Merge remote-tracking branch 'upstream/main' into trame

doc/conf.py

@@ -7,6 +7,7 @@
 from datetime import datetime, timezone
 import faulthandler
 import gc
+from importlib.metadata import metadata
 import os
 import subprocess
 import sys
@@ -1431,18 +1432,7 @@ def reset_warnings(gallery_conf, fname):
 
 # -- Dependency info ----------------------------------------------------------
 
-try:
-    from importlib.metadata import metadata  # new in Python 3.8
-
-    min_py = metadata("mne")["Requires-Python"]
-except ModuleNotFoundError:
-    from pkg_resources import get_distribution
-
-    info = get_distribution("mne").get_metadata_lines("PKG-INFO")
-    for line in info:
-        if line.strip().startswith("Requires-Python"):
-            min_py = line.split(":")[1]
-min_py = min_py.lstrip(" =<>")
+min_py = metadata("mne")["Requires-Python"].lstrip(" =<>")
 rst_prolog += f"\n.. |min_python_version| replace:: {min_py}\n"
 
 # -- website redirects --------------------------------------------------------

doc/install/check_installation.rst

@@ -16,26 +16,41 @@ a terminal::
 This should display some system information along with the versions of
 MNE-Python and its dependencies. Typical output looks like this::
 
-    Platform:      Linux-5.0.0-1031-gcp-x86_64-with-glibc2.2.5
-    Python:        3.8.1 (default, Dec 20 2019, 10:06:11)  [GCC 7.4.0]
-    Executable:    /home/travis/virtualenv/python3.8.1/bin/python
-    CPU:           x86_64: 2 cores
-    Memory:        7.8 GB
-
-    mne:           0.21.dev0
-    numpy:         1.19.0.dev0+8dfaa4a {blas=openblas, lapack=openblas}
-    scipy:         1.5.0.dev0+f614064
-    matplotlib:    3.2.1 {backend=QtAgg}
-
-    sklearn:       0.22.2.post1
-    numba:         0.49.0
-    nibabel:       3.1.0
-    cupy:          Not found
-    pandas:        1.0.3
-    dipy:          1.1.1
-    pyvista:       0.25.2 {pyvistaqt=0.1.0}
-    vtk:           9.0.0
-    qtpy:          2.0.1 {PySide6=6.2.4}
+    Platform                Windows-10-10.0.20348-SP0
+    Python                  3.10.12 | packaged by conda-forge | (main, Jun 23 2023, 22:34:57) [MSC v.1936 64 bit (AMD64)]
+    Executable              C:\Miniconda3\envs\mne\python.exe
+    CPU                     Intel64 Family 6 Model 85 Stepping 7, GenuineIntel (2 cores)
+    Memory                  7.0 GB
+
+    Core
+    ├☑ mne                  1.6.0.dev67+gb12384562
+    ├☑ numpy                1.25.2 (OpenBLAS 0.3.23.dev with 1 thread)
+    ├☑ scipy                1.11.2
+    ├☑ matplotlib           3.7.2 (backend=QtAgg)
+    ├☑ pooch                1.7.0
+    └☑ jinja2               3.1.2
+
+    Numerical (optional)
+    ├☑ sklearn              1.3.0
+    ├☑ nibabel              5.1.0
+    ├☑ nilearn              0.10.1
+    ├☑ dipy                 1.7.0
+    ├☑ openmeeg             2.5.6
+    ├☑ pandas               2.1.0
+    └☐ unavailable          numba, cupy
+
+    Visualization (optional)
+    ├☑ pyvista              0.41.1 (OpenGL 3.3 (Core Profile) Mesa 10.2.4 (git-d92815a) via Gallium 0.4 on llvmpipe (LLVM 3.4, 256 bits))
+    ├☑ pyvistaqt            0.0.0
+    ├☑ ipyvtklink           0.2.2
+    ├☑ vtk                  9.2.6
+    ├☑ qtpy                 2.4.0 (PyQt5=5.15.8)
+    ├☑ ipympl               0.9.3
+    ├☑ pyqtgraph            0.13.3
+    └☑ mne-qt-browser       0.5.2
+
+    Ecosystem (optional)
+    └☐ unavailable          mne-bids, mne-nirs, mne-features, mne-connectivity, mne-icalabel, mne-bids-pipeline
 
 
 .. dropdown:: If you get an error...

mne/commands/mne_setup_source_space.py

@@ -93,8 +93,7 @@ def run():
         "--add-dist",
         dest="add_dist",
         help='Add distances. Can be "True", "False", or "patch" '
-        "to only compute cortical patch statistics (like the "
-        "--cps option in MNE-C; requires SciPy >= 1.3)",
+        "to only compute cortical patch statistics (like the --cps option in MNE-C)",
         default="True",
     )
     parser.add_option(

mne/coreg.py

@@ -304,8 +304,8 @@ def _decimate_points(pts, res=10):
     mids = np.c_[x, y, z] + res / 2.0
 
