Dev docs nilearn/nilearn@6275389 : [FIX] coerce loaded data from NKI …

…to float in examples (#4475)

* [full doc] print list of dependencies

* fix

* add comments

dev/.buildinfo

@@ -1,4 +1,4 @@
 # Sphinx build info version 1
 # This file hashes the configuration used when building these files. When it is not found, a full rebuild will be done.
-config: 35b812e4d8cf3c1702417d577b6af5ba
+config: b23d7c84dde12d66b77cc72aa762a240
 tags: 645f666f9bcd5a90fca523b33c5a78b7

dev/_downloads/193d69e0f7475d4607f305617999a80e/plot_surf_stat_map.py

@@ -84,6 +84,9 @@
 
 timeseries = surface.load_surf_data(nki_dataset["func_left"][0])
 
+# Coercing to float is required to avoid errors withj scipy >= 0.14.0
+timeseries = timeseries.astype(float)
+
 # Extract seed region via label
 pcc_region = b"G_cingul-Post-dorsal"
 

dev/_downloads/38fb5bdda0ea836f19944f2bf2ac3796/plot_haxby_glm_decoding.ipynb

@@ -22,7 +22,7 @@
       },
       "outputs": [],
       "source": [
-        "# By default 2nd subject will be fetched\nimport numpy as np\nimport pandas as pd\n\nfrom nilearn import datasets\n\nhaxby_dataset = datasets.fetch_haxby()\n\n# repetition has to be known\nTR = 2.5"
+        "# By default 2nd subject will be fetched\nimport numpy as np\nimport pandas as pd\n\nfrom nilearn import datasets\n\nhaxby_dataset = datasets.fetch_haxby()\n\n# repetition has to be known\nt_r = 2.5"
       ]
     },
     {
@@ -58,7 +58,7 @@
       },
       "outputs": [],
       "source": [
-        "events = {}\n# events will take  the form of a dictionary of Dataframes, one per run\nfor run in unique_runs:\n    # get the condition label per run\n    conditions_run = conditions[runs == run]\n    # get the number of scans per run, then the corresponding\n    # vector of frame times\n    n_scans = len(conditions_run)\n    frame_times = TR * np.arange(n_scans)\n    # each event last the full TR\n    duration = TR * np.ones(n_scans)\n    # Define the events object\n    events_ = pd.DataFrame(\n        {\n            \"onset\": frame_times,\n            \"trial_type\": conditions_run,\n            \"duration\": duration,\n        }\n    )\n    # remove the rest condition and insert into the dictionary\n    events[run] = events_[events_.trial_type != \"rest\"]"
+        "events = {}\n# events will take  the form of a dictionary of Dataframes, one per run\nfor run in unique_runs:\n    # get the condition label per run\n    conditions_run = conditions[runs == run]\n    # get the number of scans per run, then the corresponding\n    # vector of frame times\n    n_scans = len(conditions_run)\n    frame_times = t_r * np.arange(n_scans)\n    # each event last the full TR\n    duration = t_r * np.ones(n_scans)\n    # Define the events object\n    events_ = pd.DataFrame(\n        {\n            \"onset\": frame_times,\n            \"trial_type\": conditions_run,\n            \"duration\": duration,\n        }\n    )\n    # remove the rest condition and insert into the dictionary\n    events[run] = events_[events_.trial_type != \"rest\"]"
       ]
     },
     {
@@ -76,7 +76,7 @@
       },
       "outputs": [],
       "source": [
-        "z_maps = []\nconditions_label = []\nrun_label = []\n\n# Instantiate the glm\nfrom nilearn.glm.first_level import FirstLevelModel\n\nglm = FirstLevelModel(\n    t_r=TR,\n    mask_img=haxby_dataset.mask,\n    high_pass=0.008,\n    smoothing_fwhm=4,\n    memory=\"nilearn_cache\",\n)"
+        "z_maps = []\nconditions_label = []\nrun_label = []\n\n# Instantiate the glm\nfrom nilearn.glm.first_level import FirstLevelModel\n\nglm = FirstLevelModel(\n    t_r=t_r,\n    mask_img=haxby_dataset.mask,\n    high_pass=0.008,\n    smoothing_fwhm=4,\n    memory=\"nilearn_cache\",\n)"
       ]
     },
     {

dev/_downloads/580c14414efc9faa63fadc755ef45ba9/plot_design_matrix.py

@@ -18,9 +18,11 @@
 
 from nilearn.plotting import plot_design_matrix
 
-tr = 1.0  # repetition time is 1 second
+t_r = 1.0  # repetition time is 1 second
 n_scans = 128  # the acquisition comprises 128 scans
-frame_times = np.arange(n_scans) * tr  # here are the corresponding frame times
+frame_times = (
+    np.arange(n_scans) * t_r
+)  # here are the corresponding frame times
 
