > >
> > > > > Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
> > > > >
> > > > >
> > > > > ______________________________**_________________
> > > > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > > >
> > > > > http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
> > > > _______________________________________________
> > > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > > >
> > >
> > >
> > >
> > > --
> > > Best regards,
> > >
> > > Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
> > >
> > >
> > > _______________________________________________
> > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > >
> >
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> >
> >
>
>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From zijdenbos at gmail.com  Fri Jul  6 11:00:37 2012
From: zijdenbos at gmail.com (Alex Zijdenbos)
Date: Fri, 6 Jul 2012 11:00:37 -0400
Subject: [MINC-users] xfmconcat -clobber bug
In-Reply-To: <CAG0CB23jv1iAZp-gujOOA6hPz+y4utmQ98BsJBju-0aCqACjyQ@mail.gmail.com>
References: <CABEEk17VSptKUw-yNv5SUnuHpvrrU1e63VzKHuYOPopnf2DUwA@mail.gmail.com>
	<CAG0CB23jv1iAZp-gujOOA6hPz+y4utmQ98BsJBju-0aCqACjyQ@mail.gmail.com>
Message-ID: <CABEEk16RgABj34bX4kePAcU0-0jiRL0C=g7aDAR8VKGm0NjHcg@mail.gmail.com>

Hi Andrew,

I agree - as far as I can tell the simple thing to do here would be to
remove a pre-existing grid file or symlink before the new one is
written. The clobbering behaviour is (should be) already established
here (at least xfmconcat checks it right at the outset based on
existence of the output xfm).

But that still leaves the issue that the return value of output_volume
is currently ignored; if the write (or the removal) of the grid file
fails the main prog will still exit(EXIT_SUCCESS). I think
output_one_transform() should generate a meaningful return value that
is passed up through output_transform() (line 289/291).

-- A

On Fri, Jul 6, 2012 at 2:43 AM, Andrew Janke <a.janke at gmail.com> wrote:
> Hi Alex,
>
> This really isn't a bug in xfmconcat but rather a bug in volume_io
> somewhere around here:
>
>    https://github.com/BIC-MNI/minc/blob/master/volume_io/MNI_formats/gen_xf_io.c#L187
>
> The question is though given that it's going to be a pain to get the
> value of the clobber flag into this function what is the desired
> behaviour. My thoughts would be that if the code has got to writing a
> transform it should simply write the transform irrespective of whether
> the target is a link or a file.
>
> So, do we introduce a test to see if the output grid file exists
> (already) and if it does, remove it? This should get around all
> combinations of this that I can think of.
>
>
> a
>
> On 5 July 2012 02:19, Alex Zijdenbos <zijdenbos at gmail.com> wrote:
>> Hi all,
>>
>> I just stumbled across a (actually two) problems with xfmconcat:
>>
>> 1) When run with -clobber, if the grid*.mnc files referenced in the
>> xfm are symlinks, it will try to clobber the files that are referenced
>> by the symlink, not the symlinks themselves, and
>>
>> 2) When it cannot overwrite the referent grid files (e.g., because of
>> write protection or ownership), it will throw a warning, but still
>> returns a 0 exit status.
>>
>> So, when concat_grid_0.xfm is a symlink to a non-writeable file,
>> xfmconcat nominally succeeds:
>>
>> $ ls -al concat_grid_0.mnc
>> 0 lrwxrwxrwx 1 alex alex 13 2012-07-04 12:13 concat_grid_0.mnc -> /usr/bin/last*
>> $ xfmconcat -clobber a.xfm b.xfm concat.xfm
>> (from micreate): Unable to create file 'concat_grid_0.mnc'
>> Error: opening MINC file "concat_grid_0.mnc".
>> $ echo $?
>> 0
>>
>> Obviously, the resulting xfm (+grids) will not be what the user expected.
>>
>> I think that at minimum, xfmconcat should fail when it can't write a
>> file; and I would also say that when clobbering, it should clobber a
>> symlink, not the referent file (but that latter point is up for debate
>> perhaps, which is actually why I raised it here).
>>
>> -- A
>> _______________________________________________
>> MINC-users at bic.mni.mcgill.ca
>> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From zijdenbos at gmail.com  Fri Jul  6 23:18:57 2012
From: zijdenbos at gmail.com (Alex Zijdenbos)
Date: Fri, 6 Jul 2012 23:18:57 -0400
Subject: [MINC-users] Getting max and min in a sphere centered at each
 voxel
In-Reply-To: <5F68039A26D441D4A82C01FDED75029C@gmail.com>
References: <CACGPbDwmDGtg=_WOtuiCV7ko6nVF8D0VL_YB=eUc82PZbLkqJw@mail.gmail.com>
	<4FF5EF34.7040203@gmail.com>
	<CACGPbDwj8HwsxHGY+EYfXfm19uSmExHFYP5_EBZM2csacxbQVg@mail.gmail.com>
	<4FF60BD9.1040802@gmail.com>
	<CAG0CB230cmieB3nAwLnrjLUFD_HxHaX7y3AC3bB56UngotTVMA@mail.gmail.com>
	<5F68039A26D441D4A82C01FDED75029C@gmail.com>
Message-ID: <7AFC7FF86A08427D8604F4A5107CB477@gmail.com>

I realized there were some issues with the kernel generation script I passed along (center voxel nonzero, among other things), so here is an improved version:

make_mincmorph_kernel.pl (http://cl.ly/3P0i391X0k0e2V1T2846)

It now also allows for ellipsoidal kernels. As examples, the 4 standard mincmorph kernels can be generated as follows:

make_mincmorph_kernel.pl -radius 1 3D06.ker
make_mincmorph_kernel.pl -radius 1.8 3D26.ker
make_mincmorph_kernel.pl -radius3 1 1 0 2D04.ker
make_mincmorph_kernel.pl -radius3 1.8 1.8 0 2D08.ker

