profile
viewpoint
Gabriel A. Devenyi gdevenyi Brain Imaging Center, Douglas Mental Health University Institute, @CobraLab Montreal, Quebec, Canada https://gabriel.devenyi.ca Research Computing Associate at Brain Imaging Center, Douglas Mental Health University Institute

BrkRaw/brkraw 26

BrkRaw: A comprehensive tool to access raw Bruker Biospin MRI data

carpentries-incubator/shell-extras 22

Extra Unix Shell Material

gdevenyi/AllenHumanGeneMNI 6

Careful multispectral registration of the Allen Human Brain Gene Samples in MNI ICBM NLIN SYM 09c space

devdinie/UMAGeT 1

U-Net Biomedical Image Segmentation

dglennie/mc_tissue_sphere_mega 1

Monte Carlo simulation for measuring photons detected with sphere over tissue used to calculate scattering losses as a function of optical properties

dorkylever/Ark_Pydpiper 0

Repository for Image Processing and Configuration Settings for Pydpiper Analysis for the Arkell Group

gdevenyi/2014-05-05-mcmaster 0

Software Carpentry Bootcamp Repo for McMaster May 5-6 2014

gdevenyi/2014-06-04-cornell 0

Bootcamp Repository at Cornell June 4-5

issue commentgetredash/redash

redash/redash:10.1.0.b50633 docker image fails to run on aarch64

@susodapop If I simply conditionally do these steps, will redash still build/run, or does it hard-require these bits which means changes elsewhere? https://github.com/getredash/redash/blob/master/Dockerfile#L64-L78

I can easily change the Dockerfile, but I don't know anything about the redash codebase.

gdevenyi

comment created time in 14 hours

issue commentgetredash/redash

redash/redash:10.1.0.b50633 docker image fails to run on aarch64

Will need to add a TARGETPLATFORM conditional since the MS drivers only support amd64

https://nielscautaerts.xyz/making-dockerfiles-architecture-independent.html

gdevenyi

comment created time in 14 hours

pull request commentgetredash/redash

Fix"Unable to locate package msodbcsql17"on M1

This, while it allows M1 to work, will forever break any other ARM system by hard-coding the docker builds.

M1 users should be doing this, https://stackoverflow.com/questions/65612411/forcing-docker-to-use-linux-amd64-platform-by-default-on-macos

And this Docker file needs some conditional addition of this driver so it will build on ARM.

zhengr

comment created time in 14 hours

issue commentgetredash/redash

redash/redash:10.1.0.b50633 docker image fails to run on aarch64

My guess here, is it's exactly #5638 that causes this failure after reading more carefully. Its hard-coded an amd64 compiled package, which may work on M1 (it has emulation https://www.imore.com/tests-show-apples-m1-emulates-x86-faster-intel-can-run-it-natively) but real ARM systems like a Raspberry Pi or a Oracle Ampere instance obviously don't.

gdevenyi

comment created time in 14 hours

issue commentopenzfs/zfs

ZFS send/recv with ashift 9->12 leads to data corruption

2.0.7 and 2.1.2 both include ports of 3760273, so you could go either route.

Thanks, that's helpful. Any idea of 0.8.6 is affected?

I see the issue is -c so I guess I could also avoid that.

aither64

comment created time in 16 hours

issue commentopenzfs/zfs

ZFS send/recv with ashift 9->12 leads to data corruption

Any idea here of what versions this affects? I have a 140TB migration to do from a 9->12 pool and not super happy having to upgrade to master to get the send working properly...

aither64

comment created time in 16 hours

issue commentMouse-Imaging-Centre/RMINC

mincLmer will converge when specifying summary_type = "fixef", yet will not when summary_type = "ranef" or "both"

A fortuitous discovery by another student doing the same thing. Her code runs with "|" as the ranef slope seperated (correlated slopes) and fails with the same error as the initial report if run with "||" (uncorrelated slopes).

Jeanne-Mance

comment created time in 2 days

issue commentANTsX/ANTsPy

Save regsitration at each iteration

Oops apologies I missed this was the antspy repo.

johnrr2305

comment created time in 2 days

issue commentANTsX/ANTsPy

Save regsitration at each iteration

     --write-interval-volumes <unsignedIntegerValue>
          Writes out the output volume at each iteration. It helps to present the 
          registration process as a short movie a value of 0 (the default) indicates that 
          this option should not take placeany value greater than 0 represents the 
          interval between the iterations which outputs are written to the disk. 
          <VALUES>: 0
johnrr2305

comment created time in 2 days

issue commentaces/Loris

Prevent LORIS from expiring the admin user

I don't like the idea of adding more special casing for he superuser.. especially around having it bypass a security measure.