 # %%
 # Then we define parameters related to the experimental design.

dev/_downloads/5f6b9ef2390905869d967285eb631ef7/plot_design_matrix.ipynb

@@ -22,7 +22,7 @@
       },
       "outputs": [],
       "source": [
-        "import numpy as np\n\nfrom nilearn.plotting import plot_design_matrix\n\ntr = 1.0  # repetition time is 1 second\nn_scans = 128  # the acquisition comprises 128 scans\nframe_times = np.arange(n_scans) * tr  # here are the corresponding frame times"
+        "import numpy as np\n\nfrom nilearn.plotting import plot_design_matrix\n\nt_r = 1.0  # repetition time is 1 second\nn_scans = 128  # the acquisition comprises 128 scans\nframe_times = (\n    np.arange(n_scans) * t_r\n)  # here are the corresponding frame times"
       ]
     },
     {

dev/_downloads/664b58efb158f44104589ff41775c7c5/plot_surf_stat_map.ipynb

@@ -40,7 +40,7 @@
       },
       "outputs": [],
       "source": [
-        "# Load resting state time series from nilearn\nfrom nilearn import surface\n\ntimeseries = surface.load_surf_data(nki_dataset[\"func_left\"][0])\n\n# Extract seed region via label\npcc_region = b\"G_cingul-Post-dorsal\"\n\nimport numpy as np\n\npcc_labels = np.where(parcellation == labels.index(pcc_region))[0]\n\n# Extract time series from seed region\nseed_timeseries = np.mean(timeseries[pcc_labels], axis=0)"
+        "# Load resting state time series from nilearn\nfrom nilearn import surface\n\ntimeseries = surface.load_surf_data(nki_dataset[\"func_left\"][0])\n\n# Coercing to float is required to avoid errors withj scipy >= 0.14.0\ntimeseries = timeseries.astype(float)\n\n# Extract seed region via label\npcc_region = b\"G_cingul-Post-dorsal\"\n\nimport numpy as np\n\npcc_labels = np.where(parcellation == labels.index(pcc_region))[0]\n\n# Extract time series from seed region\nseed_timeseries = np.mean(timeseries[pcc_labels], axis=0)"
       ]
     },
     {

dev/_downloads/6e9cc5ed04e5e787060f03a222e2adcb/plot_spm_multimodal_faces.py

@@ -28,7 +28,7 @@
 # Specify timing and design matrix parameters.
 
 # repetition time, in seconds
-tr = 2.0
+t_r = 2.0
 # Sample at the beginning of each acquisition.
 slice_time_ref = 0.0
 # We use a discrete cosine transform to model signal drifts.
@@ -77,7 +77,7 @@
     n_scans = img.shape[-1]
     events = pd.read_table(subject_data[f"events{idx}"])
     # Define the sampling times for the design matrix
-    frame_times = np.arange(n_scans) * tr
+    frame_times = np.arange(n_scans) * t_r
     # Build design matrix with the reviously defined parameters
     design_matrix = make_first_level_design_matrix(
         frame_times,

dev/_downloads/925433799fdaf0968e27e70b028be9d0/plot_haxby_glm_decoding.py

@@ -30,7 +30,7 @@
 haxby_dataset = datasets.fetch_haxby()
 
