Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Daniel J Wilson
    @danieljwilson
    no it doesn't look like I have nb_conda installed...
    Mihai Capotă
    @mihaic
    Have a look at the docs for that. You should be able to select the environment where you installed brainiak_tutorials with it:
    http://docs.continuum.io/anaconda/user-guide/tasks/use-jupyter-notebook-extensions/
    Daniel J Wilson
    @danieljwilson
    I made the kernel accessible in Jupyter Lab using:
    python -m ipykernel install --user --name brainiac_36 --display-name "brainIAK 3.6"
    and it is available when I launch Jupyter Lab...
    Mihai Capotă
    @mihaic
    You misspelled the name.
    According to your own pasting, it should be brainiak_36.
    Daniel J Wilson
    @danieljwilson
    was writing from memory...actual code
    python -m ipykernel install --user --name brainiak_36 --display-name "Brainiak 3.6"
    Mihai Capotă
    @mihaic
    Not sure what's happening. I would still try nb_conda, although it's probably doing something similar.
    Daniel J Wilson
    @danieljwilson
    hm, can try, not sure why it would be different? Have made my other kernels all available with this methods without issue..
    Daniel J Wilson
    @danieljwilson

    Just an update @mihaic - so I deleted the environment and recreated from scratch. Got an MPI related error (as before):

    (brainiak_36) DJW-2:brainIAK djw$ python3 -c "import brainiak.funcalign.srm"
    Traceback (most recent call last):
      File "<string>", line 1, in <module>
      File "/Users/djw/anaconda3/envs/brainiak_36/lib/python3.6/site-packages/brainiak/funcalign/srm.py", line 42, in <module>
        from mpi4py import MPI
    ImportError: dlopen(/Users/djw/anaconda3/envs/brainiak_36/lib/python3.6/site-packages/mpi4py/MPI.cpython-36m-darwin.so, 2): Library not loaded: @rpath/libmpi.1.dylib
      Referenced from: /Users/djw/anaconda3/envs/brainiak_36/lib/python3.6/site-packages/mpi4py/MPI.cpython-36m-darwin.so
      Reason: image not found

    Which was fixed by running:
    brew install mpich

    Then made the kernel accessible using nb_conda instead of installing via ipykernel and things worked.

    Mihai Capotă
    @mihaic
    Thanks for reporting back, @danieljwilson. Glad things worked. I would have expected Conda to take care of MPI, but it's good to know that installing from Brew fixes the error.
    ltetrel
    @ltetrel

    Hi guys,
    I am trying to put the brainiak tutorial on our new publication platform (binder).
    I have trouble to install brainiak with the docker image :

    Step 4/15 : COPY brainiak brainiak
    COPY failed

    (sorry I cannot paste the content)

    manojneuro
    @manojneuro
    @mihaic for more info: @ltetrel is using the dockerfile in the brainiak-tutorial repo to build.
    This is part of neurolibre, a cluster for public use, and any user will be able to run the brainiak tutorials on this cluster in binder.
    ltetrel
    @ltetrel
    Ok so the Dockerfile in brainiak/tutorial cannot work because there is no brainiak folder in this repo
    manojneuro
    @manojneuro
    @mihaic: @ltetrel took a different route, using a requirements file to build on binder (neurolibre).
    Mihai Capotă
    @mihaic
    Hello, @ltetrel! I hope you got it working. If you still want to build a Docker image, please have a look at the script in our tutorials repository:
    manojneuro
    @manojneuro
    @mihaic thanks. The route we took (due to various errors on the binder build) was to create a requirements and apt file for brainiak and tutorials: https://github.com/neurolibre/brainiak-tutorials/tree/master/binder
    The script you provided would help get the master branch of brainiak.
    Clare Grall
    @claregrall
    Hi Brainiak! First- this resource has been transformative for my career, so BIG THANK YOU! My question: Is there an "answer guide" for checking responses to the questions and exercises in the tutorials?
    Daniel J Wilson
    @danieljwilson
    @claregrall good question!
    Mihai Capotă
    @mihaic
    Hello, @claregrall! Glad to hear you're making good use of BrianIAK. I believe we don't have an answer guide, but @manojneuro, @qihongl, or @CameronTEllis might know more.
    CameronTEllis
    @CameronTEllis
    Hi @claregrall, we do have an answer guide in the works, the main slow down is finding a way to distribute it that won't sabotage any classrooms that want to use these resources. However, if there are specific questions you have or guidance you need, I am happy to answer your questions directly.
    Clare Grall
    @claregrall
    @CameronTEllis This is great, thank you. I'm interested in checking my answers against the code given by the tutorial creators. This is because I'm still growing as a python coder, so there are coding tricks used in the tutorial code that I'm unfamiliar with and have subsequently learned (side benefit of the tutorials!). This also means that my responses are probably not optimized. Therefore, I don't have any specific questions, but I'd love to have some form of the answer guide available. How will I know when one is made available?
    Qihong Lu
    @qihongl
    Thanks, @claregrall and @mihaic for at-ing me. I think this is a great question. Do we have any plan to launching some sort of semi-supervised discussion forum (e.g. piazza)? I think aside from releasing some "answer keys", having a discussion forum would be useful. And I wouldn't even describe our scripts as "answer keys", since they are also unlikely to be optimal :)
    Mihai Capotă
    @mihaic
    @qihongl, our email list doubles as a forum:
    https://groups.google.com/forum/#!forum/brainiak
    CameronTEllis
    @CameronTEllis
    @claregrall We can certainly share with you how to get the answers when they are available. For your information though, the answer key likely won't explain how to code optimally or use any more tricks. In general, all of the questions were designed to use code that we provide in the tutorials but repurposed in some way to answer the question.
    Clare Grall
    @claregrall
    @CameronTEllis @qihongl haha, pardon my use of the word optimal- not the right word. I didn't mean to imply that this would be a python coding tutorial. Ultimately, I'd love to see an answer key if only for reinforcing learning and completeness.
    AbigailNovickHoskin
    @AbigailNovickHoskin

