MAINT: Harmonize dependencies (#433)

.circleci/config.yml

@@ -10,8 +10,10 @@ _xvfb: &xvfb
 
 jobs:
     build_docs:
-      machine:
-        image: ubuntu-2004:202111-01
+      docker:
+        # Use 18.04 rather than 20.04 because MESA 20.0.8 on 18.04 has working
+        # transparency but 21.0.3 on 20.04 does not!
+        - image: cimg/base:stable-18.04
       steps:
         - checkout
         - run:
@@ -32,31 +34,33 @@ jobs:
             name: Set BASH_ENV
             command: |
               set -e
-              python3 -m venv ~/python_env --upgrade-deps
+              sudo apt update -qq
+              sudo apt install -qq libosmesa6 libglx-mesa0 libopengl0 libglx0 libdbus-1-3 \
+                libxkbcommon-x11-0 libxcb-icccm4 libxcb-image0 libxcb-keysyms1 libxcb-randr0 \
+                libxcb-render-util0 libxcb-shape0 libxcb-xfixes0 libxcb-xinerama0 \
+                graphviz optipng \
+                python3.8-venv python3-venv \
+                xvfb libxft2 ffmpeg
+              python3.8 -m venv ~/python_env
               echo "set -e" >> $BASH_ENV
-              echo "export DISPLAY=:99" >> $BASH_ENV
               echo "export OPENBLAS_NUM_THREADS=4" >> $BASH_ENV
               echo "export XDG_RUNTIME_DIR=/tmp/runtime-circleci" >> $BASH_ENV
-              source tools/get_minimal_commands.sh
+              echo "export MNE_FULL_DATE=true" >> $BASH_ENV
               echo "export MNE_3D_OPTION_ANTIALIAS=false" >> $BASH_ENV
-              echo "export MNE_3D_BACKEND=pyvista" >> $BASH_ENV
+              source tools/get_minimal_commands.sh
+              echo "export MNE_3D_BACKEND=pyvistaqt" >> $BASH_ENV
               echo "export PATH=~/.local/bin/:$PATH" >> $BASH_ENV
               echo "source ~/python_env/bin/activate" >> $BASH_ENV
               mkdir -p ~/.local/bin
               ln -s ~/python_env/bin/python ~/.local/bin/python
               echo "BASH_ENV:"
               cat $BASH_ENV
               mkdir -p ~/mne_data
-              touch pattern.txt;
-
+              touch pattern.txt
         - run:
-            name: Install 3D rendering libraries \ PyQt5 dependencies \ graphviz \ optipng (for optimized images)
+            name: check neuromag2ft
             command: |
-              sudo apt update
-              sudo apt install libosmesa6 libglx-mesa0 libopengl0 libglx0 libdbus-1-3 \
-                  libxkbcommon-x11-0 libxcb-icccm4 libxcb-image0 libxcb-keysyms1 libxcb-randr0 libxcb-render-util0 libxcb-shape0 libxcb-xfixes0 libxcb-xinerama0 \
-                  graphviz \
-                  optipng
+              neuromag2ft --version
 
         - run:
             <<: *xvfb
@@ -83,7 +87,7 @@ jobs:
         - run:
             name: Get Python running
             command: |
-              python -m pip install --upgrade --progress-bar off pip setuptools
+              python -m pip install --upgrade pip setuptools wheel
               python -m pip install --upgrade --progress-bar off --pre sphinx
               python -m pip install --upgrade --progress-bar off -r requirements.txt -r requirements_testing.txt -r requirements_doc.txt
               python -m pip install -e .

.github/workflows/linux_pip.yml

@@ -15,7 +15,7 @@ concurrency:
 jobs:
   # PIP + non-default stim channel + log level info
   job:
-    name: 'py3.8'
+    name: 'py3.10'
     runs-on: ubuntu-20.04
     if: "!contains(github.event.head_commit.message, '[skip tests]')"
     defaults:

environment.yml

@@ -7,38 +7,41 @@ dependencies:
 - numpy
 - scipy
 - matplotlib
+- tqdm
+- pooch>=1.5
+- decorator
+- h5io
+- packaging
 - numba
-- pandas==1.3.2
+- pandas
 - xlrd
 - scikit-learn
 - h5py
-- h5io
+- jinja2
 - pillow
 - statsmodels
 - jupyter
 - joblib
 - psutil
 - numexpr
 - imageio
-- tqdm
 - spyder-kernels>=1.10.0
 - imageio-ffmpeg>=0.4.1
 - vtk>=9.0.1
-- pyvista>=0.30
+- traitlets
+- pyvista>=0.32
 - pyvistaqt>=0.4
 - qdarkstyle
 - darkdetect
 - dipy
 - nibabel
+- nilearn
 - python-picard
 - pyqt!=5.15.3
 - mffpy>=0.5.7
 - ipywidgets
-- lxml
-- pytables
-- pooch
-- nilearn
 - ipyvtklink
 - mne-qt-browser
+- pymatreader
 - pip:
   - https://github.com/mne-tools/mne-python/archive/main.zip

examples/general/plot_11_hrf_measured.py

@@ -7,10 +7,12 @@
 In this example we analyse data from a real multichannel
 functional near-infrared spectroscopy (fNIRS)
 experiment (see :ref:`tut-fnirs-hrf-sim` for a simplified simulated
-analysis). The experiment consists of three conditions
-1) tapping with the left hand,
-2) tapping with the right hand,
-3) a control condition where the participant does nothing.
+analysis). The experiment consists of three conditions:
+
+1. tapping with the left hand,
+2. tapping with the right hand,
+3. a control condition where the participant does nothing.
+
 We use a GLM analysis to examine the neural activity associated with
 the different tapping conditions.
 An alternative epoching style analysis on the same data can be
@@ -27,12 +29,12 @@
    :local:
    :depth: 2
 
