Jupyter notebook

[arnodelorme]

Check figure compatibilty

[claude]

🤖 Not actionable as written — need more detail before I can proceed.

The title "Jupyter notebook" and body "Check figure compatibility" don't specify what's broken, where to look, or how to verify a fix. Multiple plausible interpretations:

1. Interactive figures produced by EEGPrep pop_* plotting/GUI functions (e.g. topoplot, ICA viewers, eegplot-style scrollers) should display and remain interactive inside a Jupyter notebook rather than opening a separate Qt window or crashing.
2. The checked-in notebooks under sample_notebooks/ (pipeline.ipynb, oct2pytest.ipynb, playground2.ipynb, sandbox.ipynb, test.ipynb, test_mne_iclabel.ipynb) don't render figures correctly when re-run.
3. Sphinx-executed notebook examples in docs/source/ produce broken figures.
4. Something about matplotlib backend selection (qtagg vs inline vs widget) when EEGPrep is imported in a kernel.

To pick the right fix, please share:

• Which figure(s) — a specific function (pop_topoplot, pop_selectcomps, eegplot, pop_prop_extended, …) or a specific notebook path.
• Which notebook environment — classic Jupyter, JupyterLab, VS Code, Colab; and which matplotlib backend (%matplotlib inline, widget, qt).
• What "compatibility" means here — figure doesn't render, renders blank, crashes the kernel, opens a Qt window instead of inline, loses interactivity, differs from EEGLAB output, fails on notebook re-execution, etc.
• Reproduction — minimal cell(s) that trigger the problem, expected vs actual output, and (if applicable) a screenshot of the broken figure.
• Definition of done — one figure fixed, all sample_notebooks/ runnable end-to-end, a new smoke test, docs updates, or something else.

Holding off on any PR until scope is clarified.

[innaamogolonova]

Jupyter notebook compatibility checked with uv run --with jupyterlab jupyter lab.

Jupyter notebook compatibility as expected:

[arnodelorme]

Investigate spectopo