brainiak · mihaic · Oct 3, 2023 · Sep 3, 2021 · Jan 24, 2022 · Jan 24, 2022
diff --git a/.conda/build.sh b/.conda/build.sh
@@ -3,7 +3,7 @@
 # Install from PyPI because there is no current conda package for the
 # following. Explicitly install dependencies with no conda package as well
 # because otherwise conda-build does not include them in the output package.
-PIP_NO_INDEX=False $PYTHON -m pip install pymanopt
+PIP_NO_INDEX=False $PYTHON -m pip install "pymanopt<=0.2.5"
 
 # NOTE: This is the recommended way to install packages
 $PYTHON setup.py install --single-version-externally-managed --record=record.txt
diff --git a/.conda/meta.yaml b/.conda/meta.yaml
@@ -1,5 +1,5 @@
 {% set conda_package_nonexistent = (
-    "pymanopt",
+    "pymanopt<=0.2.5",
 ) %}
 {% set data = load_setup_py_data() %}
 
@@ -34,19 +34,24 @@ requirements:
 
   host:
     - python
-    - setuptools
     - pip
     - mpich
-    - openmp
-    {% for req in data.get('setup_requires', [])
-        if req not in conda_package_nonexistent -%}
-    - {{req}}
-    {% endfor %}
+    - llvm-openmp
+    - setuptools>=42 
+    - wheel
+    - pybind11>=2.9.0
+    - scipy!=1.0.0
+    - cython
+    - numpy<=1.23.1
+    - setuptools_scm
 
   run:
     - python
+    - numpy<=1.23.1
     - mpich
-    - openmp
+    - llvm-openmp
+    - tensorflow
+    - tensorflow-probability
     {% for req in data.get('install_requires', [])
         if req not in conda_package_nonexistent -%}
     - {{req}}
@@ -55,7 +60,6 @@ requirements:
 test:
   commands:
     - find $BRAINIAK_HOME/tests | grep pycache | xargs rm -rf
-    - pip install tensorflow tensorflow-probability
     - mpiexec -n 2 pytest $BRAINIAK_HOME
 
       # Known issue: https://github.com/travis-ci/travis-ci/issues/4704#issuecomment-348435959
@@ -64,4 +68,5 @@ test:
     - conda inspect objects -p $PREFIX brainiak  # [osx]
   requires:
     - pytest
+    - testbook
     - numdifftools
diff --git a/.github/workflows/della_notebooks.yml b/.github/workflows/della_notebooks.yml
@@ -0,0 +1,20 @@
+# Workflow file for test runs on Princeton's della cluster. This only runs on self
+# hosted runners. The pr-check script detects that it is running on della and 
+# runs tests of notebooks that require larger resources than GitHub Actions hosted
+# runners provide. It would probably be better to move a lot of that logic to this
+# file in the future.
+
+on: [pull_request]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  notebook_tests:
+    runs-on: self-hosted
+    steps:
+      - uses: actions/checkout@v2
+      - run: |
+          chmod a+x pr-check.sh
+          ./pr-check.sh
diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml
@@ -1,4 +1,9 @@
-on: [pull_request]
+on: [pull_request, push]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
 jobs:
   pypi:
     env:
@@ -7,7 +12,7 @@ jobs:
     strategy:
       matrix:
         os: [ubuntu-20.04, macos-latest]
-        python-version: [3.8]
+        python-version: ["3.8", "3.9", "3.10"]
     steps:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v2
@@ -21,11 +26,12 @@ jobs:
           ./pr-check.sh
       - if: ${{ contains(matrix.os, 'macos') }}
         run: |
-          export CC=$(brew --prefix)/opt/llvm/bin/clang
-          export CXX=$(brew --prefix)/opt/llvm/bin/clang++
-          export LDFLAGS="-L/usr/local/opt/llvm/lib
-          -Wl,-rpath,/usr/local/opt/llvm/lib $LDFLAGS"
-          export CPPFLAGS="-I/usr/local/opt/llvm/include $CPPFLAGS"
+          export CLANG_PREFIX=$(brew --prefix llvm@15)
+          export CC=$CLANG_PREFIX/bin/clang
+          export CXX=$CLANG_PREFIX/bin/clang++
+          export LDFLAGS="-L$CLANG_PREFIX/lib
+          -Wl,-rpath,$CLANG_PREFIX/lib $LDFLAGS -L/usr/local/opt/libomp/lib"
+          export CPPFLAGS="-I$CLANG_PREFIX/include $CPPFLAGS -I/usr/local/opt/libomp/include"
           ./pr-check.sh
       - uses: codecov/codecov-action@v1
   conda:
@@ -36,10 +42,12 @@ jobs:
         python-version: [3.8]
     steps:
       - uses: actions/checkout@v2
-      - uses: conda-incubator/setup-miniconda@v2
+      - uses: mamba-org/setup-micromamba@v1
         with:
-          channels: conda-forge
-          python-version: ${{ matrix.python-version }}
+          environment-name: test-env
+          create-args: >-
+            python=${{ matrix.python-version }}
+            boa
       - run: |
           export CONDA_HOME=$CONDA
           conda install conda-build

diff --git a/.gitignore b/.gitignore
@@ -38,3 +38,9 @@ __MACOSX
 brainiak/fcma/cython_blas.c
 brainiak/eventseg/_utils.c
 examples/fcma/face_scene/
+
+
+docs/examples/fmrisim/Corr_MVPA_archive.tar.gz
+docs/examples/iem/RademakerEtAl2019_WM_S05_avgTime.npz
+docs/examples/isc/brainiak-aperture-isc-data.tgz
+docs/examples/srm/brainiak-aperture-srm-data.tgz
diff --git a/brainiak/factoranalysis/htfa.py b/brainiak/factoranalysis/htfa.py
@@ -498,7 +498,7 @@ def _init_prior_posterior(self, rank, R, n_local_subj):
 
         if rank == 0:
             idx = np.random.choice(n_local_subj, 1)
-            self.global_prior_, self.global_centers_cov,\
+            self.global_prior_, self.global_centers_cov, \
                 self.global_widths_var = self.get_template(R[idx[0]])
             self.global_centers_cov_scaled =\
                 self.global_centers_cov / float(self.n_subj)
@@ -724,7 +724,7 @@ def _fit_htfa(self, data, R):
         m = 0
         outer_converged = np.array([0])
         while m < self.max_global_iter and not outer_converged[0]:
-            if(self.verbose):
+            if (self.verbose):
                 logger.info("HTFA global iter %d " % (m))
             # root broadcast first 4 fields of global_prior to all nodes
             self.comm.Bcast(self.global_prior_, root=0)

diff --git a/brainiak/fcma/preprocessing.py b/brainiak/fcma/preprocessing.py
@@ -370,7 +370,7 @@ def prepare_searchlight_mvpa_data(images, conditions, data_type=np.float32,
 
     logger.info('start to apply masks and separate epochs')
     for sid, f in enumerate(images):
-        data = f.get_data().astype(data_type)
+        data = f.get_fdata().astype(data_type)
         [d1, d2, d3, d4] = data.shape
         if random == RandomType.REPRODUCIBLE:
             data = data.reshape((d1 * d2 * d3, d4))

diff --git a/brainiak/funcalign/fastsrm.py b/brainiak/funcalign/fastsrm.py
@@ -238,7 +238,7 @@ def _check_imgs_array(imgs):
     for i in range(n_subjects):
         for j in range(n_sessions):
             if not (isinstance(imgs[i, j], str) or isinstance(
-                    imgs[i, j], np.str_) or isinstance(imgs[i, j], np.str)):
+                    imgs[i, j], np.str_) or isinstance(imgs[i, j], str)):
                 raise ValueError("imgs[%i, %i] is stored using "
                                  "type %s which is not a str" %
                                  (i, j, type(imgs[i, j])))
@@ -337,7 +337,7 @@ def check_atlas(atlas, n_components=None):
         return None
 
     if not (isinstance(atlas, np.ndarray) or isinstance(atlas, str)
-            or isinstance(atlas, np.str_) or isinstance(atlas, np.str)):
+            or isinstance(atlas, np.str_) or isinstance(atlas, str)):
         raise ValueError("Atlas is stored using "
                          "type %s which is neither np.ndarray or str" %
                          type(atlas))

diff --git a/brainiak/funcalign/rsrm.py b/brainiak/funcalign/rsrm.py
@@ -191,8 +191,8 @@ def transform(self, X):
         return r, s
 
     def _transform_new_data(self, X, subject):
-        """Transform new data for a subjects by projecting to the shared subspace and
-        computing the individual information.
+        """Transform new data for a subjects by projecting to the shared
+        subspace and computing the individual information.
 
