Issue with running test_physiology.py #26
Comments
|
Hello, A couple of quick questions to confirm your setup.
Also, the connectivity between T-stellate cells and bushy cells has not been measured quantitatively, so we did not include it in the connectivity table; for that you would need to have some idea of the divergence and convergence (in frequency and in terms of numbers of synapses), as well has the dynamics of the synapses. In this implementation, the bushy and tstellate cells receive their excitatory input from the auditory nerve, and, if you instantiate connectivity, from inhibitory tuberculoventral and D-Stellate cells. A name would be good. Kind Regards, --Paul |
|
Hi, Thank you for the answer. I tried what you suggested and some test failed.
So I think there is a problem with codes for mechanisms and making errors for synapses... Best regards, cnmodel/mechanisms/tests/test_mechanisms.py::test_max_open_probability FAILED [ 52%] cnmodel/synapses/tests/test_psd.py::test_sgc_bushy_psd PASSED [ 55%] cnmodel/synapses/tests/test_synapses.py::test_sgc_bushy FAILED [ 65%] cnmodel/util/tests/test_expfitting.py::test_fit1 PASSED [ 84%] __________________________ test_max_open_probability ___________________________ Comparing data array, shapes match: True |
|
Hi Jin, This appears to be a problem with either compilation of some of the .mod files or versions of some of the supporting libraries. I am unable to reproduce it - however I found several other issues with doing the installation. I have fixed these, and the installed version passes all the tests (and will push up later today after some more testing). You also might want to try the latest version that I am actively using at https://github.com/pbmanis/cnmodel.git (python3 branch). This version works with Python3.9 and the latest Neuron (8.0). It wil also help if you tell me which platform (linux, macosx, windows) you are using, and if you are doing the installation with anaconda or with a virtual environment. I almost exclusively use the venv approach, and have not attempted the conda installation in some time. I also use unix/macosx, and haven't verified the windows installation (which uses a batch file to create the virtual environment) in several months. When using the unix-systems based virtual environment install (make_env.sh/requirements.txt), errors during install will appear in red text on the screen, although the installation will proceed beyond that. The cause of the error has to be fixed first, and the installation restarted after removing the virtual environment folder. In the current version on the cnmodel site, there are a few errors in the requirements file that will break the installation (as mentioned above). Thanks, --Paul |
|
Hi Paul, I am using Ubuntu and conda. I tried the code in pbmanis/cnmodel.git, and I am getting same issues. But when I ran the setup.py, it seems like there is an error in the end. You can find the whole output from setup.py after the output from nrninvmodl command. Thank you for your help! Best regards, This is the output from setup.py... |
|
Hi Jin, A couple of things:
I don't have time right now to figure out why the test results are different, but will look into it. Meanwhile, you should be able to get this version working, and then build your models. As I noted before, the design is that the code in cnmodel itself is not changed, but that cnmodel is imported using a project in an "adjacent" directory, which might start from something like the code in examples/test_physiology.py; when you run this you should be sure you have activated the cnmodel_venv first. If you need additional modules (say, plotly instead of matplotlib), you should either add them to the requirements.txt file, or make copy the make_env.sh and requirements.txt to your working (adjacent) directory, and modify those instead of modifying the ones in cnmodel. --Paul |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi!
I am trying to see how tstellate cells are activated with stimuli with example file test_physiology.py.
I added two lines of code:
tstellate_cell_ids = self.tstellate.select(cells_per_band, cf=f, create=True)
self.tstellate.resolve_inputs(depth=2)
as shown in below:
However, it seems like tstellate cell is not firing at all.
I also noticed that the table in data/connectivity.py file has 0 values for connection from tstellate to bushy. Does this mean bushy cells dont get any input from tstellate cells with current configuration?
Still, I would expect that tstellate cells would fire as they get input from other types of cells.
Could you help me to tackle this problem?
I wonder if I am missing something or if its simply because tstellate cells are not getting enough input to fire....
The text was updated successfully, but these errors were encountered: