junifer 0.0.8.dev257__tar.gz → 0.0.8.dev270__tar.gz

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/Dockerfile-ci

@@ -115,11 +115,11 @@ RUN set -eux; \
     echo "${DATA_VERSION}" > "$HOME/junifer-data-version"
 
 # --- nilearn datasets used by junifer.testing (no downloads at test time) ---
-ARG NILEARN_VERSION=0.10.4
+ARG NILEARN_VERSION=0.14.1
 RUN <<'EOF'
 set -eux
 python3 -m venv /tmp/nilearn-venv
-/tmp/nilearn-venv/bin/pip install --no-cache-dir -q "nilearn==${NILEARN_VERSION}" "numpy<2.4"
+/tmp/nilearn-venv/bin/pip install --no-cache-dir -q "nilearn==${NILEARN_VERSION}"
 /tmp/nilearn-venv/bin/python - <<'PY'
 from nilearn import datasets
 datasets.fetch_oasis_vbm(n_subjects=10)
@@ -131,4 +131,65 @@ rm -rf /tmp/nilearn-venv
 du -sh "$HOME/nilearn_data"
 EOF
 
+# --- Files derived by junifer.testing and templateflow templates ---
+# Generated with junifer itself (from the build context), so they match what
+# the tests would otherwise create in every job: shortened BOLD files next to
+# the nilearn datasets, the UCLA CNP VBM maps (OpenNeuro ds000030) and the
+# templates used by the tests.
+COPY --chown=docker:docker pyproject.toml setup.py README.md /tmp/junifer-src/
+COPY --chown=docker:docker junifer /tmp/junifer-src/junifer
+RUN <<'EOF'
+set -eux
+python3 -m venv /tmp/junifer-venv
+# Editable: without .git, setuptools-scm cannot list the subpackages for a
+# regular install
+SETUPTOOLS_SCM_PRETEND_VERSION=0.0.0 \
+    /tmp/junifer-venv/bin/pip install --no-cache-dir -q -e /tmp/junifer-src
+/tmp/junifer-venv/bin/python - <<'PY'
+from templateflow import api as tflow
+
+from junifer.testing.datagrabbers import (
+    ADHDTestingDataGrabber,
+    PartlyCloudyTestingDataGrabber,
+    UCLACNPVBMTestingDataGrabber,
+)
+
+# Shortened BOLD (and confounds) used by the tests
+with ADHDTestingDataGrabber() as dg:
+    dg["sub-01"]
+with PartlyCloudyTestingDataGrabber(n_timepoints=50) as dg:
+    dg["sub-01"]
+
+# VBM maps (downloaded from OpenNeuro) resampled to the 2mm grid
+with UCLACNPVBMTestingDataGrabber() as dg:
+    for element in dg.get_elements():
+        dg[element]
+
+# Templates used by the tests
+for space, res in [
+    ("MNI152Lin", 1),
+    ("MNI152Lin", 2),
+    ("MNI152NLin2009aAsym", 1),
+    ("MNI152NLin2009cAsym", 1),
+    ("MNI152NLin2009cAsym", 2),
+    ("MNI152NLin6Asym", 1),
+    ("MNI152NLin6Asym", 2),
+]:
+    tflow.get(space, resolution=res, suffix="T1w", desc=None, extension="nii.gz")
+for space in ("MNI152Lin", "MNI152NLin2009cAsym", "MNI152NLin6Asym"):
+    tflow.get(space, resolution=2, desc="brain", suffix="mask", extension="nii.gz")
+for label in ("CSF", "GM", "WM"):
+    tflow.get(
+        "MNI152NLin2009cAsym",
+        resolution=2,
+        label=label,
+        suffix="probseg",
+        extension="nii.gz",
+    )
+PY
+rm -rf /tmp/junifer-venv /tmp/junifer-src
+find "$HOME/nilearn_data" -name "*_tp-*" -o -name "*_res-*"
+du -sh "$HOME/.cache/templateflow"
+EOF
+
 ENTRYPOINT ["/start.sh"]