         Parameters
         ----------

diff --git a/brainiak/funcalign/srm.py b/brainiak/funcalign/srm.py
@@ -51,7 +51,8 @@
 
 
 def _init_w_transforms(data, features, random_states, comm=MPI.COMM_SELF):
-    """Initialize the mappings (Wi) for the SRM with random orthogonal matrices.
+    """Initialize the mappings (Wi) for the SRM with random orthogonal
+    matrices.
 
     Parameters
     ----------
@@ -240,8 +241,8 @@ def fit(self, X, y=None):
                     "Not all ranks have same number of subjects")
 
         # Collect size information
-        shape0 = np.zeros((number_subjects,), dtype=np.int)
-        shape1 = np.zeros((number_subjects,), dtype=np.int)
+        shape0 = np.zeros((number_subjects,), dtype=int)
+        shape1 = np.zeros((number_subjects,), dtype=int)
 
         for subject in range(number_subjects):
             if X[subject] is not None:
@@ -480,7 +481,8 @@ def save(self, file):
         )
 
     def _srm(self, data):
-        """Expectation-Maximization algorithm for fitting the probabilistic SRM.
+        """Expectation-Maximization algorithm for fitting the probabilistic
+        SRM.
 
         Parameters
         ----------
@@ -854,7 +856,8 @@ def transform_subject(self, X):
         return w
 
     def _srm(self, data):
-        """Expectation-Maximization algorithm for fitting the probabilistic SRM.
+        """Expectation-Maximization algorithm for fitting the probabilistic
+        SRM.
 
         Parameters
         ----------

diff --git a/brainiak/funcalign/sssrm.py b/brainiak/funcalign/sssrm.py
@@ -32,6 +32,16 @@
 from sklearn.utils import assert_all_finite
 from sklearn.utils.validation import NotFittedError
 from sklearn.utils.multiclass import unique_labels
+
+# Workaround for Theano for numpy after 1.20.3, see:
+# https://github.com/numpy/numpy/issues/21079
+try:
+    import numpy.distutils
+    blas_info = np.__config__.blas_ilp64_opt_info  # type: ignore
+    numpy.distutils.__config__.blas_opt_info = blas_info  # type: ignore
+except Exception:
+    pass
+
 import theano
 import theano.tensor as T
 import theano.compile.sharedvalue as S

diff --git a/brainiak/image.py b/brainiak/image.py
@@ -24,7 +24,7 @@
 
 import itertools
 
-from typing import Iterable, Sequence, Type, TypeVar
+from typing import Optional, Iterable, Sequence, Type, TypeVar
 
 import numpy as np
 
@@ -104,8 +104,11 @@ def extract_labels(self) -> np.ndarray:
         return condition_idxs[unique_epoch_idxs]
 
 
-def mask_image(image: SpatialImage, mask: np.ndarray, data_type: type = None
-               ) -> np.ndarray:
+def mask_image(
+        image: SpatialImage,
+        mask: np.ndarray,
+        data_type: Optional[type] = None
+        ) -> np.ndarray:
     """Mask image after optionally casting its type.
 
