- mne/parallel.py: return False early in _running_in_joblib_context()
  on emscripten; joblib parallel backends are unavailable in the browser
- mne/utils/config.py: catch Exception (not just ValueError) when
  loading the MNE config JSON; Pyodide's json parser raises SyntaxError
  on a corrupt or absent config file

Tutorials and examples that call interactive Qt backends (raw.plot(),
epochs.plot(), ica.plot_sources(), etc.) or depend on large datasets not
bundled in JupyterLite will hang or error in Pyodide. Wrap them with
sys.platform guards so they are skipped when running in the browser.

Interactive Qt plots (skip on emscripten):
- tutorials/intro/10_overview.py: raw.plot(), stc.plot()
- tutorials/intro/15_inplace.py: original_raw.plot(), rereferenced_raw.plot()
- tutorials/intro/20_events_from_raw.py: raw.copy().pick().plot(), raw.plot()
- tutorials/intro/40_sensor_locations.py: mne.viz.plot_alignment()
- tutorials/evoked/40_whitened.py: raw.plot(), epochs.plot()
- examples/preprocessing/muscle_ica.py: all ica.plot_* calls

Large datasets unavailable in the browser (raise RuntimeError on emscripten):
- tutorials/io/60_ctf_bst_auditory.py: BST auditory dataset (~2.9 GB)
- tutorials/io/70_reading_eyetracking_data.py: EyeLink misc dataset
- examples/visualization/eyetracking_plot_heatmap.py: EyeLink dataset

…erLite

- doc/conf.py: Fix lzma mock to use real stdlib lzma when available in
  Pyodide instead of LZMAFile=object which broke joblib's compressor
  registration
- 10_overview.py: Guard ica.plot_properties() which opens an interactive
  Qt window
- 15_inplace.py: Guard set_eeg_reference block which fails under
  Python 3.13 in Pyodide
- 20_events_from_raw.py: Guard STIM channel plot and all EEGLAB sections
  that require the unavailable testing dataset
- 40_sensor_locations.py: Guard ssvep dataset loading and sphere plot
  that require the unavailable ssvep dataset
- 50_configure_mne.py: Guard KIT test data loading whose test files are
  stripped from the Pyodide wheel
- 70_report.py: Skip Report.save() file-writing in browser, guard
  nibabel-dependent add_bem, 3D methods (add_trans/add_stc/add_forward/
  add_inverse_operator), missing ECG/events files, pandas-dependent
  make_metadata, and the HDF5 round-trip section

All intro tutorials (10, 15, 20, 30, 40, 50, 70) now run cleanly in
JupyterLite/Pyodide without errors.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

for more information, see https://pre-commit.ci

…Lite

for more information, see https://pre-commit.ci

This reverts commit 60fcc7f.

This reverts commit 825740e.

…ct issue

It asks make_field_map to subdivide the helmet mesh through VTK, and plot_field
needs the interactive viewer, so neither blocker is reachable from the browser.

plot_alignment and plot_bem locate their surfaces by probing the filesystem
before any reader runs, so the browser has to fetch the candidates up front.
Also hides the launch badge on 14 notebooks that need the Brain widget layer
or data too large to bundle.

read_raw_nirx and read_raw_egi are handed a folder rather than a file, so the
copy leaves a manifest beside it for the browser to walk. Files past a size
limit are skipped rather than bundled, and the seven notebooks that need
fsaverage, mne_bids or a multi-hundred-MB volume model are excluded instead.

…rcle full]

The curated lite_data archive only carries the files it was published with, so
anything added since went missing from the browser bundle without failing the
build. Prefer whatever CI already restored and fall back to the archive.

The tutorials no longer branch on sys.platform: the data those blocks skipped
is served now, and reading a KIT file or a BrainVision header goes through the
setup cell like everything else. 70_report keeps too much it cannot render, so
its badge is hidden instead.

The full build reports 379 MB and 251 MB for these, so the copy step skips
them and the badge would have had nothing to load.

Its raw alone is 344 MB, 404 MB with the forward and surfaces, which is more
than those six pages are worth on every docs deploy.

Most MNE readers validate a path through _check_fname before opening it, so
hooking that covers read_info, read_evokeds, read_cov and the rest at once;
read_label, read_epochs and read_raw_edf open directly and are wrapped
individually. Also serves talairach.xfm, the auditory/visual labels and the
EEGBCI runs CI provides, and installs mffpy and python-picard.

An oct-6 source space meant 8196 meshes and 8196 actors in a single scene,
which ran the browser tab out of memory. Also honor glyph_resolution/height
and solid_transform, so the MRI fiducials come out 5 mm rather than a metre.

The tutorials that build a renderer themselves reach for renderer.figure,
which the pyvista backend exposes alongside scene(). Honour a fig argument
too, so plot_alignment and plot_dipole_locations composite into one scene
instead of opening a second one.

…e full]

mne.gui.coregistration places fiducials by clicking the scalp, and the vtk.js
renderer has no picker, so that one cell cannot run. Swap it for a note via
sphinx-gallery's notebook_modification_function, which only touches the
JupyterLite copies, and the other nine cells keep working.

…ircle full]

conf.py asserts sphinx_gallery_conf is serializable and sphinx rejects function
objects, so passing the hook directly broke the build before Sphinx started.
sphinx-gallery imports a dotted path instead, so move it to sphinxext.

first_notebook_cell is applied when sphinx-gallery generates a notebook, so the
piplite cell also landed in the .ipynb offered for download, where it fails on
the first line. Prepend it at copy time instead, alongside the cell notes, so
only the JupyterLite copies carry it.

Twelve of the readers only need their file fetched before MNE opens it, and
each had its own six-line copy. One wrapper plus a table of readers does the
same job; the ones that need more than a fetch keep their own shim.