-- A

On Thursday, 5 July, 2012 at 8:10 PM, Alex Zijdenbos wrote:  
> Ah - so mincmorph actually does grey level morphology? Somehow I've only ever used it for binary morphology. Nice.  
>  
> However, I do think you meant:
>  
> mincmorph -3D26 -erosion in.mnc min.mnc
> mincmorph -3D26 -dilation in.mnc max.mnc
>  
> Then with this perl script:
>  
>  ?make_mincmorph_kernel.pl (http://cl.ly/2420323b382t2z3p3E3N)
>  
> you can generate spherical kernels of arbitrary size; but yes it will get a bit slow for large ones. I have generated 10x10x10 kernels though and that is all still very doable (you couldn't finish your coffee in the time it takes to run).
>  
> -- A
>  
> On Thursday, 5 July, 2012 at 5:53 PM, Andrew Janke wrote:
> > mincmorph will also do this but only behaves well for small
> > structuring elements (up to 5x5). Beyond that it's a tad slow.
> >  
> > You should be able to get a rough hack at what you want with something
> > like this for a 3x3x3 structuring element to start with:
> >  
> > mincmorph -3D26 -erode in.mnc max.mnc
> > mincmorph -3D26 -dilate in.mnc min.mnc
> >  
> > This should then get what you want across the whole volume in min.mnc
> > and max.mnc.
> >  
> >  
> > a
> >  
> > On 6 July 2012 07:49, Vladimir S. FONOV <vladimir.fonov at gmail.com (mailto:vladimir.fonov at gmail.com)> wrote:
> > > Then you would need to use gray-scale morphology and perform dilation to get
> > > max and erosion for min. I have a tool based on ITK which does that .
> > >  
> > >  
> > > On 12-07-05 04:39 PM, francois hebert wrote:
> > > >  
> > > > Hi,
> > > >  
> > > > you suggestion seems good for one voxel but I would like a filter that
> > > > will
> > > > go through all the voxels.
> > > >  
> > > > Francois Hebert
> > > >  
> > > > 2012/7/5 Vladimir S. FONOV <vladimir.fonov at gmail.com (mailto:vladimir.fonov at gmail.com)>
> > > >  
> > > > > Hello,
> > > > >  
> > > > >  
> > > > > On 12-07-05 03:38 PM, francois hebert wrote:
> > > > >  
> > > > > > I would like to know if there is a minc tool that allow to get the max
> > > > > > and
> > > > > > min inside a sphere of a predefined diameter at each voxel.
> > > > > >  
> > > > >  
> > > > >  
> > > > > i would think make_phantom to create a spherical mask followed by
> > > > > mincstats -min -max
> > > > >  
> > > > > --
> > > > > Best regards,
> > > > >  
> > > > > Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
> > > > >  
> > > > >  
> > > > > ______________________________**_________________
> > > > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > > >  
> > > > > http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
> > > > _______________________________________________
> > > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > > >  
> > >  
> > >  
> > >  
> > > --
> > > Best regards,
> > >  
> > > Vladimir S. FONOV ~ vladimir <dot> fonov <at> gmail <dot> com
> > >  
> > >  
> > > _______________________________________________
> > > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > >  
> >  
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca (mailto:MINC-users at bic.mni.mcgill.ca)
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> >  
> >  
> >  
>  
>  


From vladimir.fonov at gmail.com  Sun Jul  8 00:02:05 2012
From: vladimir.fonov at gmail.com (Vladimir S. Fonov)
Date: Sun, 08 Jul 2012 00:02:05 -0400
Subject: [MINC-users] Official Installation Instructions
In-Reply-To: <CAOq9ZoFyKMJG+5Q8SOwbJ6-Qy7Bg6wZ5-Jp3yp3t6FTj+Bc8tQ@mail.gmail.com>
References: <CAB-GmL0BjeNqn5YJwUtDa-WU_3o1rt+h1j1ZVXRETZ_hfauckw@mail.gmail.com>
	<4FF19819.2070607@gmail.com>
	<CAB-GmL0r1OpAkiBKcPjX6tniyY+guATavWo3P=3FnELvzEPNCw@mail.gmail.com>
	<4FF4D0B7.5030701@gmail.com>
	<CADPHwxAVc4MnCkyVJVUvJC0qbBsyN5ifNV58Jx_gcsQ3Z-pVBw@mail.gmail.com>
	<CAG0CB22M4tiDbN_3R3kyWTOkCav5uD-U5Mt5M0B5ytHPLFxJUA@mail.gmail.com>
	<4FF58536.8030400@gmail.com>
	<CAOq9ZoFyKMJG+5Q8SOwbJ6-Qy7Bg6wZ5-Jp3yp3t6FTj+Bc8tQ@mail.gmail.com>
Message-ID: <4FF9063D.9020000@gmail.com>

Hello Everybody,


I discovered that my 0.3.10 version of minc-toolkit had problem with 
MINC2 format support - it didn't properly handle files with internal 
compression. I fixed it in version 0.3.11 ( 
http://www.bic.mni.mcgill.ca/ServicesSoftware/ServicesSoftwareMincToolKit )

-- 
Best regards,
  Vladimir S. Fonov ~ vladimir.fonov <at> gmail.com



From mishkind at gmail.com  Sat Jul 14 23:13:34 2012
From: mishkind at gmail.com (Mishkin Derakhshan)
Date: Sat, 14 Jul 2012 23:13:34 -0400
Subject: [MINC-users] installing brain-view2
Message-ID: <CAB-GmL3ekes-zjriKCQZ=0AvD2qVcnELZNn8KOYsUC9se+ti0A@mail.gmail.com>

I had a hell of time installing brain-view2 so I documented my errors
and solutions/hacks in case others come across the same problems
(read: when I have to do this again I can look back at this post).

This was on a ubuntu 12.04 64 bit machine.

I followed the instruction here (thank you jason and sghanavati):
https://wiki.phenogenomics.ca/display/MICePub/brain-view

Pre-requisites
1. MINC. I used the minc-toolkit binaries and only ran into one minor problem.
opus[~]$ sudo apt-get install libc6 libstdc++6 imagemagick perl
freeglut3 libgl1 libxcb1 libxdmcp6 libx11-6 libxext6 libxau6 libuuid1
libjpeg62 libexpat1 libtiff4
Reading package lists... Done
Building dependency tree
Reading state information... Done
Package libgl1 is a virtual package provided by:
  libgl1-mesa-swx11 8.0.2-0ubuntu3.1
  libgl1-mesa-glx 8.0.2-0ubuntu3.1
You should explicitly select one to install.

E: Package 'libgl1' has no installation candidate

I chose libgl1-mesa-glx at random, and then everything else worked.

2. Qt4: sudo apt-get install qt4-qmake libqt4-dev libqt4-core
3. Boost: sudo apt-get install libboost1.48
4.Coin:
wget http://ftp.coin3d.org/coin/src/all/Coin-3.1.3.tar.gz
tar xvzf Coin-3.1.3.tar.gz
mkdir coin-build
cd coin-build/
../Coin-3.1.3/configure
make
sudo make install

5. Quarter:
wget http://ftp.coin3d.org/coin/src/all/Quarter-1.0.0.tar.gz
tar xvzf Quarter-1.0.0.tar.gz
mkdir quarter-build
cd quarter-build/
../Quarter-1.0.0/configure
make

ERROR:../../../Quarter-1.0.0/src/Quarter/Quarter.cpp:148:13: error:
'stderr' was not declared in this scope
SOLUTION:Added
#include "stdio.h"
to this file: ../Quarter-1.0.0/src/Quarter/Quarter.cpp

make
sudo make install

6. HDF5 1.8.9:
Downloaded a build of HDF5 1.8.9:
http://www.hdfgroup.org/HDF5/release/obtain5.html
tar xvzf hdf5-1.8.9-linux-x86_64-static.tar.gz

7. BicInventor:
wget http://packages.bic.mni.mcgill.ca/tgz/bicInventor-0.3.1.tar.gz
cd bicInventor-0.3.1
./configure --prefix=/opt/minc --with-build-path=/opt/minc
--includedir=/opt/minc/include --libdir=/opt/minc/lib --with-minc2
ERROR: (lots of erros when I tried to configure)
SOLUTION: sudo apt-get install build-essential realpath zlib1g-dev
bison flex libx11-dev libxmu-dev byacc libsoqt4-dev automake libtool
libdbi- perl libblas-dev liblapack-dev libblitz0-dev libreadline-dev
bzr bzrtools octave3.2 imagemagick