If LORIS ever intends to offer managed cloud instances which don't require daily babysitting, you may want to reconsider allowing the superuser to be expired unable to login requiring database edits.

gdevenyi

comment created time in 2 days

startedrmcelreath/stat_rethinking_2022

started time in 3 days

pull request commentzlib-ng/zlib-ng

Memcpy/memcmp with unaligned header

Oracle cloud has free amprere arm VMs

nmoinvaz

comment created time in 5 days

issue openedaces/Loris

Prevent LORIS from expiring the admin user

Discussed in https://github.com/aces/Loris/discussions/7949

<div type='discussions-op-text'>

<sup>Originally posted by gdevenyi January 12, 2022</sup> I have inherited a LORIS instance that has been lightly used in the past year, and which I have not logged into.

I tried to login today to find: image

This shouldn't happen.

created time in 5 days

GollumEvent

issue commentBrkRaw/brkraw

xlrd 2.0.1 Dropped Supports of XLSX

HTML5 based GUI

GUIs are fine and all, but the workflows I use brkraw for are command-line and preferably automated, I'd like to see brkraw move in the direction of https://github.com/nipy/heudiconv for automation

cecilyen

comment created time in 6 days

issue commentSimpleITK/SimpleITK

"Streaming" a pythonic-iterator to access voxels?

Continuing brainstorming here. I would be interesting to have a python iterator for an image/file which generates image regions. This could be used with map to process image chunks.

This is exactly what I'm thinking about this.

Long term -- the idea here is wrapping SimpleITK + statsmodels to do voxel-wise statistics without loading everything.

gdevenyi

comment created time in 6 days

issue openedgetredash/redash

redash/redash:10.1.0.b50633 docker image fails to run on aarch64

Issue Summary

Using the https://github.com/getnewdash/setup setup script (which has been updated with new versions of everything) on an Oracle Cloud arrach64 instance:

$ uname -a
Linux instance-20220113-1107 5.11.0-1022-oracle #23~20.04.1-Ubuntu SMP Fri Nov 12 15:45:47 UTC 2021 aarch64 aarch64 aarch64 GNU/Linux

The create_db step in the setup causes the following error:

Status: Downloaded newer image for redash/redash:10.1.0.b50633
Creating redash_postgres_1 ... done
Creating redash_redis_1    ... done
Creating redash_server_run ... done
standard_init_linux.go:228: exec user process caused: exec format error
ERROR: 1

Steps to Reproduce

  1. Create oracle cloud Ampere instance
  2. Run setup.sh

Any other info e.g. Why do you consider this to be a bug? What did you expect to happen instead?

Technical details:

  • Redash Version: redash/redash:10.1.0.b50633
  • Browser/OS:
  • How did you install Redash: https://github.com/getnewdash/setup

created time in 6 days

startedgetredash/redash

started time in 6 days

PR opened defencore/gpd-pocket-3-linux

Fix path
+1 -1

0 comment

1 changed file

pr created time in 6 days

push eventgdevenyi/gpd-pocket-3-linux

Gabriel A. Devenyi

commit sha 55e2150cda587c4210aa32d7f3568a584adb125d

Fix path

view details

push time in 6 days

starteddefencore/gpd-pocket-3-linux

started time in 6 days

issue commentzlib-ng/zlib-ng

LTO should be enabled on architecture specific source files when using "native" optimizations

CMake can do this right

https://stackoverflow.com/a/47370726

KungFuJesus

comment created time in 7 days

issue openedSimpleITK/SimpleITK

"Streaming" a pythonic-iterator to access voxels?

I see from the examples here: https://simpleitk.readthedocs.io/en/master/link_AdvancedImageReading_docs.html

That some ImageIO interfaces offer the ability to load partial files. (I'm guessing the limitations listed are due to the underlying ITK implementations)

I think an interesting feature for (the python part at least) of SimpleITK would be to provide a generator/iterator which could be used to read all voxels in a streaming manner using this interface.

created time in 7 days

push eventgdevenyi/minc-diff-labels

Gabriel A. Devenyi

commit sha 9f6ad5a6726c74d2c1d960466a2f9a0f8d567fb9

Switch to wide images

view details

push time in 8 days

GollumEvent
GollumEvent
GollumEvent

issue commentMouse-Imaging-Centre/RMINC

Can't run pMincApply example code: object 'fun' of mode 'function' was not found

I've tried to generate a reproducer using the test data and been unable to. Looking into how to provide a pared down dataset.

gdevenyi

comment created time in 8 days

issue openedMouse-Imaging-Centre/RMINC

mincGetVoxel iterator?

As an alternative to solving #299 problems because we would like to move forward, I was looking at doing the dumbest iteration over all voxels in the inputs one at a time. To my suprise I can't find anything equivalent to an iterator I could use to get one voxel at a time sequentially. Does something like this exist (or is there a workaround?)

created time in 9 days

more