Welcome!

Welcome!#

CCN Software Workshop, Feb 2026#

We are excited to see everyone at the Flatiron Center for Computational Neuroscience workshop on using open source packages to analyze and visualize neural data! You should have received an email with logistical information, including the schedule and link to the slack channel where we will be communicating the workshop. If you did not receive this email, please let us know!

Over the course of this two-day workshop, we will walk you through the notebooks included on this site in order to demonstrate how to use pynapple and NeMoS to analyze and visualize your data.

Before the workshop, please try to follow the setup instructions below to install everything on your personal laptop. If you run into issues, first check the troubleshooting section on this page to see if there’s a solution for your problem and then, if you are unable to solve the problem, come to our installation help session at the Flatiron Institute, 160 5th Ave, in the 3rd floor conference center, before breakfast on Tuesday, February 3rd, 2026. We’ll be there starting at 8:30am!

We will use jupyter lab throughout the workshop. If you are unfamiliar with jupyter lab, please check out the “Intro to Jupyter Lab” page.

This website#

This website contains rendered versions of the notebooks we will be working through during this workshop. Several different versions of the notebooks are presented in the sidebar:

The Full notebooks section shows versions of the live coding notebook that include explanatory text, code, and plots. These are intended for you to check out after the workshop or to send to your labmates, so you can see everything that we covered.
The Group projects section contains the notebooks you will be working through during the group project sections of the workshop. They contain some pre-filled code, some notes, and a lot of problems that you will be solving as a group. If you follow the setup instructions below, you will have editable copies of these notebooks on your laptop, and you are expected to follow along using these notebooks.
The Limited live coding notebooks section contains two different versions of the live coding notebooks:
- For users contains the notebooks you attendees should be working though. These notebooks have some code pre-filled, as well as brief notes to help orient you. If you follow the setup instructions below, you will have editable copies of these notebooks on your laptop, and you are expected to follow along using these notebooks.
- For presenters is there to help if you miss something or fall behind. These notebooks includes the completed code blocks (along with some notes), so you can catch up.

We also have completed versions of the group projects notebooks; if you are interested, ask us after the workshop and we will send them to you.

Troubleshooting#

When creating the conda environment, you may get errors related to accessing the defaults channel. Due to recent changes to Anaconda’s Terms of Service, the Simons Foundation cannot use the defaults conda channel and it is blocked on all Flatiron Institute wireless networks. You need to specify conda-forge instead (which is community-managed and open to all). (The instructions above should do so.)
If you are on Mac and get an error related to ruamel.yaml (or clang) when running pip install -e ., we think this can be fixed by updating your Xcode Command Line Tools.
On Windows, you may receive an error saying “running scripts is disabled on this system” when trying to activate the virtual environment. If so, run Set-ExecutionPolicy -Scope CurrentUser and enter Unrestricted, then press Y. (You may have to do this every time you open powershell.)
If you have multiple jupyter installs on your path (because e.g., because you have an existing jupyter installation in a conda environment and you then used uv to setup the virtual environment for this workshop), jupyter can get confused. (You can check if this is the case by running which -a jupyter on Mac / Linux.) To avoid this problem, either make sure you only have one virtual environment active (e.g., by running conda deactivate) or prepend JUPYTER_DATA_DIR=$(realpath ..)/.venv/share/jupyter/ to your jupyter command above:
```
JUPYTER_DATA_DIR=$(realpath ..)/.venv/share/jupyter/ jupyter lab
```
(On Windows, replace $(realpath ..) with the path to the ccn-software-feb-2026 directory.)
We have noticed jupyter notebooks behaving a bit odd in Safari — if you are running/editing jupyter in Safari and the behavior seems off (scrolling not smooth, lag between creation and display of cells), try a different browser. We’ve had better luck with Firefox or using the arrow keys to navigate between cells.
On Windows + uv: When running the setup script, if you run into an error like: ImportError: DLL load failed while importing _jax: The specified module could not be found, it is possible that there is an issue with how uv is solving the environment. Try going through the installation with miniforge and conda instead.
On Windows + conda: if after installing conda the path are not correctly set, you may encounter this error message:
```
conda : The term 'conda' is not recognized as the name of a cmdlet, function, script file, or operable program. Check the spelling of the name, or if a path was included, verify that the path is correct and try again.
```
First, make sure you’re running all conda commands in either Miniforge Prompt or Anaconda Prompt, depending on which is installed. Otherwise, if you want to use powershell, you can try the following steps:
- Locate the path to the condabin folder. The path should look like: some-folder-path\Miniforge3\condabin.
  