./configure --prefix=/opt/minc --with-build-path=/opt/minc
--includedir=/opt/minc/include --libdir=/opt/minc/lib --with-minc2
make
sudo make install

ACTUALLY INSTALLING BRAIN-VIEW2
mkdir bv2
cd bv2
git init
git remote add origin git at github.com:sghanavati/brain-view2.git
git pull -u origin master

qmake-qt4 MINCDIR=/opt/minc/
HDF5DIR=/home/mishkin/hdf5-1.8.9-linux-x86_64-static/ brain-view2.pro
make
ERROR: GeometryNode.h:25:30: fatal error: H5Cpp.h: No such file or directory
SOLUTION: sudo apt-get install libhdf5-serial-dev (probably a bad hack
because it installed 1.8.4 but it worked)

ERROR: H5AtomType.cpp:(.text+0x257): undefined reference to `H5Tget_pad'
SOLUTION: put -lhdf5_cpp before -lhdf5 in the g++ command
 g++ -m64 -o brain-view2 MainWindow.o BrainQuarter.o TreeItem.o
TreeModel.o ResourceForm.o GeometryScene.o GeometryNode.o main.o
textureColumn.o textureFileItem.o tagFileItem.o tagPointItem.o
moc_MainWindow.o moc_BrainQuarter.o moc_TreeItem.o moc_TreeModel.o
moc_ResourceForm.o moc_GeometryNode.o moc_textureColumn.o
qrc_colourbars.o    -L/usr/lib/x86_64-linux-gnu -L/usr/X11R6/lib64
-L/usr/local/lib -lCoin -L/lib/ -lQuarter -L/opt/minc//lib -loobicpl
-lpcre++ -lbicpl -lvolume_io2 -lminc2 -L/lib -lbicInventor
-L/home/mishkin/hdf5-1.8.9-linux-x86_64-static//lib/ -lhdf5_cpp -lhdf5
-lQtOpenGL -lQtGui -lQtCore -lGL -lpthread

And that gave me the brain-view2 binary which i copied into /opt/minc/bin.

From mfriedel at phenogenomics.ca  Wed Jul 18 10:38:19 2012
From: mfriedel at phenogenomics.ca (Miriam Friedel)
Date: Wed, 18 Jul 2012 10:38:19 -0400
Subject: [MINC-users] minc development tools on github
Message-ID: <3873843.204271342622299000.JavaMail.root@mail2.phenogenomics.ca>

Hello everyone,  
  
Here at MICe we are working to consolidate the 
minc-based scripts and tools we've written into a single repository on 
github. We have many different useful pieces of code that we'd like to 
consolidate in one place. In particular, we'd like to do this in such a 
way that all members of the minc community feel welcome to use this repo
 and contribute their own tools as well.   
  
To this end, we 
thought an appropriate place to do this might be adding to the existing 
conglomerate repository, which is currently on github:  
  
https://github.com/BIC-MNI/conglomerate  
  
In
 order to contribute to conglomerate, is it acceptable for new code to 
be compatible with minc2 files only, or must it be compatible with minc1
 as well?  If everything in conglomerate needs to be compatible with 
both minc1 and 2, then we will likely start a new repo, since much of 
our new code works with minc2 files only.   
  
Thanks!  
Miriam  
 

From vladimir.fonov at gmail.com  Wed Jul 18 11:00:40 2012
From: vladimir.fonov at gmail.com (Vladimir S. FONOV)
Date: Wed, 18 Jul 2012 11:00:40 -0400
Subject: [MINC-users] minc development tools on github
In-Reply-To: <3873843.204271342622299000.JavaMail.root@mail2.phenogenomics.ca>
References: <3873843.204271342622299000.JavaMail.root@mail2.phenogenomics.ca>
Message-ID: <CAMCwi7MXL=AzFQLorJEOHF4wCNG=qnSxpB8B1rAQsOy6hMnE6A@mail.gmail.com>

Hello Miriam,

I would think that it's better to start a new repository. After all
conglomerate has a very long history and most of the code haven't been
touched for long time. It would be easier for you to manage a separate
code base.

On Wed, Jul 18, 2012 at 10:38 AM, Miriam Friedel
<mfriedel at phenogenomics.ca> wrote:
> Hello everyone,
>
> Here at MICe we are working to consolidate the
> minc-based scripts and tools we've written into a single repository on
> github. We have many different useful pieces of code that we'd like to
> consolidate in one place. In particular, we'd like to do this in such a
> way that all members of the minc community feel welcome to use this repo
>  and contribute their own tools as well.
>
> To this end, we
> thought an appropriate place to do this might be adding to the existing
> conglomerate repository, which is currently on github:
>
> https://github.com/BIC-MNI/conglomerate
>
> In
>  order to contribute to conglomerate, is it acceptable for new code to
> be compatible with minc2 files only, or must it be compatible with minc1
>  as well?  If everything in conglomerate needs to be compatible with
> both minc1 and 2, then we will likely start a new repo, since much of
> our new code works with minc2 files only.
>
> Thanks!
> Miriam
>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users



-- 
Best regards,

 Vladimir S. Fonov ~ vladimir <dot> fonov <at> gmail <dot> com

From mfriedel at phenogenomics.ca  Wed Jul 18 11:14:24 2012
From: mfriedel at phenogenomics.ca (Miriam Friedel)
Date: Wed, 18 Jul 2012 11:14:24 -0400
Subject: [MINC-users] minc development tools on github
In-Reply-To: <CAMCwi7MXL=AzFQLorJEOHF4wCNG=qnSxpB8B1rAQsOy6hMnE6A@mail.gmail.com>
Message-ID: <25361963.204851342624464001.JavaMail.root@mail2.phenogenomics.ca>

Thank you Vladimir! We'll start a new repo.  
  
----- Original Message -----  
From: Vladimir S. FONOV <vladimir.fonov at gmail.com>  
Sent: Wed, 7/18/2012 11:00am  
To: MINC users mailing list <minc-users at bic.mni.mcgill.ca>  
Subject: Re: [MINC-users] minc development tools on github  
  
Hello Miriam,  
  
I would think that it's better to start a new repository. After all  
conglomerate has a very long history and most of the code haven't been  
touched for long time. It would be easier for you to manage a separate  
code base.  
  
