Hi,
In nestml/doc/tutorials/stdp_windows notebook, in cell 3, I see an example of fetching model name, neuron name, and synapse name from nestml/models/neurons/iaf_psc_delta_neuron.nestml, which appears to have only model inside of it. It tries to extract the names from a format of
1. model <name>: 2. neuron <name>: 3. synapse <name>:
But the iaf_psc_delta_neuron.nestml file from the latest build doesn't have any structure as neuron <name>: to my best knowledge. Please let me know if there is any fix for this problem.
I went ahead and tried to edit the generate_code_for function from NESTCodeGeneratorUtils class. It can be found here https://pastebin.com/Qz8JwinA. But it gives an error as follows:
Best,
Ashraf Ul Islam Shihab
Graduate Research Assistant
Department of Electrical and Computer Engineering Boise State University (BSU)
Boise, Idaho 83725
Email: ashrafshihab@u.boisestate.edu
Hi,
Thanks for your inquiry! I'm not sure from your description what the actual problem is, you are running the stdp_windows tutorial notebook, and there is an error? Could you post a full log?
It would be great if you could use the latest NESTML version from GitHub (try running python3 -m pip install git+https://github.com/nest/nestml/).
Note that the notebook (in cell 3) contains the path "../../../models/neurons/iaf_psc_delta_neuron.nestml", so you should first change directory (cd) to "doc/tutorials/stdp_windows" and run Jupyter Notebook or Python from that directory. There is no need to change any internal code inside the generate_code_for() function.
For the error importing the "nest" module in Python, please make sure that the environment is set up correctly before starting the Jupyter Notebook server. Typically you can do this by sourcing nest_vars.sh (see bottom of this page https://nest-simulator.readthedocs.io/en/stable/installation/noenv_install.html#noenv), which sets up the correct PYTHONPATH environment variable. How did you install NEST?
With kind regards, Charl
On Wed, May 8, 2024, at 00:54, Ashraf Ul Islam Shihab wrote:
Hi,
I installed NEST using the following code.
import condacolab
condacolab.install() !conda install -c conda-forge nest-simulator
import nest
nest.__version__
!pip install nestml
import pynestml pynestml.__version__
Then, I manually imported the path to iaf_psc_delta_neuron.nestml. It gave me this error:
---------------------------------------------------------------------------
IndexError Traceback (most recent call last)
<ipython-input-10-7981312f3233> https://localhost:8080/# in <cell line: 1>()----> 1 module_name, neuron_model_name, synapse_model_name = NESTCodeGeneratorUtils.generate_code_for( 2 "/content/iaf_psc_delta_neuron.nestml", 3 nestml_stdp_model, 4 post_ports=["post_spikes"]) 5
/usr/local/lib/python3.10/site-packages/pynestml/codegeneration/nest_code_generator_utils.py https://localhost:8080/# in generate_code_for(cls, nestml_neuron_model, nestml_synapse_model, module_name, target_path, post_ports, mod_ports, logging_level) 87 88 # update neuron model name inside the file---> 89 neuron_model_name = re.findall(r"neuron [^:\s]*:", nestml_neuron_model)[0][7:-1] 90 neuron_fn = neuron_model_name + ".nestml" 91 with open(neuron_fn, "w") as f:
IndexError: list index out of range
This is on NEST version 3.7.0 and NESTML version 7.0.2.
Best,
Ashraf
On Wed, May 8, 2024 at 4:59 AM Charl Linssen nest-users@turingbirds.com wrote:
Hi, sorry, but this is turning out to be a little bit tricky.
First, there is a problem with the NEST environment: when NESTML runs and tries to query the NEST version, the correct $PYTHONPATH is not picked up so there is an error when it tries to run "import nest" from a subprocess.
I manually worked around that, but then there is an error with the build step: "The CMAKE_CXX_COMPILER: /usr/local/bin/x86_64-conda-linux-gnu-c++ is not a full path to an existing compiler tool." I believe this is a known issue with running via conda. I was not able to bypass this error by overriding the CMake flags.
These issues might take a considerable time to solve on our side. Would it be possible for you to use an alternative way of running NEST/NESTML? Personally I often use the Docker containers for deployment. Possibly, you could also obtain access to the European EBRAINS cloud via your institution.
Sorry for the difficulty, but we do not have experience using Google Collab for running NEST/NESTML and a lot of workarounds are needed. Steffen: as you are maintaining the Anaconda packages, do you have any thoughts on this issue? Why doesn't everything (including the shell environment) run from inside the conda environment? I thought everything in our Anaconda distribution is set up right for using NESTML.
Cheers! Charl
On Wed, May 8, 2024, at 20:01, Ashraf Ul Islam Shihab wrote:
Hi Linssen,
I used the EBRAINS cloud platform to implement the same thing. It still gives me the same error.
Best,
Ashraf
[image: image.png]
On Thu, May 9, 2024 at 3:15 AM Charl Linssen nest-users@turingbirds.com wrote:
-
Ashraf Ul Islam Shihab -
Charl Linssen