{junifer-0.0.8.dev257/junifer.egg-info → junifer-0.0.8.dev270}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: junifer
-Version: 0.0.8.dev257
+Version: 0.0.8.dev270
 Summary: JUelich NeuroImaging FEature extractoR
 Author-email: Fede Raimondo <f.raimondo@fz-juelich.de>, Synchon Mandal <s.mandal@fz-juelich.de>
 Maintainer-email: Fede Raimondo <f.raimondo@fz-juelich.de>, Synchon Mandal <s.mandal@fz-juelich.de>
@@ -34,7 +34,7 @@ Requires-Dist: scipy<1.17.0,>=1.10.0
 Requires-Dist: datalad<1.3.0,>=1.0.0
 Requires-Dist: pandas<2.4.0,>=2.0.0
 Requires-Dist: nibabel<5.4.0,>=5.2.0
-Requires-Dist: nilearn<=0.10.4,>=0.10.3
+Requires-Dist: nilearn<0.15,>=0.14.0
 Requires-Dist: sqlalchemy<=2.1.0,>=2.0.25
 Requires-Dist: ruamel.yaml<0.19,>=0.17
 Requires-Dist: h5py>=3.10
@@ -44,7 +44,7 @@ Requires-Dist: lapy<2.0.0,>=1.0.0
 Requires-Dist: lazy_loader==0.4
 Requires-Dist: importlib_metadata; python_version < "3.11"
 Requires-Dist: looseversion==1.3.0; python_version >= "3.12"
-Requires-Dist: junifer-data>=1.4.0
+Requires-Dist: junifer-data>=1.5.0
 Requires-Dist: structlog<26.0.0,>=25.0.0
 Requires-Dist: pydantic>=2.11.4
 Requires-Dist: aenum<3.2.0,>=3.1.0

junifer-0.0.8.dev270/docs/changes/newsfragments/515.doc

@@ -0,0 +1 @@
+Use the UCLA CNP VBM testing data and the Tian parcellations in the examples and the codeless documentation, so no example needs warping by `Fede Raimondo`_

junifer-0.0.8.dev270/docs/changes/newsfragments/515.enh

@@ -0,0 +1 @@
+Add :class:`.UCLACNPVBMTestingDataGrabber` with grey matter probability maps from OpenNeuro ds000030 in MNI152NLin2009cAsym, add ``n_subjects`` to :class:`.PartlyCloudyTestingDataGrabber` and use the native 3mm resolution by default in :class:`.ADHDTestingDataGrabber` by `Fede Raimondo`_

junifer-0.0.8.dev270/docs/changes/newsfragments/515.enh.1

@@ -0,0 +1 @@
+Speed up :class:`.ParcelAggregation` by not passing a target affine to the masker, which made ``nilearn`` crop and copy the input image even when the parcellation was already on its grid by `Fede Raimondo`_

junifer-0.0.8.dev270/docs/changes/newsfragments/515.misc

@@ -0,0 +1 @@
+Speed up the external tests by using fewer subjects, the UCLA CNP VBM testing data and ``junifer-data`` 1.5.0, and bake the derived testing data and templateflow templates into the CI image by `Fede Raimondo`_

junifer-0.0.8.dev270/docs/changes/newsfragments/515.misc.1

@@ -0,0 +1 @@
+Speed up the test suite by freezing the objects alive after collection, so the full garbage collections forced by ``nilearn`` are much faster by `Fede Raimondo`_

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/docs/extending/marker.rst

@@ -142,7 +142,7 @@ and the values would be a dictionary of storage type specific key-value pairs.
         # Create a masker
         masker = NiftiLabelsMasker(
             labels_img=t_parcellation,
-            standardize=True,
+            standardize="zscore_sample",
             memory="nilearn_cache",
             verbose=5,
         )
@@ -232,7 +232,7 @@ Finally, we need to register the Marker using the ``@register_marker`` decorator
             # Create a masker
             masker = NiftiLabelsMasker(
                 labels_img=t_parcellation,
-                standardize=True,
+                standardize="zscore_sample",
                 memory="nilearn_cache",
                 verbose=5,
             )