On Wed, Jul 18, 2012 at 10:38 AM, Miriam Friedel  
<mfriedel at phenogenomics.ca> wrote:  
> Hello everyone,  
>  
> Here at MICe we are working to consolidate the  
> minc-based scripts and tools we've written into a single repository on  
> github. We have many different useful pieces of code that we'd like to  
> consolidate in one place. In particular, we'd like to do this in such a  
> way that all members of the minc community feel welcome to use this repo  
>  and contribute their own tools as well.  
>  
> To this end, we  
> thought an appropriate place to do this might be adding to the existing  
> conglomerate repository, which is currently on github:  
>  
> https://github.com/BIC-MNI/conglomerate  
>  
> In  
>  order to contribute to conglomerate, is it acceptable for new code to  
> be compatible with minc2 files only, or must it be compatible with minc1  
>  as well?  If everything in conglomerate needs to be compatible with  
> both minc1 and 2, then we will likely start a new repo, since much of  
> our new code works with minc2 files only.  
>  
> Thanks!  
> Miriam  
>  
> _______________________________________________  
> MINC-users at bic.mni.mcgill.ca  
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users  
  
  
  
--   
Best regards,  
  
 Vladimir S. Fonov ~ vladimir <dot> fonov <at> gmail <dot> com  
_______________________________________________  
MINC-users at bic.mni.mcgill.ca  
http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users  

From jason at phenogenomics.ca  Wed Jul 18 13:09:23 2012
From: jason at phenogenomics.ca (Jason Lerch)
Date: Wed, 18 Jul 2012 13:09:23 -0400
Subject: [MINC-users] installing brain-view2
In-Reply-To: <CAB-GmL3ekes-zjriKCQZ=0AvD2qVcnELZNn8KOYsUC9se+ti0A@mail.gmail.com>
References: <CAB-GmL3ekes-zjriKCQZ=0AvD2qVcnELZNn8KOYsUC9se+ti0A@mail.gmail.com>
Message-ID: <EE957E5F-D367-4317-A40B-2676DB96D28B@phenogenomics.ca>

Thanks Mishkin!

I recognize that we need to update the installation instructions - but in the meantime I've taken the liberty of adding your email in its entirety to the bottom of that wiki page. Please take a look and let me know in case you have any issues with that.

And one of these days, hopefully soon, we'll write more complete installation instructions.

Jason

On 2012-07-14, at 11:13 PM, Mishkin Derakhshan wrote:

> I had a hell of time installing brain-view2 so I documented my errors
> and solutions/hacks in case others come across the same problems
> (read: when I have to do this again I can look back at this post).
> 
> This was on a ubuntu 12.04 64 bit machine.
> 
> I followed the instruction here (thank you jason and sghanavati):
> https://wiki.phenogenomics.ca/display/MICePub/brain-view
> 
> Pre-requisites
> 1. MINC. I used the minc-toolkit binaries and only ran into one minor problem.
> opus[~]$ sudo apt-get install libc6 libstdc++6 imagemagick perl
> freeglut3 libgl1 libxcb1 libxdmcp6 libx11-6 libxext6 libxau6 libuuid1
> libjpeg62 libexpat1 libtiff4
> Reading package lists... Done
> Building dependency tree
> Reading state information... Done
> Package libgl1 is a virtual package provided by:
>  libgl1-mesa-swx11 8.0.2-0ubuntu3.1
>  libgl1-mesa-glx 8.0.2-0ubuntu3.1
> You should explicitly select one to install.
> 
> E: Package 'libgl1' has no installation candidate
> 
> I chose libgl1-mesa-glx at random, and then everything else worked.
> 
> 2. Qt4: sudo apt-get install qt4-qmake libqt4-dev libqt4-core
> 3. Boost: sudo apt-get install libboost1.48
> 4.Coin:
> wget http://ftp.coin3d.org/coin/src/all/Coin-3.1.3.tar.gz
> tar xvzf Coin-3.1.3.tar.gz
> mkdir coin-build
> cd coin-build/
> ../Coin-3.1.3/configure
> make
> sudo make install
> 
> 5. Quarter:
> wget http://ftp.coin3d.org/coin/src/all/Quarter-1.0.0.tar.gz
> tar xvzf Quarter-1.0.0.tar.gz
> mkdir quarter-build
> cd quarter-build/
> ../Quarter-1.0.0/configure
> make
> 
> ERROR:../../../Quarter-1.0.0/src/Quarter/Quarter.cpp:148:13: error:
> 'stderr' was not declared in this scope
> SOLUTION:Added
> #include "stdio.h"
> to this file: ../Quarter-1.0.0/src/Quarter/Quarter.cpp
> 
> make
> sudo make install
> 
> 6. HDF5 1.8.9:
> Downloaded a build of HDF5 1.8.9:
> http://www.hdfgroup.org/HDF5/release/obtain5.html
> tar xvzf hdf5-1.8.9-linux-x86_64-static.tar.gz
> 
> 7. BicInventor:
> wget http://packages.bic.mni.mcgill.ca/tgz/bicInventor-0.3.1.tar.gz
> cd bicInventor-0.3.1
> ./configure --prefix=/opt/minc --with-build-path=/opt/minc
> --includedir=/opt/minc/include --libdir=/opt/minc/lib --with-minc2
> ERROR: (lots of erros when I tried to configure)
> SOLUTION: sudo apt-get install build-essential realpath zlib1g-dev
> bison flex libx11-dev libxmu-dev byacc libsoqt4-dev automake libtool
> libdbi- perl libblas-dev liblapack-dev libblitz0-dev libreadline-dev
> bzr bzrtools octave3.2 imagemagick
> 
> ./configure --prefix=/opt/minc --with-build-path=/opt/minc
> --includedir=/opt/minc/include --libdir=/opt/minc/lib --with-minc2
> make
> sudo make install
> 
> ACTUALLY INSTALLING BRAIN-VIEW2
> mkdir bv2
> cd bv2
> git init
> git remote add origin git at github.com:sghanavati/brain-view2.git
> git pull -u origin master
> 
> qmake-qt4 MINCDIR=/opt/minc/
> HDF5DIR=/home/mishkin/hdf5-1.8.9-linux-x86_64-static/ brain-view2.pro
> make
> ERROR: GeometryNode.h:25:30: fatal error: H5Cpp.h: No such file or directory
> SOLUTION: sudo apt-get install libhdf5-serial-dev (probably a bad hack
> because it installed 1.8.4 but it worked)
> 
> ERROR: H5AtomType.cpp:(.text+0x257): undefined reference to `H5Tget_pad'
> SOLUTION: put -lhdf5_cpp before -lhdf5 in the g++ command
> g++ -m64 -o brain-view2 MainWindow.o BrainQuarter.o TreeItem.o
> TreeModel.o ResourceForm.o GeometryScene.o GeometryNode.o main.o
> textureColumn.o textureFileItem.o tagFileItem.o tagPointItem.o
> moc_MainWindow.o moc_BrainQuarter.o moc_TreeItem.o moc_TreeModel.o
> moc_ResourceForm.o moc_GeometryNode.o moc_textureColumn.o
> qrc_colourbars.o    -L/usr/lib/x86_64-linux-gnu -L/usr/X11R6/lib64
> -L/usr/local/lib -lCoin -L/lib/ -lQuarter -L/opt/minc//lib -loobicpl
> -lpcre++ -lbicpl -lvolume_io2 -lminc2 -L/lib -lbicInventor
> -L/home/mishkin/hdf5-1.8.9-linux-x86_64-static//lib/ -lhdf5_cpp -lhdf5
> -lQtOpenGL -lQtGui -lQtCore -lGL -lpthread
> 
> And that gave me the brain-view2 binary which i copied into /opt/minc/bin.
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users


From a.janke at gmail.com  Thu Jul 19 05:47:24 2012
From: a.janke at gmail.com (Andrew Janke)
Date: Thu, 19 Jul 2012 19:47:24 +1000
Subject: [MINC-users] minc development tools on github
In-Reply-To: <CAMCwi7MXL=AzFQLorJEOHF4wCNG=qnSxpB8B1rAQsOy6hMnE6A@mail.gmail.com>
References: <3873843.204271342622299000.JavaMail.root@mail2.phenogenomics.ca>
	<CAMCwi7MXL=AzFQLorJEOHF4wCNG=qnSxpB8B1rAQsOy6hMnE6A@mail.gmail.com>
Message-ID: <CAG0CB21avms1rZzmGygTOeKtOU=xTOGzmQgp6ki7-1K+xO3Bgg@mail.gmail.com>

Hi Miriam,

On 19 July 2012 01:00, Vladimir S. FONOV <vladimir.fonov at gmail.com> wrote:
> I would think that it's better to start a new repository. After all
> conglomerate has a very long history and most of the code haven't been
> touched for long time. It would be easier for you to manage a separate
> code base.