-.. note:: Parts of this tutorial require the latest development version of MNE-Python. See these instructions for  
+.. note:: Parts of this tutorial require the latest development version of MNE-Python. See these instructions for
           `how to upgrade <https://mne.tools/dev/install/updating.html>`__.
           But basically boils down to running
           ``pip install -U --no-deps https://github.com/mne-tools/mne-python/archive/main.zip``.
           Sections of the code that require this version will be noted below.
-          
+
 """
 # sphinx_gallery_thumbnail_number = 9
 
@@ -98,10 +100,12 @@
 # Next we update the annotations by assigning names to each trigger ID.
 # Then we crop the recording to the section containing our
 # experimental conditions.
-
+#
+# Because of limitations with ``nilearn``, we use ``'_'`` to separate conditions
+# rather than the standard ``'/'``.
 raw_intensity.annotations.rename({'1.0': 'Control',
-                                  '2.0': 'Tapping/Left',
-                                  '3.0': 'Tapping/Right'})
+                                  '2.0': 'Tapping_Left',
+                                  '3.0': 'Tapping_Right'})
 raw_intensity.annotations.delete(raw_intensity.annotations.description == '15.0')
 raw_intensity.annotations.set_durations(5)
 
@@ -166,7 +170,7 @@
 # .. sidebar:: Relevant literature
 #
 #    For further discussion on design matrices see
-#    the Nilearn examples. Specifically the 
+#    the Nilearn examples. Specifically the
 #    `first level model example <http://nilearn.github.io/auto_examples/04_glm_first_level/plot_first_level_details.html>`_.
 #
 # Next we create a model to fit our data to.
@@ -221,7 +225,7 @@
 # Examine expected response
 # -------------------------
 #
-# The matrices above can be a bit abstract as they encompase multiple 
+# The matrices above can be a bit abstract as they encompase multiple
 # conditions and regressors.
 # Instead we can examine a single condition.
 # Here we observe the boxcar function for a single condition,
@@ -239,7 +243,7 @@
 
 s = mne_nirs.experimental_design.create_boxcar(raw_intensity, stim_dur=5.0)
 plt.plot(raw_intensity.times, s[:, 1])
-plt.plot(design_matrix['Tapping/Left'])
+plt.plot(design_matrix['Tapping_Left'])
 plt.xlim(180, 300)
 plt.legend(["Stimulus", "Expected Response"])
 plt.xlabel("Time (s)")
@@ -336,7 +340,7 @@
 # negative of HbO as expected.
 
 glm_est = run_glm(raw_haemo, design_matrix)
-glm_est.plot_topo(conditions=['Tapping/Left', 'Tapping/Right'])
+glm_est.plot_topo(conditions=['Tapping_Left', 'Tapping_Right'])
 
 
 # %%
@@ -364,7 +368,7 @@
 fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(10, 6), gridspec_kw=dict(width_ratios=[0.92, 1]))
 
 glm_hbo = glm_est.copy().pick(picks="hbo")
-conditions = ['Tapping/Right']
+conditions = ['Tapping_Right']
 
 glm_hbo.plot_topo(axes=axes[0], colorbar=False, conditions=conditions)
 
@@ -380,7 +384,7 @@
 # Another way to view the data is to project the GLM estimates to the nearest
 # cortical surface
 
-glm_est.copy().surface_projection(condition="Tapping/Right", view="dorsal", chroma="hbo")
+glm_est.copy().surface_projection(condition="Tapping_Right", view="dorsal", chroma="hbo")
 
 
 # %%
@@ -412,7 +416,7 @@
 groups = dict(Left_ROI=picks_pair_to_idx(raw_haemo, left),
               Right_ROI=picks_pair_to_idx(raw_haemo, right))
 
-conditions = ['Control', 'Tapping/Left', 'Tapping/Right']
+conditions = ['Control', 'Tapping_Left', 'Tapping_Right']
 
 df = glm_est.to_dataframe_region_of_interest(groups, conditions)
 
@@ -439,7 +443,7 @@
 contrast_matrix = np.eye(design_matrix.shape[1])
 basic_conts = dict([(column, contrast_matrix[i])
                    for i, column in enumerate(design_matrix.columns)])
-contrast_LvR = basic_conts['Tapping/Left'] - basic_conts['Tapping/Right']
+contrast_LvR = basic_conts['Tapping_Left'] - basic_conts['Tapping_Right']
 
 contrast = glm_est.compute_contrast(contrast_LvR)
 contrast.plot_topo()
@@ -474,8 +478,8 @@
 # the tapping, but we do expect 5% or less for the false positive rate.
 
 (df
- .query('Condition in ["Control", "Tapping/Left", "Tapping/Right"]')
- .groupby(['Condition', 'Chroma'])
+ .query('Condition in ["Control", "Tapping_Left", "Tapping_Right"]')
+ .drop(['df', 'mse', 'p_value', 't'], axis=1)
+ .groupby(['Condition', 'Chroma', 'ch_name'])
  .agg(['mean'])
- .drop(['df', 'mse', 'p_value', 't'], 1)
  )