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/docs/understanding/preprocess.rst

@@ -108,8 +108,10 @@ parameters:
      - Apply detrending on timeseries, before confound removal.
      - activated
    * - ``standardize``
-     - Scale signals to unit variance.
-     - activated
+     - | Strategy to standardize the signals: ``zscore_sample`` (zero mean
+       | and unit variance), ``psc`` (percent signal change) or ``null``
+       | (no standardization).
+     - ``zscore_sample``
    * - ``low_pass``
      - Low cutoff frequencies, in Hertz.
      - deactivated

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/docs/using/codeless.rst

@@ -31,7 +31,7 @@ as well as some general parameters.
 
 As an example, we will generate the configuration file for a pipeline that will
 extract the mean ``VBM_GM`` values using two different parcellations and one set
-of coordinates, from the ``Oasis VBM Testing dataset`` included in ``junifer``.
+of coordinates, from the ``UCLA CNP VBM Testing dataset`` included in ``junifer``.
 
 
 General Parameters
@@ -96,12 +96,12 @@ task.
     kind: DataladAOMICPIOP1
     tasks: restingstate
 
-In the ``Oasis VBM Testing dataset`` example, the section will look like this:
+In the ``UCLA CNP VBM Testing dataset`` example, the section will look like this:
 
 .. code-block:: yaml
 
   datagrabber:
-    kind: OasisVBMTestingDataGrabber
+    kind: UCLACNPVBMTestingDataGrabber
 
 
 Data Reader
@@ -121,7 +121,7 @@ keys to pass parameters to the DataReader constructor:
       kind: DefaultDataReader
 
 
-For the ``Oasis VBM Testing dataset`` example, we will not specify a
+For the ``UCLA CNP VBM Testing dataset`` example, we will not specify a
 ``datareader`` step.
 
 Preprocess
@@ -146,10 +146,10 @@ need to specify its name as the ``kind`` key, as well as its parameters.
         global_signal: basic
       spike: 0.2
       detrend: false
-      standardize: true
+      standardize: zscore_sample
 
 
-For the ``Oasis VBM Testing dataset`` example, we will not specify a
+For the ``UCLA CNP VBM Testing dataset`` example, we will not specify a
 preprocessing step.
 
 
@@ -160,22 +160,22 @@ The ``markers`` section like the ``preprocess`` section expects a list of
 markers. Each Marker has a name that we can use to refer to it later,
 and a set of parameters that will be passed to the Marker.
 
-For the ``Oasis VBM Testing dataset`` example, we want to compute the mean
-``VBM_GM`` value for each parcel using the ``Schaefer parcellation (100 parcels,
-7 networks)``, ``Schaefer parcellation (200 parcels, 7 networks)``, and the
-``DMNBuckner`` network, using ``5mm`` spheres. Thus, we will configure the
-``markers`` section as follows:
+For the ``UCLA CNP VBM Testing dataset`` example, we want to compute the mean
+``VBM_GM`` value for each parcel using the ``Tian subcortical parcellation
+(scale I, 16 parcels)``, ``Tian subcortical parcellation (scale II, 32
+parcels)``, and the ``DMNBuckner`` network, using ``5mm`` spheres. Thus, we will
+configure the ``markers`` section as follows:
 
 .. code-block:: yaml
 
   markers:
-    - name: Schaefer100x7_mean
+    - name: TianxS1_mean
       kind: ParcelAggregation
-      parcellation: Schaefer100x7
+      parcellation: TianxS1x3TxMNInonlinear2009cAsym
       method: mean
-    - name: Schaefer200x7_mean
+    - name: TianxS2_mean
       kind: ParcelAggregation
-      parcellation: Schaefer200x7
+      parcellation: TianxS2x3TxMNInonlinear2009cAsym
       method: mean
     - name: DMNBuckner_5mm_mean
       kind: SphereAggregation
@@ -198,13 +198,13 @@ specify where we want to store the results:
 
     storage:
       kind: HDF5FeatureStorage
-      uri: /data/junifer/example/oasis_vbm_testing.hdf5
+      uri: /data/junifer/example/uclacnp_vbm_testing.hdf5
 
 
 Complete Example
 ----------------
 
-This is how the full ``Oasis VBM Testing dataset`` example configuration file
+This is how the full ``UCLA CNP VBM Testing dataset`` example configuration file
 looks like:
 
 .. code-block:: yaml
@@ -214,16 +214,16 @@ looks like:
   workdir: /tmp
 
   datagrabber:
-    kind: OasisVBMTestingDataGrabber
+    kind: UCLACNPVBMTestingDataGrabber
 
   markers:
-    - name: Schaefer100x7_mean
+    - name: TianxS1_mean
       kind: ParcelAggregation
-      parcellation: Schaefer100x7
+      parcellation: TianxS1x3TxMNInonlinear2009cAsym
       method: mean
-    - name: Schaefer200x7_mean
+    - name: TianxS2_mean
       kind: ParcelAggregation
-      parcellation: Schaefer200x7
+      parcellation: TianxS2x3TxMNInonlinear2009cAsym
       method: mean
     - name: DMNBuckner_5mm_mean
       kind: SphereAggregation
@@ -233,4 +233,4 @@ looks like:
 
   storage:
     kind: HDF5FeatureStorage
-    uri: /data/junifer/example/oasis_vbm_testing.hdf5
+    uri: /data/junifer/example/uclacnp_vbm_testing.hdf5

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/examples/run_compute_parcel_mean.py

@@ -1,12 +1,10 @@
 """
-Compute Sphere and Parcel Aggregation.
-======================================
+Compute Parcel Aggregation.
+===========================
 
-This example uses the ``SphereAggregation`` marker to compute the mean of
-spheres around the extended default mode network coordinates (16 ROIs) for a
-3D NIfTI and the ``ParcelAggregation`` marker to compute the mean of each
-parcel using the Schaefer parcellation (100 rois, 7 Yeo networks) for a 4D
-NIfTI.
+This example uses the ``ParcelAggregation`` marker to compute the mean of each
+parcel using the Tian subcortical parcellation (scale I, 16 ROIs) for a 3D
+NIfTI and the Schaefer parcellation (100 rois, 7 Yeo networks) for a 4D NIfTI.
 
 Authors: Federico Raimondo, Synchon Mandal
 
@@ -15,11 +13,11 @@ License: BSD 3 clause
 
 from junifer.testing.datagrabbers import (
     ADHDTestingDataGrabber,
-    OasisVBMTestingDataGrabber,
+    UCLACNPVBMTestingDataGrabber,
 )
 from junifer.datagrabber import DataType
 from junifer.datareader import DefaultDataReader
-from junifer.markers import ParcelAggregation, SphereAggregation
+from junifer.markers import ParcelAggregation
 from junifer.stats import AggFunc
 from junifer.utils import configure_logging
 
@@ -29,24 +27,22 @@ from junifer.utils import configure_logging
 configure_logging(level="INFO")
 
 ###############################################################################
-# Perform sphere aggregation on VBM GM data (3D) from OASIS dataset
-with OasisVBMTestingDataGrabber() as dg:
+# Perform parcel aggregation on VBM GM data (3D) from the UCLA CNP dataset
+with UCLACNPVBMTestingDataGrabber() as dg:
     # Get the first element
     element = dg.get_elements()[0]
     # Read the element
     element_data = DefaultDataReader().fit_transform(dg[element])
     # Initialize marker
-    marker = SphereAggregation(
-        coords="extDMN",
-        radius=5.0,
+    marker = ParcelAggregation(
+        parcellation="TianxS1x3TxMNInonlinear2009cAsym",
         method=AggFunc.Mean,
-        masks="compute_brain_mask",
     )
     # Compute feature
     feature = marker.fit_transform(element_data)
     # Print the output
     print(feature.keys())
-    print(feature["VBM_GM"]["aggregation"]["data"].shape)  # Shape is (1 x spheres)
+    print(feature["VBM_GM"]["aggregation"]["data"].shape)  # Shape is (1 x parcels)
 
 ###############################################################################
 # Perform parcel aggregation on BOLD data (4D) from ADHD dataset

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/examples/run_junifer_julearn.py

@@ -2,10 +2,10 @@
 Run junifer and julearn.
 ========================
 
-This example uses a SphereAggregation marker to compute the mean of spheres
-around the extended default mode network coordinates (16 ROIs) for a 3D nifti
-to extract some features for machine learning using julearn to predict some
-other data.
+This example uses a ParcelAggregation marker to compute the mean of each parcel
+of the Tian subcortical parcellation (scale II, 32 ROIs) for a 3D nifti to
+extract some features for machine learning using julearn to predict some other
+data.
 