I'd second what Vlad has already said, there are a lot of wierd and
wonderful things in conglomerate, many of which I have no idea what
they do and there is precious little documentation beyond the source
code for them!


a

From pgravel at bic.mni.mcgill.ca  Tue Jul 24 11:20:00 2012
From: pgravel at bic.mni.mcgill.ca (Paul GRAVEL)
Date: Tue, 24 Jul 2012 11:20:00 -0400 (EDT)
Subject: [MINC-users] EMMA tools for Cygwin
In-Reply-To: <CAG0CB21avms1rZzmGygTOeKtOU=xTOGzmQgp6ki7-1K+xO3Bgg@mail.gmail.com>
References: <3873843.204271342622299000.JavaMail.root@mail2.phenogenomics.ca>
	<CAMCwi7MXL=AzFQLorJEOHF4wCNG=qnSxpB8B1rAQsOy6hMnE6A@mail.gmail.com>
	<CAG0CB21avms1rZzmGygTOeKtOU=xTOGzmQgp6ki7-1K+xO3Bgg@mail.gmail.com>
Message-ID: <Pine.LNX.4.64.1207241116180.6790@cassio.bic.mni.mcgill.ca>

Dear All,

Is there any version of the emma tools for cygwin? The reason is that I 
have installed the tools on my machine (running Windows XP x64), and I can 
actually read a minc files in matlab using the tools, but when I come to 
writing a file using newimage, I get the following error:

Output: 'micreateimage' is not recognized as an internal or external 
command, operable program or batch file.

Please note that the path are set properly (at least I hope!).

Any help would be appreciated!

Best Regards,

Paul

From a.janke at gmail.com  Tue Jul 24 18:13:51 2012
From: a.janke at gmail.com (Andrew Janke)
Date: Wed, 25 Jul 2012 08:13:51 +1000
Subject: [MINC-users] EMMA tools for Cygwin
In-Reply-To: <Pine.LNX.4.64.1207241116180.6790@cassio.bic.mni.mcgill.ca>
References: <3873843.204271342622299000.JavaMail.root@mail2.phenogenomics.ca>
	<CAMCwi7MXL=AzFQLorJEOHF4wCNG=qnSxpB8B1rAQsOy6hMnE6A@mail.gmail.com>
	<CAG0CB21avms1rZzmGygTOeKtOU=xTOGzmQgp6ki7-1K+xO3Bgg@mail.gmail.com>
	<Pine.LNX.4.64.1207241116180.6790@cassio.bic.mni.mcgill.ca>
Message-ID: <CAG0CB22xW4Tk6+FBLcV=SYK5mdtj06WAZ9p1+4gc8ae_348jCw@mail.gmail.com>

Hi Paul

On 25 July 2012 01:20, Paul GRAVEL <pgravel at bic.mni.mcgill.ca> wrote:
> Is there any version of the emma tools for cygwin?

Not that I know of.  In any case EMMA has been deprecated for quite
some time now in favour of Pierre Bellecs NIAK python package

   http://code.google.com/p/niak/


a

From pgravel at bic.mni.mcgill.ca  Wed Jul 25 12:15:14 2012
From: pgravel at bic.mni.mcgill.ca (Paul GRAVEL)
Date: Wed, 25 Jul 2012 12:15:14 -0400 (EDT)
Subject: [MINC-users] EMMA tools for Cygwin
In-Reply-To: <CAG0CB22xW4Tk6+FBLcV=SYK5mdtj06WAZ9p1+4gc8ae_348jCw@mail.gmail.com>
References: <3873843.204271342622299000.JavaMail.root@mail2.phenogenomics.ca>
	<CAMCwi7MXL=AzFQLorJEOHF4wCNG=qnSxpB8B1rAQsOy6hMnE6A@mail.gmail.com>
	<CAG0CB21avms1rZzmGygTOeKtOU=xTOGzmQgp6ki7-1K+xO3Bgg@mail.gmail.com>
	<Pine.LNX.4.64.1207241116180.6790@cassio.bic.mni.mcgill.ca>
	<CAG0CB22xW4Tk6+FBLcV=SYK5mdtj06WAZ9p1+4gc8ae_348jCw@mail.gmail.com>
Message-ID: <Pine.LNX.4.64.1207251152570.14307@cassio.bic.mni.mcgill.ca>

Many thanks Andrew!

I did install the NIAK tools and added the path(s) in Matlab, but I now 
face another problem. When I run the niak_read_minc (in Matlab) command, I 
get the following error message:

##################################################
??? Error using ==> niak_read_hdr_minc at 96
niak:read: Could not parse the minc header !

Error in ==> niak_read_minc at 70
hdr = niak_read_hdr_minc(file_name);
##################################################

where line 96 of niak_read_hdr_minc is returning an error baseb on line 87 
which is:
[flag,str_header] = system(cat(2,'mincheader ',file_name));

When I run this command from Matlab:
file_name = 'qfal_003_0_pet_fr1_32_sm9mm_VoWal_basFunc_k2.mnc'
[flag,str_header] = system(cat(2,'mincheader ',file_name))

I get the following error message:
##################################################
flag =

      1


str_header =

'mincheader' is not recognized as an internal or external command,
operable program or batch file.
##################################################

while it runs fine in cygwin. And yes I have added the cygwin and MINC 
tools paths (C:\cygwin\bin;C:\cygwin\usr\local\mni\bin) to my Windows XP 
path environment variable. For example, when I call mincinfo from within 
Matlab, it runs fine.

In addition, I have noticed that when I do an long listing of the MINC 
tools folder from cygwin: ls -l /usr/local/mni/bin
Most of the MINC tools have the .exe extension, e.g. mincinfo.exe, while 
mincheader does not. I renamed mincheader to mincheader.exe but it did not 
fix the problem. However, the error message in Matlab differed: 
##################################################
flag =

      1


str_header =

The image file C:\cygwin\usr\local\mni\bin\mincheader.exe is valid, but is 
for a machine type other than the current machine.
##################################################

Could this be a result of how it was compiled, and thus cause the error 
explained above? And if so, is there another version of mincheader that 
could solve this problem?

I am using the following MINC tools package:
http://packages.bic.mni.mcgill.ca/cygwin/mni-2006-11-15.tar.gz

and I am running MATLAB Version 7.10.0.499 (R2010a)

Any help will be greatly appreciated!

Best Regards,

Paul


On Wed, 25 Jul 2012, Andrew Janke wrote:

> Hi Paul
>
> On 25 July 2012 01:20, Paul GRAVEL <pgravel at bic.mni.mcgill.ca> wrote:
>> Is there any version of the emma tools for cygwin?
>
> Not that I know of.  In any case EMMA has been deprecated for quite
> some time now in favour of Pierre Bellecs NIAK python package
>
>   http://code.google.com/p/niak/
>
>
> a
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From a.janke at gmail.com  Wed Jul 25 17:55:05 2012
From: a.janke at gmail.com (Andrew Janke)
Date: Thu, 26 Jul 2012 07:55:05 +1000
Subject: [MINC-users] EMMA tools for Cygwin
In-Reply-To: <Pine.LNX.4.64.1207251152570.14307@cassio.bic.mni.mcgill.ca>
References: <3873843.204271342622299000.JavaMail.root@mail2.phenogenomics.ca>
	<CAMCwi7MXL=AzFQLorJEOHF4wCNG=qnSxpB8B1rAQsOy6hMnE6A@mail.gmail.com>
	<CAG0CB21avms1rZzmGygTOeKtOU=xTOGzmQgp6ki7-1K+xO3Bgg@mail.gmail.com>
	<Pine.LNX.4.64.1207241116180.6790@cassio.bic.mni.mcgill.ca>
	<CAG0CB22xW4Tk6+FBLcV=SYK5mdtj06WAZ9p1+4gc8ae_348jCw@mail.gmail.com>
	<Pine.LNX.4.64.1207251152570.14307@cassio.bic.mni.mcgill.ca>
Message-ID: <CAG0CB21ShcgMkyRAKtfyBzuuPtuJULCtHnQLexdXiv_gicUS1A@mail.gmail.com>

Hi Paul,

> ##################################################
> ??? Error using ==> niak_read_hdr_minc at 96
> niak:read: Could not parse the minc header !
>
> Error in ==> niak_read_minc at 70
> hdr = niak_read_hdr_minc(file_name);
> ##################################################


> Could this be a result of how it was compiled, and thus cause the error
> explained above? And if so, is there another version of mincheader that
> could solve this problem?
>
> I am using the following MINC tools package:
> http://packages.bic.mni.mcgill.ca/cygwin/mni-2006-11-15.tar.gz

>From these two bits of information I have two guesses as to what is up.

1. It's a MINC 2 files (HDF) and I'll bet that my old 2006 build of
minc on cygwin isn't a MINC2 build.  What does this return?:

   $ file qfal_003_0_pet_fr1_32_sm9mm_VoWal_basFunc_k2.mnc

2. You are using a 64 bit install of windows and the old cygwin build
will be 32 bit.

Both of these things mean that you are going to have to recompile the
MINC tools on cygwin for your machine. I haven't done this in a while
as I find it a lot easier to simply use the neurodebian VM on windows
machines.


a

From pgravel at bic.mni.mcgill.ca  Thu Jul 26 10:32:30 2012
From: pgravel at bic.mni.mcgill.ca (Paul GRAVEL)
Date: Thu, 26 Jul 2012 10:32:30 -0400 (EDT)
Subject: [MINC-users] EMMA tools for Cygwin
In-Reply-To: <CAG0CB21ShcgMkyRAKtfyBzuuPtuJULCtHnQLexdXiv_gicUS1A@mail.gmail.com>
References: <3873843.204271342622299000.JavaMail.root@mail2.phenogenomics.ca>
	<CAMCwi7MXL=AzFQLorJEOHF4wCNG=qnSxpB8B1rAQsOy6hMnE6A@mail.gmail.com>
	<CAG0CB21avms1rZzmGygTOeKtOU=xTOGzmQgp6ki7-1K+xO3Bgg@mail.gmail.com>
	<Pine.LNX.4.64.1207241116180.6790@cassio.bic.mni.mcgill.ca>
	<CAG0CB22xW4Tk6+FBLcV=SYK5mdtj06WAZ9p1+4gc8ae_348jCw@mail.gmail.com>
	<Pine.LNX.4.64.1207251152570.14307@cassio.bic.mni.mcgill.ca>
	<CAG0CB21ShcgMkyRAKtfyBzuuPtuJULCtHnQLexdXiv_gicUS1A@mail.gmail.com>
Message-ID: <Pine.LNX.4.64.1207261029020.11827@cassio.bic.mni.mcgill.ca>

Many thanks Andrew!

I think someone has asked the following question previously, but just to 
bu sure: where can I find the latest MINC source files to be recompiled 
for Cygwin?

As for the MINC1 vs MINC2 format, the file is actually in MINC1 format:
$file qfal_003_0_pet_fr1_32_sm9mm_VoWal_basFunc_k2.mnc
qfal_003_0_pet_fr1_32_sm9mm_VoWal_basFunc_k2.mnc: NetCDF Data Format data

Best Regards,

Paul


On Thu, 26 Jul 2012, Andrew Janke wrote:

> Hi Paul,
>
>> ##################################################
>> ??? Error using ==> niak_read_hdr_minc at 96
>> niak:read: Could not parse the minc header !
>>
>> Error in ==> niak_read_minc at 70
>> hdr = niak_read_hdr_minc(file_name);
>> ##################################################
>
>
>> Could this be a result of how it was compiled, and thus cause the error
>> explained above? And if so, is there another version of mincheader that
>> could solve this problem?
>>
>> I am using the following MINC tools package:
>> http://packages.bic.mni.mcgill.ca/cygwin/mni-2006-11-15.tar.gz
>
>> From these two bits of information I have two guesses as to what is up.
>
> 1. It's a MINC 2 files (HDF) and I'll bet that my old 2006 build of
> minc on cygwin isn't a MINC2 build.  What does this return?:
>
>   $ file qfal_003_0_pet_fr1_32_sm9mm_VoWal_basFunc_k2.mnc
>
> 2. You are using a 64 bit install of windows and the old cygwin build
> will be 32 bit.
>
> Both of these things mean that you are going to have to recompile the
> MINC tools on cygwin for your machine. I haven't done this in a while
> as I find it a lot easier to simply use the neurodebian VM on windows
> machines.
>
>
> a
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From claude at bic.mni.mcgill.ca  Thu Jul 26 10:56:12 2012
From: claude at bic.mni.mcgill.ca (Claude LEPAGE)
Date: Thu, 26 Jul 2012 10:56:12 -0400
Subject: [MINC-users] EMMA tools for Cygwin
In-Reply-To: <Pine.LNX.4.64.1207261029020.11827@cassio.bic.mni.mcgill.ca>
Message-ID: <201207261456.q6QEuCr3017094@cassio.bic.mni.mcgill.ca>

Paul,

> I think someone has asked the following question previously, but just to 
> bu sure: where can I find the latest MINC source files to be recompiled 
> for Cygwin?

I have some here:

http://www.bic.mni.mcgill.ca/users/claude/cygwin_nt-5.1-1.7.7.tar.gz

Basic minc tools, but likely enough for most users. This is a non-official
package. Use at your own risks. It's 32 bits.