  The following powershell command could be useful (note that i am starting form C: as a root, but you can change that):
```
Get-ChildItem -Path C:\ -Directory -Recurse -Filter "condabin" -ErrorAction SilentlyContinue | Select-Object -ExpandProperty FullName
```
- Temporarily add conda to the paths:
```
$env:Path += ";some-folder-path\Miniforge3\condabin"
```
- Initialize conda:
```
conda init powershell 
```
- Restart the powershell and check that conda is in the path. Run for example conda --version.
If you see sys:1: DeprecationWarning: Call to deprecated function (or staticmethod) _destroy. when running python scripts/setup.py, we don’t think this is actually a problem. As long as check_setup.py says everything looks good, you’re fine!
If you get an issue during the creation of the conda environment, try making sure the ~/.condarc file looks like the following (that file might not be created; create it if needed):
```
channels:
- conda-forge
ssl_verify: false
channel_priority: true
```
If you see an error that mentions SSL verification and are using conda, add the following line to your ~/.condarc: ssl_verify: false. Then restart your terminal and run the command again. (If you are using uv, I’m not sure how to set this configuration option.)
On an ARM-based (newer) Mac using conda, during check_setup.py, if you get This version of jaxlib was built using AVX instructions, uninstall using pip and install using conda: pip uninstall jax jaxlib; conda install -c conda-forge jax jaxlib.
- If not using conda, not sure how to avoid this issue, you may have to switch.
If you see a ModuleNotFoundError: No module named '_sqlite3' or ModuleNotFoundError: No module named pysqlite2 when running scripts/setup.py or jupyter lab, this is likely a sign that your python installation was not properly built. Reinstalling the python interpreter may help. How to do this depends on whether you are using uv or conda; come to the installation help session for assistance.

If you have an NVIDIA GPU and jax is configured to use it, when running the import blocks you may get a warning about being out of memory and then a big scary-looking error that says XlaRuntimeError: INTERNAL: No supported devices found for platofrm CUDA. This happens because the GPU has run out of memory: if you re-run the import block, it will run without a problem and then you can run the rest of the notebook (jax will just use the CPU and so be slightly slower). Alternatively, you can free up space on your GPU, probably by shutting down the kernels of notebooks you are no longer using (NOTE: this is not the same thing as just closing the notebook).
- By default, jax pre-allocates memory aggressively. We have tried to disable this behavior (by setting the environmental variable XLA_PYTHON_CLIENT_PREALLOCATE=false), but that depends on running import workshop_utils before import jax or import nemos (this has been done in our notebooks, but you may run into this issue on your own).
If you run an import cell in our notebooks and it hangs for a long time, and then you hit “Stop” and the error says something about permission denied, you have a permission issue with your virtual environment. You should either modify the permissions of the folder containing your environment variable (where this folder is and how this should be done depends on which tool you used and your Operating System), or run jupyter lab with elevated permissions:
- On Windows: Open the start menu and right click on “Anaconda Prompt” or “Powershell” (whichever you use) and click “Run as Administrator”. You may also need to open the browser as administrator.
- On Mac/Linux: run sudo jupyter lab instead of jupyter lab.

Binder#

A binder instance (a virtual environment running on Flatiron’s cluster) is provided in case we cannot get your installation working. To access it, click the “launch binder” button in the top left of this site or click here.

You must login with the email address you provided when registering for the workshop. If you get a 403 Forbidden error or would like to use a different email, send Billy Broderick a message on the workshop slack.

Some usage notes:

You are only allowed to have a single binder instance running at a time, so if you get the “already have an instance running error”, go to the binderhub page (or click on “check your currently running servers” on the right of the page) to join your running instance.
If you lose connection halfway through the workshop, go to the binderhub page to join your running instance rather than restarting the image.
This is important because if you restart the image, you will lose all data and progress.
The binder will be shutdown automatically after 1 day of inactivity or 7 days of total usage. Data will not persist after the binder instance shuts down, so please download any notebooks you want to keep.
I will destroy this instance in 2 weeks. You can download your notebooks to keep them after the fact.
You may run out of memory, see comment above.

Full notebooks

Group projects

Limited live coding notebooks

Welcome!

Contents

Welcome!#

CCN Software Workshop, Feb 2026#

This website#

Setup#

Create a virtual environment with python 3.12#

Install dependencies and setup notebooks#

Optional: Install pynaviz#

Troubleshooting#

Binder#