     # each point belongs to at most one voxel center, so figure those out
-    # (cKDTree faster than BallTree for these small problems)
-    tree = _DistanceQuery(mids, method="cKDTree")
+    # (KDTree faster than BallTree for these small problems)
+    tree = _DistanceQuery(mids, method="KDTree")
     _, mid_idx = tree.query(pts)
 
     # then figure out which to actually use based on proximity

mne/morph.py

@@ -116,7 +116,7 @@ def compute_source_morph(
         Number of iterations for the smoothing of the surface data.
         If None, smooth is automatically defined to fill the surface
         with non-zero values. Can also be ``'nearest'`` to use the nearest
-        vertices on the surface (requires SciPy >= 1.3).
+        vertices on the surface.
 
         .. versionchanged:: 0.20
            Added support for 'nearest'.

mne/morph_map.py

@@ -223,7 +223,7 @@ def _make_morph_map_hemi(subject_from, subject_to, subjects_dir, reg_from, reg_t
     _normalize_vectors(to_rr)
 
     # from surface: get nearest neighbors, find triangles for each vertex
-    nn_pts_idx = _compute_nearest(from_rr, to_rr, method="cKDTree")
+    nn_pts_idx = _compute_nearest(from_rr, to_rr, method="KDTree")
     from_pt_tris = _triangle_neighbors(from_tri, len(from_rr))
     from_pt_tris = [from_pt_tris[pt_idx].astype(int) for pt_idx in nn_pts_idx]
     from_pt_lens = np.cumsum([0] + [len(x) for x in from_pt_tris])

mne/source_space/_source_space.py

@@ -1508,7 +1508,7 @@ def setup_source_space(
     add_dist : bool | str
         Add distance and patch information to the source space. This takes some
         time so precomputing it is recommended. Can also be 'patch' to only
-        compute patch information (requires SciPy 1.3+).
+        compute patch information.
 
         .. versionchanged:: 0.20
            Support for ``add_dist='patch'``.
@@ -2575,7 +2575,7 @@ def _filter_source_spaces(surf, limit, mri_head_t, src, n_jobs=None, verbose=Non
                 idx = idx[mask]
                 check_r1s = check_r1s[mask]
             dists = _compute_nearest(
-                surf["rr"], check_r1s, return_dists=True, method="cKDTree"
+                surf["rr"], check_r1s, return_dists=True, method="KDTree"
             )[1]
             close = dists < limit / 1000.0
             omit_limit = np.sum(close)

mne/surface.py

@@ -570,7 +570,7 @@ class _DistanceQuery:
     """Wrapper for fast distance queries."""
 
     def __init__(self, xhs, method="BallTree", allow_kdtree=False):
-        assert method in ("BallTree", "cKDTree", "cdist")
+        assert method in ("BallTree", "KDTree", "cdist")
 
         # Fastest for our problems: balltree
         if method == "BallTree":
@@ -581,7 +581,7 @@ def __init__(self, xhs, method="BallTree", allow_kdtree=False):
                     "Nearest-neighbor searches will be significantly "
                     "faster if scikit-learn is installed."
                 )
-                method = "cKDTree"
+                method = "KDTree"
             else:
                 self.query = partial(
                     _safe_query,
@@ -590,18 +590,11 @@ def __init__(self, xhs, method="BallTree", allow_kdtree=False):
                     return_distance=True,
                 )
 
-        # Then cKDTree
-        if method == "cKDTree":
-            try:
-                from scipy.spatial import cKDTree
-            except ImportError:
-                method = "cdist"
-            else:
-                self.query = cKDTree(xhs).query
+        # Then KDTree
+        if method == "KDTree":
+            from scipy.spatial import KDTree
 
-        # KDTree is really only faster for huge (~100k) sets,
-        # (e.g., with leafsize=2048), and it's slower for small (~5k)
-        # sets. We can add it later if we think it will help.
+            self.query = KDTree(xhs).query
 
         # Then the worst: cdist
         if method == "cdist":
@@ -1408,7 +1401,7 @@ def _decimate_surface_sphere(rr, tris, n_triangles):
     sphere_rr, _ = read_surface(qsphere)
     norms = np.linalg.norm(sphere_rr, axis=1, keepdims=True)
     sphere_rr /= norms
-    idx = _compute_nearest(sphere_rr, ico_surf["rr"], method="cKDTree")
+    idx = _compute_nearest(sphere_rr, ico_surf["rr"], method="KDTree")
     n_dup = len(idx) - len(np.unique(idx))
     if n_dup:
         raise RuntimeError(

mne/tests/test_morph.py

@@ -35,7 +35,7 @@
 from mne.minimum_norm import apply_inverse, read_inverse_operator, make_inverse_operator
 from mne.source_space._source_space import _add_interpolator, _grid_interp
 from mne.transforms import quat_to_rot
-from mne.utils import check_version, catch_logging, _record_warnings
+from mne.utils import catch_logging, _record_warnings
 
 # Setup paths
 
@@ -244,10 +244,6 @@ def test_surface_source_morph_round_trip(smooth, lower, upper, n_warn, dtype):
     if dtype is complex:
         stc.data = 1j * stc.data
         assert_array_equal(stc.data.real, 0.0)
-    if smooth == "nearest" and not check_version("scipy", "1.3"):
-        with pytest.raises(ValueError, match="required to use nearest"):
-            morph = compute_source_morph(stc, "sample", "fsaverage", **kwargs)
-        return
     with _record_warnings() as w:
         morph = compute_source_morph(stc, "sample", "fsaverage", **kwargs)
     w = [ww for ww in w if "vertices not included" in str(ww.message)]

mne/tests/test_surface.py

@@ -105,15 +105,15 @@ def test_compute_nearest():
     y = x[nn_true]
 
     nn1 = _compute_nearest(x, y, method="BallTree")
-    nn2 = _compute_nearest(x, y, method="cKDTree")
+    nn2 = _compute_nearest(x, y, method="KDTree")
     nn3 = _compute_nearest(x, y, method="cdist")
     assert_array_equal(nn_true, nn1)
     assert_array_equal(nn_true, nn2)
     assert_array_equal(nn_true, nn3)
 
     # test distance support
     nnn1 = _compute_nearest(x, y, method="BallTree", return_dists=True)
-    nnn2 = _compute_nearest(x, y, method="cKDTree", return_dists=True)
+    nnn2 = _compute_nearest(x, y, method="KDTree", return_dists=True)
     nnn3 = _compute_nearest(x, y, method="cdist", return_dists=True)
     assert_array_equal(nnn1[0], nn_true)
     assert_array_equal(nnn1[1], np.zeros_like(nn1))  # all dists should be 0

mne/viz/utils.py

@@ -2838,7 +2838,7 @@ def _generate_default_filename(ext=".png"):
 
 
 def _prop_kw(kind, val):
-    # Can be removed in when we depend on matplotlib 3.4.3+
+    # Can be removed in when we depend on matplotlib 3.5+
     # https://github.com/matplotlib/matplotlib/pull/20585
     from matplotlib.widgets import SpanSelector
 
@@ -2863,39 +2863,17 @@ def _handle_precompute(precompute):
 def _set_3d_axes_equal(ax):
     """Make axes of 3D plot have equal scale on all dimensions.
 
-    This way spheres appear as actual spheres, cubes as cubes, etc..
-    This is one possible solution to Matplotlib's ``ax.set_aspect('equal')``
-    and ``ax.axis('equal')`` not working for 3D.
+    This way spheres appear as actual spheres, cubes as cubes, etc.
 
     Parameters
     ----------
     ax: matplotlib.axes.Axes
         A matplotlib 3d axis to use.
-
-    Notes
-    -----
-    modified from:
-    https://stackoverflow.com/q/13685386
-
-    Should no longer be necessary for matplotlib >= 3.3.0:
-    https://matplotlib.org/stable/users/prev_whats_new/whats_new_3.3.0.html#axes3d-no-longer-distorts-the-3d-plot-to-match-the-2d-aspect-ratio
     """
-    x_lim, y_lim, z_lim = ax.get_xlim3d(), ax.get_ylim3d(), ax.get_zlim3d()
-
-    def get_range(lim):
-        return lim[1] - lim[0], np.mean(lim)
-
-    x_range, x_mean = get_range(x_lim)
-    y_range, y_mean = get_range(y_lim)
-    z_range, z_mean = get_range(z_lim)
-
-    # The plot bounding box is a sphere in the sense of the infinity
-    # norm, hence I call half the max range the plot radius.
-    plot_radius = 0.5 * max([x_range, y_range, z_range])
-
-    ax.set_xlim3d([x_mean - plot_radius, x_mean + plot_radius])
-    ax.set_ylim3d([y_mean - plot_radius, y_mean + plot_radius])
-    ax.set_zlim3d([z_mean - plot_radius, z_mean + plot_radius])
+    ranges = tuple(
+        np.abs(np.diff(getattr(ax, f"get_{d}lim")())).item() for d in ("x", "y", "z")
+    )
+    ax.set_box_aspect(ranges)
 
 
 def _check_type_projs(projs):

requirements.txt

@@ -1,7 +1,7 @@
 # requirements for full MNE-Python functionality (other than raw/epochs export)
 numpy>=1.15.4
-scipy>=1.6.3
-matplotlib>=3.4.0
+scipy>=1.7.1
+matplotlib>=3.4.3
 tqdm
 pooch>=1.5
 decorator

tutorials/intro/40_sensor_locations.py

@@ -17,9 +17,6 @@
 import matplotlib.pyplot as plt
 import numpy as np
 
-# the following import is required for matplotlib < 3.2:
-from mpl_toolkits.mplot3d import Axes3D  # noqa
-
 import mne
 
 # %%