 Authors: Leonard Sasse, Sami Hamdan, Nicolas Nieto, Synchon Mandal
 
@@ -14,13 +14,12 @@ License: BSD 3 clause
 
 import tempfile
 
-import nilearn
-import pandas as pd
 from julearn import run_cross_validation, PipelineCreator
 
 import junifer.testing.registry  # noqa: F401
 from junifer.api import collect, run
 from junifer.storage import HDF5FeatureStorage
+from junifer.testing.datagrabbers import UCLACNPVBMTestingDataGrabber
 from junifer.utils import configure_logging
 
 
@@ -34,20 +33,16 @@ configure_logging(level="INFO")
 
 marker_dicts = [
     {
-        "name": "extDMN_TrimMean80",
-        "kind": "SphereAggregation",
-        "coords": "extDMN",
-        "radius": 5.0,
-        "masks": "compute_brain_mask",
+        "name": "TianxS2_TrimMean80",
+        "kind": "ParcelAggregation",
+        "parcellation": "TianxS2x3TxMNInonlinear2009cAsym",
         "method": "trim_mean",
         "method_params": {"proportiontocut": 0.2},
     },
     {
-        "name": "extDMN_Mean",
-        "kind": "SphereAggregation",
-        "coords": "extDMN",
-        "radius": 5.0,
-        "masks": "compute_brain_mask",
+        "name": "TianxS2_Mean",
+        "kind": "ParcelAggregation",
+        "parcellation": "TianxS2x3TxMNInonlinear2009cAsym",
         "method": "mean",
     },
 ]
@@ -60,15 +55,10 @@ confound = "sex"
 
 
 ###############################################################################
-# Load the VBM phenotype data for machine learning data:
-# - Fetch the Oasis dataset
-oasis_dataset = nilearn.datasets.fetch_oasis_vbm()
-age = oasis_dataset.ext_vars[y][:10]
-sex = (
-    pd.Series(oasis_dataset.ext_vars["mf"][:10])
-    .map(lambda x: 1 if x == "F" else 0)
-    .values
-)
+# Load the phenotype data of the UCLA CNP subjects (OpenNeuro ds000030) for
+# machine learning:
+with UCLACNPVBMTestingDataGrabber() as dg:
+    participants = dg.get_participants()
 
 
 ###############################################################################
@@ -78,7 +68,7 @@ with tempfile.TemporaryDirectory() as tmpdir:
     # run the defined junifer feature extraction pipeline
     run(
         workdir="/tmp",
-        datagrabber={"kind": "OasisVBMTestingDataGrabber"},
+        datagrabber={"kind": "UCLACNPVBMTestingDataGrabber"},
         markers=marker_dicts,
         storage=storage,
     )
@@ -88,17 +78,19 @@ with tempfile.TemporaryDirectory() as tmpdir:
     collect(storage)
     db = HDF5FeatureStorage(uri=storage["uri"])
 
-    df_vbm = db.read_df(feature_name="VBM_GM_extDMN_Mean_aggregation")
-    oasis_subjects = [x[0] for x in df_vbm.index]
-    df_vbm.index = oasis_subjects
+    df_vbm = db.read_df(feature_name="VBM_GM_TianxS2_Mean_aggregation")
+    df_vbm.index = [x[0] for x in df_vbm.index]
 