    Sorry if this isn't the right venue, but I have a question about brainiak searchlights :)

    I have epi data in a .nii file, a full brain mask in a .nii file, as well as an array of TR labels in a .npy file. I'm trying to run a searchlight, and I get a message saying my searchlight object has no attribute 'blocks'. What does blocks refer to? I'm assuming its # of runs? I only have 1 run, but 4 different stimuli presented within that run. Is there a place I need to specify this?

    Thanks so much for your help!

    File "/jukebox/cohen/abby/fMRIMultitasking/BrainiakSearchlight/BrainiakTest.py", line 184, in <module>
    sl_result = sl.run_searchlight(calc_svm, pool_size=pool_size)
    File "/jukebox/pkgs/PYGER/neu350/lib/python3.6/site-packages/brainiak/searchlight/searchlight.py", line 470, in run_searchlight
    pool_size)
    File "/jukebox/pkgs/PYGER/neu350/lib/python3.6/site-packages/brainiak/searchlight/searchlight.py", line 398, in run_block_function
    for idx, block in enumerate(self.blocks):
    AttributeError: 'Searchlight' object has no attribute 'blocks'

    Mihai Capotă
    @mihaic
    Hello, @AbigailNovickHoskin! This is certainly the right venue. Are you calling sl.distribute before sl.run_searchlight?
    AbigailNovickHoskin
    @AbigailNovickHoskin
    I am not!
    great catch, thanks Mihai :)
    Mihai Capotă
    @mihaic
    Glad to help. :) While the documentation is not very clear on that, the examples show it. The tutorial does an even better job:
    https://brainiak.org/notebooks/tutorials/html/07-searchlight.html#create_obj
    AbigailNovickHoskin
    @AbigailNovickHoskin
    Thanks again! Code seems to be running smoothly now :D
    Michael Shvartsman
    @mshvartsman
    Hi folks, I’m poking at the matrix-normal PR again and having issues installing brainiak on OSX. I ran into the command line tools issue someone mentioned above (solved by installing macOS_SDK_headers_for_macOS_10.14.pkg as instructed) but am now getting a different error: in the fcma cython compilation I’m getting an error brainiak/fcma/cython_blas.pyx:18:8: 'scipy/linalg/cython_blas.pxd' not found. This looks a lot like my environment is broken in some way but I’ve uninstalled/reinstalled scipy from both conda and pip and getting the same issue. I’ll try a totally fresh env but wondering if people have any other ideas.
    Michael Shvartsman
    @mshvartsman
    Note that I’m trying to use a conda env for this instead of homebrew python and virtualenv (but this is how I worked on brainiak stuff when I was at PNI and had no problems).
    Mihai Capotă
    @mihaic
    Hey, @mshvartsman! Nice to hear from you! I think we should start with a simple test to figure out what's wrong. Can you run pr-check.sh on the master branch?
    Michael Shvartsman
    @mshvartsman
    Ok, so outside of the fact that with newest conda you need conda activate instead of source activate and likewise for deactivate, things looked like they worked? Also some test errors in eventseg
    Michael Shvartsman
    @mshvartsman
    ok, looks like the pr_check env is fine, so I just cloned it as my main dev env, thanks for the help. No idea what I was doing wrong.
    (ah, and either source or conda is cool if you set up your shell correctly for conda, I was in zsh without having it set up correctly)
    Mihai Capotă
    @mihaic
    Good to hear. I have a branch that fixes the eventseg errors; waiting for an update to Jenkins to open a PR.
    SebastianSpeer
    @SebastianSpeer
    hello brainiak team
    I've been trying to install brainiak and use it in the Jupyter notebook. It seems to have installed fine via conda but when I try to import it, it can't find the module
    Do you have any suggestions as to how to fix this issue?
    I've also tried to install brainiak with pip
    but it also didnot work
    Mingbo Cai
    @lcnature
    Hi Sebastian. Can you show one example command of importing you have tried?
    Michael Shvartsman
    @mshvartsman
    Did you install brainiak in a conda environment that’s not your root env? Last time I checked, Jupyter doesn’t play nicely with conda envs by default. You need https://docs.anaconda.com/anaconda/user-guide/tasks/use-jupyter-notebook-extensions/#notebook-conda or another similar workaround.