 # repetition has to be known
-TR = 2.5
+t_r = 2.5
 
 # %%
 # Load the behavioral data
@@ -59,9 +59,9 @@
     # get the number of scans per run, then the corresponding
     # vector of frame times
     n_scans = len(conditions_run)
-    frame_times = TR * np.arange(n_scans)
+    frame_times = t_r * np.arange(n_scans)
     # each event last the full TR
-    duration = TR * np.ones(n_scans)
+    duration = t_r * np.ones(n_scans)
     # Define the events object
     events_ = pd.DataFrame(
         {
@@ -87,7 +87,7 @@
 from nilearn.glm.first_level import FirstLevelModel
 
 glm = FirstLevelModel(
-    t_r=TR,
+    t_r=t_r,
     mask_img=haxby_dataset.mask,
     high_pass=0.008,
     smoothing_fwhm=4,

dev/_downloads/a28e5e1d886cf6349700a2aee34d4efb/plot_spm_multimodal_faces.ipynb

@@ -40,7 +40,7 @@
       },
       "outputs": [],
       "source": [
-        "# repetition time, in seconds\ntr = 2.0\n# Sample at the beginning of each acquisition.\nslice_time_ref = 0.0\n# We use a discrete cosine transform to model signal drifts.\ndrift_model = \"Cosine\"\n# The cutoff for the drift model is 0.01 Hz.\nhigh_pass = 0.01\n# The hemodynamic response function\nhrf_model = \"spm + derivative\""
+        "# repetition time, in seconds\nt_r = 2.0\n# Sample at the beginning of each acquisition.\nslice_time_ref = 0.0\n# We use a discrete cosine transform to model signal drifts.\ndrift_model = \"Cosine\"\n# The cutoff for the drift model is 0.01 Hz.\nhigh_pass = 0.01\n# The hemodynamic response function\nhrf_model = \"spm + derivative\""
       ]
     },
     {
@@ -112,7 +112,7 @@
       },
       "outputs": [],
       "source": [
-        "for idx, img in enumerate(fmri_img, start=1):\n    # Build experimental paradigm\n    n_scans = img.shape[-1]\n    events = pd.read_table(subject_data[f\"events{idx}\"])\n    # Define the sampling times for the design matrix\n    frame_times = np.arange(n_scans) * tr\n    # Build design matrix with the reviously defined parameters\n    design_matrix = make_first_level_design_matrix(\n        frame_times,\n        events,\n        hrf_model=hrf_model,\n        drift_model=drift_model,\n        high_pass=high_pass,\n    )\n\n    # put the design matrices in a list\n    design_matrices.append(design_matrix)"
+        "for idx, img in enumerate(fmri_img, start=1):\n    # Build experimental paradigm\n    n_scans = img.shape[-1]\n    events = pd.read_table(subject_data[f\"events{idx}\"])\n    # Define the sampling times for the design matrix\n    frame_times = np.arange(n_scans) * t_r\n    # Build design matrix with the reviously defined parameters\n    design_matrix = make_first_level_design_matrix(\n        frame_times,\n        events,\n        hrf_model=hrf_model,\n        drift_model=drift_model,\n        high_pass=high_pass,\n    )\n\n    # put the design matrices in a list\n    design_matrices.append(design_matrix)"
       ]
     },
     {

dev/_downloads/a5651a769eacc82cae220652665c23ea/plot_hrf.py

@@ -57,12 +57,12 @@
 from scipy.stats import gamma
 
 
-def mion_response_function(tr, oversampling=16, onset=0.0):
+def mion_response_function(t_r, oversampling=16, onset=0.0):
     """Implement the MION response function model.
 
     Parameters
     ----------
-    tr: float
+    t_r: float
         scan repeat time, in seconds
     oversampling: int, optional
         temporal oversampling factor
@@ -71,10 +71,10 @@ def mion_response_function(tr, oversampling=16, onset=0.0):
 
     Returns
     -------
-    response_function: array of shape(length / tr * oversampling, dtype=float)
+    response_function: array of shape(length / t_r * oversampling, dtype=float)
         response_function sampling on the oversampled time grid
     """
-    dt = tr / oversampling
+    dt = t_r / oversampling
     time_stamps = np.linspace(
         0, time_length, np.rint(time_length / dt).astype(int)
     )
@@ -91,25 +91,25 @@ def mion_response_function(tr, oversampling=16, onset=0.0):
     return response_function
 
 
-def mion_time_derivative(tr, oversampling=16.0):
+def mion_time_derivative(t_r, oversampling=16.0):
     """Implement the MION time derivative response function model.
 