Claude


From pgravel at bic.mni.mcgill.ca  Thu Jul 26 11:39:55 2012
From: pgravel at bic.mni.mcgill.ca (Paul GRAVEL)
Date: Thu, 26 Jul 2012 11:39:55 -0400 (EDT)
Subject: [MINC-users] EMMA tools for Cygwin
In-Reply-To: <201207261456.q6QEuCr3017094@cassio.bic.mni.mcgill.ca>
References: <201207261456.q6QEuCr3017094@cassio.bic.mni.mcgill.ca>
Message-ID: <Pine.LNX.4.64.1207261131210.11781@cassio.bic.mni.mcgill.ca>

Many Thanks Claude!

Unfortunately, I still have the same problem with mincheader, and indeed 
it seems to be the 64-bit architecture that is causing the problem... But 
the question remains: why would it run fine in Cygwin (which is installed 
on the Windows XP x64 machine), but not when called via Windows, while 
tools like mincinfo runs fine in both cygwin and matlab?

In the meantime I will verify if there is a way to call commands from 
matlab by specifying a 32-bit flag...

Best,

Paul

On Thu, 26 Jul 2012, Claude LEPAGE wrote:

> Paul,
>
>> I think someone has asked the following question previously, but just to
>> bu sure: where can I find the latest MINC source files to be recompiled
>> for Cygwin?
>
> I have some here:
>
> http://www.bic.mni.mcgill.ca/users/claude/cygwin_nt-5.1-1.7.7.tar.gz
>
> Basic minc tools, but likely enough for most users. This is a non-official
> package. Use at your own risks. It's 32 bits.
>
> Claude
>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From pgravel at bic.mni.mcgill.ca  Thu Jul 26 12:39:57 2012
From: pgravel at bic.mni.mcgill.ca (Paul GRAVEL)
Date: Thu, 26 Jul 2012 12:39:57 -0400 (EDT)
Subject: [MINC-users] EMMA tools for Cygwin
In-Reply-To: <Pine.LNX.4.64.1207261131210.11781@cassio.bic.mni.mcgill.ca>
References: <201207261456.q6QEuCr3017094@cassio.bic.mni.mcgill.ca>
	<Pine.LNX.4.64.1207261131210.11781@cassio.bic.mni.mcgill.ca>
Message-ID: <Pine.LNX.4.64.1207261237450.25626@cassio.bic.mni.mcgill.ca>

Hi All!

I think I got it!

Line 87 of niak_read_hdr_minc.m was modified from:
[flag,str_header] = system(cat(2,'mincheader.exe ',file_name));

to:

[flag,str_header] = system(cat(2,'bash -c "mincheader ',file_name,'"'));

so it ensures calling the mincheader command using cygwin's bash, instead 
of windows.

Many thanks again!

Paul


On Thu, 26 Jul 2012, Paul GRAVEL wrote:

> Many Thanks Claude!
>
> Unfortunately, I still have the same problem with mincheader, and indeed it 
> seems to be the 64-bit architecture that is causing the problem... But the 
> question remains: why would it run fine in Cygwin (which is installed on the 
> Windows XP x64 machine), but not when called via Windows, while tools like 
> mincinfo runs fine in both cygwin and matlab?
>
> In the meantime I will verify if there is a way to call commands from matlab 
> by specifying a 32-bit flag...
>
> Best,
>
> Paul
>
> On Thu, 26 Jul 2012, Claude LEPAGE wrote:
>
>> Paul,
>> 
>>> I think someone has asked the following question previously, but just to
>>> bu sure: where can I find the latest MINC source files to be recompiled
>>> for Cygwin?
>> 
>> I have some here:
>> 
>> http://www.bic.mni.mcgill.ca/users/claude/cygwin_nt-5.1-1.7.7.tar.gz
>> 
>> Basic minc tools, but likely enough for most users. This is a non-official
>> package. Use at your own risks. It's 32 bits.
>> 
>> Claude
>> 
>> _______________________________________________
>> MINC-users at bic.mni.mcgill.ca
>> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>> 
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From pbellec at bic.mni.mcgill.ca  Thu Jul 26 14:41:11 2012
From: pbellec at bic.mni.mcgill.ca (Pierre Bellec)
Date: Thu, 26 Jul 2012 14:41:11 -0400
Subject: [MINC-users] EMMA tools for Cygwin
In-Reply-To: <Pine.LNX.4.64.1207261237450.25626@cassio.bic.mni.mcgill.ca>
References: <201207261456.q6QEuCr3017094@cassio.bic.mni.mcgill.ca>
	<Pine.LNX.4.64.1207261131210.11781@cassio.bic.mni.mcgill.ca>
	<Pine.LNX.4.64.1207261237450.25626@cassio.bic.mni.mcgill.ca>
Message-ID: <CADPHwxCPLza6a-A6A2CLj2_bRgXf-y8e_MwtReVhufus_ZPvXw@mail.gmail.com>

Dear Paul,

Sorry for taking so long to jump in. I tested the niak reader and writer of
MINC on cygwin years ago, but I ran into various issues at the time. Since
MINC was not officially supported on this platform, I decided to abandon
windows altogether and assume people would install a ubuntu virtual machine
if they needed to read/write minc. I am thus very happy to see you managed
to get it to work, that's great news. However, the change you suggested is
what's in the code already.
http://code.google.com/p/niak/source/browse/trunk/commands/read_write/niak_read_hdr_minc.m

Maybe you were working on a version that had been edited for windows but
somehow ended up breaking things with cygwin.

Best,

Pierre Bellec, PhD
Research Centre of the Montreal Geriatric Institute
& Department of Computer Science and Operations Research
University of Montreal, Qu?bec, Canada
http://simexp-lab.org/brainwiki/doku.php?id=pierrebellec
(001)(514) 340 3540 #3367


2012/7/26 Paul GRAVEL <pgravel at bic.mni.mcgill.ca>

> Hi All!
>
> I think I got it!
>
> Line 87 of niak_read_hdr_minc.m was modified from:
> [flag,str_header] = system(cat(2,'mincheader.exe ',file_name));
>
> to:
>
> [flag,str_header] = system(cat(2,'bash -c "mincheader ',file_name,'"'));
>
> so it ensures calling the mincheader command using cygwin's bash, instead
> of windows.
>
> Many thanks again!
>
> Paul
>
>
>
> On Thu, 26 Jul 2012, Paul GRAVEL wrote:
>
>  Many Thanks Claude!
>>
>> Unfortunately, I still have the same problem with mincheader, and indeed
>> it seems to be the 64-bit architecture that is causing the problem... But
>> the question remains: why would it run fine in Cygwin (which is installed
>> on the Windows XP x64 machine), but not when called via Windows, while
>> tools like mincinfo runs fine in both cygwin and matlab?
>>
>> In the meantime I will verify if there is a way to call commands from
>> matlab by specifying a 32-bit flag...
>>
>> Best,
>>
>> Paul
>>
>> On Thu, 26 Jul 2012, Claude LEPAGE wrote:
>>
>>  Paul,
>>>
>>>  I think someone has asked the following question previously, but just to
>>>> bu sure: where can I find the latest MINC source files to be recompiled
>>>> for Cygwin?
>>>>
>>>
>>> I have some here:
>>>
>>> http://www.bic.mni.mcgill.ca/**users/claude/cygwin_nt-5.1-1.**7.7.tar.gz<http://www.bic.mni.mcgill.ca/users/claude/cygwin_nt-5.1-1.7.7.tar.gz>
>>>
>>> Basic minc tools, but likely enough for most users. This is a
>>> non-official
>>> package. Use at your own risks. It's 32 bits.
>>>
>>> Claude
>>>
>>> ______________________________**_________________
>>> MINC-users at bic.mni.mcgill.ca
>>> http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
>>>
>>>  ______________________________**_________________
>> MINC-users at bic.mni.mcgill.ca
>> http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
>>
>>  ______________________________**_________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/**mailman/listinfo/minc-users<http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users>
>