     Parameters
@@ -127,19 +130,21 @@ def mask_image(image: SpatialImage, mask: np.ndarray, data_type: type = None
     ValueError
         Image data and masks have different shapes.
     """
-    image_data = image.get_data()
+    image_data = image.get_fdata()
     if image_data.shape[:3] != mask.shape:
         raise ValueError("Image data and mask have different shapes.")
     if data_type is not None:
-        cast_data = image_data.astype(data_type)
+        cast_data: np.ndarray = image_data.astype(data_type)
     else:
         cast_data = image_data
     return cast_data[mask]
 
 
-def multimask_images(images: Iterable[SpatialImage],
-                     masks: Sequence[np.ndarray], image_type: type = None
-                     ) -> Iterable[Sequence[np.ndarray]]:
+def multimask_images(
+        images: Iterable[SpatialImage],
+        masks: Sequence[np.ndarray],
+        image_type: Optional[type] = None
+        ) -> Iterable[Sequence[np.ndarray]]:
     """Mask images with multiple masks.
 
     Parameters
@@ -161,7 +166,7 @@ def multimask_images(images: Iterable[SpatialImage],
 
 
 def mask_images(images: Iterable[SpatialImage], mask: np.ndarray,
-                image_type: type = None) -> Iterable[np.ndarray]:
+                image_type: Optional[type] = None) -> Iterable[np.ndarray]:
     """Mask images.
 
     Parameters
@@ -178,5 +183,5 @@ def mask_images(images: Iterable[SpatialImage], mask: np.ndarray,
     np.ndarray
         Masked image.
     """
-    for images in multimask_images(images, (mask,), image_type):
-        yield images[0]
+    for imgs in multimask_images(images, (mask,), image_type):  # type: ignore
+        yield imgs[0]  # type: ignore
diff --git a/brainiak/io.py b/brainiak/io.py
@@ -22,7 +22,7 @@
 ]
 
 from pathlib import Path
-from typing import Callable, Iterable, List, Union
+from typing import Optional, Callable, Iterable, List, Union
 
 import logging
 import nibabel as nib
@@ -69,7 +69,7 @@ def load_images_from_dir(in_dir: Union[str, Path], suffix: str = "nii.gz",
         logger.debug(
             'Starting to read file %s', f
         )
-        yield nib.load(str(f))
+        yield nib.load(str(f))  # type: ignore
 
 
 def load_images(image_paths: Iterable[Union[str, Path]]
@@ -100,11 +100,12 @@ def load_images(image_paths: Iterable[Union[str, Path]]
         logger.debug(
             'Starting to read file %s', string_path
         )
-        yield nib.load(string_path)
+        yield nib.load(string_path)  # type: ignore
 
 
 def load_boolean_mask(path: Union[str, Path],
-                      predicate: Callable[[np.ndarray], np.ndarray] = None
+                      predicate: Optional[
+                          Callable[[np.ndarray], np.ndarray]] = None
                       ) -> np.ndarray:
     """Load boolean nibabel.SpatialImage mask.
 
@@ -123,11 +124,11 @@ def load_boolean_mask(path: Union[str, Path],
     """
     if not isinstance(path, str):
         path = str(path)
-    data = nib.load(path).get_data()
+    data = nib.load(path).get_fdata()  # type: ignore
     if predicate is not None:
         mask = predicate(data)
     else:
-        mask = data.astype(np.bool)
+        mask = data.astype(bool)
     return mask
 
 

diff --git a/brainiak/reprsimil/brsa.py b/brainiak/reprsimil/brsa.py
@@ -4080,13 +4080,13 @@ def _check_scan_onsets_GBRSA(self, scan_onsets, X):
         return scan_onsets
 
     def _bin_exp(self, n_bin, scale=1.0):
-        """ Calculate the bin locations to approximate exponential distribution.
-            It breaks the cumulative probability of exponential distribution
-            into n_bin equal bins, each covering 1 / n_bin probability. Then it
-            calculates the center of mass in each bins and returns the
-            centers of mass. So, it approximates the exponential distribution
-            with n_bin of Delta function weighted by 1 / n_bin, at the
-            locations of these centers of mass.
+        """ Calculate the bin locations to approximate exponential
+            distribution. It breaks the cumulative probability of
+            exponential distribution into n_bin equal bins, each covering
+            1 / n_bin probability. Then it calculates the center of mass in
+            each bins and returns the centers of mass. So, it approximates the
+            exponential distribution with n_bin of Delta function weighted by
+            1 / n_bin, at the locations of these centers of mass.
         Parameters:
         -----------
         n_bin: int