 
 ###############################################################################
 # Using julearn for machine learning:
 # We predict the age given our vbm features and sex as a confound.
 X = list(df_vbm.columns)
-df_vbm[y] = age
-df_vbm[confound] = sex
+# Match the phenotypes to the subjects by their identifier
+df_vbm[y] = participants.loc[df_vbm.index, "age"].to_numpy()
+df_vbm[confound] = (
+    participants.loc[df_vbm.index, "gender"] == "F"
+).to_numpy(dtype=int)
 
 X_types = {
     "features": X,

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/examples/run_run_gmd_mean.py

@@ -1,6 +1,6 @@
 """
-UKB VBM GMD Extraction
-======================
+UCLA CNP VBM GMD Extraction
+===========================
 
 Authors: Federico Raimondo
 
@@ -13,33 +13,27 @@ from junifer.api import run
 
 
 datagrabber = {
-    "kind": "OasisVBMTestingDataGrabber",
+    "kind": "UCLACNPVBMTestingDataGrabber",
 }
 
 markers = [
     {
-        "name": "extDMN_TrimMean80",
-        "kind": "SphereAggregation",
-        "coords": "extDMN",
-        "radius": 5.0,
-        "masks": "compute_brain_mask",
+        "name": "TianxS2_TrimMean80",
+        "kind": "ParcelAggregation",
+        "parcellation": "TianxS2x3TxMNInonlinear2009cAsym",
         "method": "trim_mean",
         "method_params": {"proportiontocut": 0.2},
     },
     {
-        "name": "extDMN_Mean",
-        "kind": "SphereAggregation",
-        "coords": "extDMN",
-        "radius": 5.0,
-        "masks": "compute_brain_mask",
+        "name": "TianxS2_Mean",
+        "kind": "ParcelAggregation",
+        "parcellation": "TianxS2x3TxMNInonlinear2009cAsym",
         "method": "mean",
     },
     {
-        "name": "extDMN_Std",
-        "kind": "SphereAggregation",
-        "coords": "extDMN",
-        "radius": 5.0,
-        "masks": "compute_brain_mask",
+        "name": "TianxS2_Std",
+        "kind": "ParcelAggregation",
+        "parcellation": "TianxS2x3TxMNInonlinear2009cAsym",
         "method": "std",
     },
 ]

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/examples/yamls/gmd_mean.yaml

@@ -2,22 +2,22 @@ with: junifer.testing.registry
 workdir: /tmp
 
 datagrabber:
-  kind: OasisVBMTestingDataGrabber
+  kind: UCLACNPVBMTestingDataGrabber
 elements:
 markers:
-  - name: Schaefer1000x7_TrimMean80
+  - name: TianxS2_TrimMean80
     kind: ParcelAggregation
-    parcellation: Schaefer1000x7
+    parcellation: TianxS2x3TxMNInonlinear2009cAsym
     method: trim_mean
     method_params:
       proportiontocut: 0.2
-  - name: Schaefer1000x7_Mean
+  - name: TianxS2_Mean
     kind: ParcelAggregation
-    parcellation: Schaefer1000x7
+    parcellation: TianxS2x3TxMNInonlinear2009cAsym
     method: mean
-  - name: Schaefer1000x7_Std
+  - name: TianxS2_Std
     kind: ParcelAggregation
-    parcellation: Schaefer1000x7
+    parcellation: TianxS2x3TxMNInonlinear2009cAsym
     method: std
 storage:
   kind: HDF5FeatureStorage

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/examples/yamls/gmd_mean_htcondor.yaml

@@ -2,11 +2,11 @@ with: junifer.testing.registry
 workdir: /tmp
 
 datagrabber:
-  kind: OasisVBMTestingDataGrabber
+  kind: UCLACNPVBMTestingDataGrabber
 markers:
-  - name: Schaefer1000x7_Mean
+  - name: TianxS2_Mean
     kind: ParcelAggregation
-    parcellation: Schaefer1000x7
+    parcellation: TianxS2x3TxMNInonlinear2009cAsym
     method: mean
 storage:
   kind: HDF5FeatureStorage

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/junifer/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.0.8.dev257'
-__version_tuple__ = version_tuple = (0, 0, 8, 'dev257')
+__version__ = version = '0.0.8.dev270'
+__version_tuple__ = version_tuple = (0, 0, 8, 'dev270')
 