From antoine.pattin at gmail.com  Fri Jul 27 11:12:47 2012
From: antoine.pattin at gmail.com (antoine pattin)
Date: Fri, 27 Jul 2012 17:12:47 +0200
Subject: [MINC-users] [MINC-USERS] getting 4D volume datas
Message-ID: <CACWOJq4+Az0d3TtaoDGbCNcFMStYx6=qmuH6XNz+Mt95JedoTQ@mail.gmail.com>

Hello everybody !

Is it possible that the miget_real_value_hyperslab() function isn't
compatible with 4D volumes ?
(You can see more details about it at the link :
http://minc.sourcearchive.com/documentation/2.1.00-1/hyper_8c_bac30780b4cebea089eddb8f687ce3c0.html
)

My objectif is to store dynamic volume datas in a hyperslab (1D array).
The solution could be inspired from the following and very helpful tutorial
:
http://en.wikibooks.org/wiki/MINC/Tutorials/Programming05

I tested it with a static volume (3D) and it perfectly worked.
I then adapted the code for a 4D data volume and got this executing message
:

*HDF5-DIAG: Error detected in HDF5 library version: 1.6.1 thread 0.  Back
trace follows.
  #000: H5Dio.c line 356 in H5Dread(): selection+offset not within extent
    major(14): Dataspace interface
    minor(04): Out of range
(from miicv_get): Can't read dataset /minc-2.0/image/0/image
Error getting hyperslab*

Is it because of the 4 dimensions or is it about other parameters linked to
the volume tested ?

Any help would be appreciated!

Best Regards,
Antoine

From baghdadi at phenogenomics.ca  Fri Jul 27 12:24:52 2012
From: baghdadi at phenogenomics.ca (Leila Baghdadi)
Date: Fri, 27 Jul 2012 12:24:52 -0400
Subject: [MINC-users] [MINC-USERS] getting 4D volume datas
In-Reply-To: <CACWOJq4+Az0d3TtaoDGbCNcFMStYx6=qmuH6XNz+Mt95JedoTQ@mail.gmail.com>
Message-ID: <1080932.303591343406292138.JavaMail.root@mail2.phenogenomics.ca>

Antoine,

I am not at work today so can not check for sure but I can recall having an example for 4D volumes (minc2) in libsrc2/test

perhaps you should check that out

HTH

Leila

----- Original Message -----
From: antoine pattin <antoine.pattin at gmail.com>
Sent: Fri, 7/27/2012 11:12am
To: minc-users at bic.mni.mcgill.ca
Subject: [MINC-users] [MINC-USERS] getting 4D volume datas

Hello everybody !

Is it possible that the miget_real_value_hyperslab() function isn't
compatible with 4D volumes ?
(You can see more details about it at the link :
http://minc.sourcearchive.com/documentation/2.1.00-1/hyper_8c_bac30780b4cebea089eddb8f687ce3c0.html
)

My objectif is to store dynamic volume datas in a hyperslab (1D array).
The solution could be inspired from the following and very helpful tutorial
:
http://en.wikibooks.org/wiki/MINC/Tutorials/Programming05

I tested it with a static volume (3D) and it perfectly worked.
I then adapted the code for a 4D data volume and got this executing message
:

*HDF5-DIAG: Error detected in HDF5 library version: 1.6.1 thread 0.  Back
trace follows.
  #000: H5Dio.c line 356 in H5Dread(): selection+offset not within extent
    major(14): Dataspace interface
    minor(04): Out of range
(from miicv_get): Can't read dataset /minc-2.0/image/0/image
Error getting hyperslab*

Is it because of the 4 dimensions or is it about other parameters linked to
the volume tested ?

Any help would be appreciated!

Best Regards,
Antoine
_______________________________________________
MINC-users at bic.mni.mcgill.ca
http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users

From antoine.pattin at insa-lyon.fr  Mon Jul 30 09:50:48 2012
From: antoine.pattin at insa-lyon.fr (antoine pattin)
Date: Mon, 30 Jul 2012 15:50:48 +0200
Subject: [MINC-users] [MINC-USERS] getting 4D volume datas
In-Reply-To: <1080932.303591343406292138.JavaMail.root@mail2.phenogenomics.ca>
References: <CACWOJq4+Az0d3TtaoDGbCNcFMStYx6=qmuH6XNz+Mt95JedoTQ@mail.gmail.com>
	<1080932.303591343406292138.JavaMail.root@mail2.phenogenomics.ca>
Message-ID: <CACWOJq5E7_WYa7p=5v9EfcS=ADRjUsDHXd1m2j1-UnWqtMcS8Q@mail.gmail.com>

Thanks Leila for such a quick answer

I checked the file you mentioned and there are indeed a few examples that
might help me.

Best regards,
Antoine


2012/7/27 Leila Baghdadi <baghdadi at phenogenomics.ca>

> Antoine,
>
> I am not at work today so can not check for sure but I can recall having
> an example for 4D volumes (minc2) in libsrc2/test
>
> perhaps you should check that out
>
> HTH
>
> Leila
>
> ----- Original Message -----
> From: antoine pattin <antoine.pattin at gmail.com>
> Sent: Fri, 7/27/2012 11:12am
> To: minc-users at bic.mni.mcgill.ca
> Subject: [MINC-users] [MINC-USERS] getting 4D volume datas
>
> Hello everybody !
>
> Is it possible that the miget_real_value_hyperslab() function isn't
> compatible with 4D volumes ?
> (You can see more details about it at the link :
>
> http://minc.sourcearchive.com/documentation/2.1.00-1/hyper_8c_bac30780b4cebea089eddb8f687ce3c0.html
> )
>
> My objectif is to store dynamic volume datas in a hyperslab (1D array).
> The solution could be inspired from the following and very helpful tutorial
> :
> http://en.wikibooks.org/wiki/MINC/Tutorials/Programming05
>
> I tested it with a static volume (3D) and it perfectly worked.
> I then adapted the code for a 4D data volume and got this executing message
> :
>
> *HDF5-DIAG: Error detected in HDF5 library version: 1.6.1 thread 0.  Back
> trace follows.
>   #000: H5Dio.c line 356 in H5Dread(): selection+offset not within extent
>     major(14): Dataspace interface
>     minor(04): Out of range
> (from miicv_get): Can't read dataset /minc-2.0/image/0/image
> Error getting hyperslab*
>
> Is it because of the 4 dimensions or is it about other parameters linked to
> the volume tested ?
>
> Any help would be appreciated!
>
> Best Regards,
> Antoine
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>