-__commit_id__ = commit_id = 'g19ea0a56d'
+__commit_id__ = commit_id = 'gcf60784bf'

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/junifer/api/tests/test_functions.py

@@ -35,6 +35,8 @@ yaml.indent(mapping=2, sequence=4, offset=2)
 # Kept for parametrizing
 _datagrabber = {
     "kind": "PartlyCloudyTestingDataGrabber",
+    # Fewer subjects for faster tests
+    "n_subjects": 4,
 }
 _bids_ses_datagrabber = {
     "kind": "PatternDataladDataGrabber",
@@ -334,10 +336,13 @@ def test_run_and_collect(
     )
     # Get datagrabber
     dg = PipelineComponentRegistry().build_component_instance(
-        step="datagrabber", name=datagrabber["kind"], baseclass=BaseDataGrabber
+        step="datagrabber",
+        name=datagrabber["kind"],
+        baseclass=BaseDataGrabber,
+        init_params={k: v for k, v in datagrabber.items() if k != "kind"},
     )
     elements = dg.get_elements()  # type: ignore
-    # This should create 10 files
+    # This should create one file per element
     files = list(tmp_path.glob("*.sqlite"))
     assert len(files) == len(elements)
     # But the test.sqlite file should not exist

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/junifer/cli/tests/data/gmd_mean.yaml

@@ -2,12 +2,12 @@ with: junifer.testing.registry
 workdir: /tmp
 
 datagrabber:
-  kind: OasisVBMTestingDataGrabber
-elements: [1, 2]
+  kind: UCLACNPVBMTestingDataGrabber
+elements: [sub-10206, sub-10339]
 markers:
-  - name: Schaefer1000x7_Mean
+  - name: TianxS2_Mean
     kind: ParcelAggregation
-    parcellation: Schaefer1000x7
+    parcellation: TianxS2x3TxMNInonlinear2009cAsym
     method: mean
 storage:
   kind: SQLiteFeatureStorage

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/junifer/cli/tests/data/gmd_mean_htcondor.yaml

@@ -2,11 +2,11 @@ with: junifer.testing.registry
 workdir: /tmp
 
 datagrabber:
-  kind: OasisVBMTestingDataGrabber
+  kind: UCLACNPVBMTestingDataGrabber
 markers:
-  - name: Schaefer1000x7_Mean
+  - name: TianxS2_Mean
     kind: ParcelAggregation
-    parcellation: Schaefer1000x7
+    parcellation: TianxS2x3TxMNInonlinear2009cAsym
     method: mean
 storage:
   kind: SQLiteFeatureStorage

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/junifer/cli/tests/data/partly_cloudy_agg_mean_tian.yml

@@ -4,6 +4,7 @@ with: junifer.testing.registry
 
 datagrabber:
   kind: PartlyCloudyTestingDataGrabber
+  n_subjects: 4
 
 markers:
   - kind: ParcelAggregation

{junifer-0.0.8.dev257 → junifer-0.0.8.dev270}/junifer/conftest.py

@@ -4,6 +4,8 @@
 #          Synchon Mandal <s.mandal@fz-juelich.de>
 # License: AGPL
 
+import gc
+
 import pytest
 
 from junifer.utils.singleton import Singleton
@@ -14,6 +16,24 @@ from junifer.utils.singleton import Singleton
 collect_ignore = ["external/h5io", "external/BrainPrint"]
 
 
+def pytest_collection_finish(session: pytest.Session) -> None:
+    """Exclude the objects alive after collection from garbage collection.
+
+    nilearn (< 0.14) forces a full garbage collection every time it gets
+    the data of an image, which has to go over all the objects of the test
+    session (modules, collected tests, etc.). Freezing them makes these
+    collections much faster.
+
+    Parameters
+    ----------
+    session : pytest.Session
+        The pytest session.
+
+    """
+    gc.collect()
+    gc.freeze()
+
+
 @pytest.fixture(autouse=True)
 def reset_singletons() -> None:
     """Reset all singletons."""