     Parameters
     ----------
-    tr: float
+    t_r: float
         scan repeat time, in seconds
     oversampling: int, optional
         temporal oversampling factor, optional
 
     Returns
     -------
-    drf: array of shape(time_length / tr * oversampling, dtype=float)
+    drf: array of shape(time_length / t_r * oversampling, dtype=float)
         derived_response_function sampling on the provided grid
     """
     do = 0.1
     drf = (
-        mion_response_function(tr, oversampling)
-        - mion_response_function(tr, oversampling, do)
+        mion_response_function(t_r, oversampling)
+        - mion_response_function(t_r, oversampling, do)
     ) / do
 
     return drf

dev/_downloads/ec36d74d1c921804ff3412db1424031b/plot_surf_stat_map_experimental.py

@@ -91,6 +91,9 @@
 # Load resting state time series from nilearn
 timeseries = nki_dataset[0].data.parts[hemi].T
 
+# Coercing to float is required to avoid errors withj scipy >= 0.14.0
+timeseries = timeseries.astype(float)
+
 # Extract seed region via label
 pcc_region = "G_cingul-Post-dorsal"
 

dev/_downloads/efb4722f21ca44d748365b3274f7c9db/plot_hrf.ipynb

@@ -69,7 +69,7 @@
       },
       "outputs": [],
       "source": [
-        "from scipy.stats import gamma\n\n\ndef mion_response_function(tr, oversampling=16, onset=0.0):\n    \"\"\"Implement the MION response function model.\n\n    Parameters\n    ----------\n    tr: float\n        scan repeat time, in seconds\n    oversampling: int, optional\n        temporal oversampling factor\n    onset: float, optional\n        hrf onset time, in seconds\n\n    Returns\n    -------\n    response_function: array of shape(length / tr * oversampling, dtype=float)\n        response_function sampling on the oversampled time grid\n    \"\"\"\n    dt = tr / oversampling\n    time_stamps = np.linspace(\n        0, time_length, np.rint(time_length / dt).astype(int)\n    )\n    time_stamps -= onset\n\n    # parameters of the gamma function\n    delay = 1.55\n    dispersion = 5.5\n\n    response_function = gamma.pdf(time_stamps, delay, loc=0, scale=dispersion)\n    response_function /= response_function.sum()\n    response_function *= -1\n\n    return response_function\n\n\ndef mion_time_derivative(tr, oversampling=16.0):\n    \"\"\"Implement the MION time derivative response function model.\n\n    Parameters\n    ----------\n    tr: float\n        scan repeat time, in seconds\n    oversampling: int, optional\n        temporal oversampling factor, optional\n\n    Returns\n    -------\n    drf: array of shape(time_length / tr * oversampling, dtype=float)\n        derived_response_function sampling on the provided grid\n    \"\"\"\n    do = 0.1\n    drf = (\n        mion_response_function(tr, oversampling)\n        - mion_response_function(tr, oversampling, do)\n    ) / do\n\n    return drf"
+        "from scipy.stats import gamma\n\n\ndef mion_response_function(t_r, oversampling=16, onset=0.0):\n    \"\"\"Implement the MION response function model.\n\n    Parameters\n    ----------\n    t_r: float\n        scan repeat time, in seconds\n    oversampling: int, optional\n        temporal oversampling factor\n    onset: float, optional\n        hrf onset time, in seconds\n\n    Returns\n    -------\n    response_function: array of shape(length / t_r * oversampling, dtype=float)\n        response_function sampling on the oversampled time grid\n    \"\"\"\n    dt = t_r / oversampling\n    time_stamps = np.linspace(\n        0, time_length, np.rint(time_length / dt).astype(int)\n    )\n    time_stamps -= onset\n\n    # parameters of the gamma function\n    delay = 1.55\n    dispersion = 5.5\n\n    response_function = gamma.pdf(time_stamps, delay, loc=0, scale=dispersion)\n    response_function /= response_function.sum()\n    response_function *= -1\n\n    return response_function\n\n\ndef mion_time_derivative(t_r, oversampling=16.0):\n    \"\"\"Implement the MION time derivative response function model.\n\n    Parameters\n    ----------\n    t_r: float\n        scan repeat time, in seconds\n    oversampling: int, optional\n        temporal oversampling factor, optional\n\n    Returns\n    -------\n    drf: array of shape(time_length / t_r * oversampling, dtype=float)\n        derived_response_function sampling on the provided grid\n    \"\"\"\n    do = 0.1\n    drf = (\n        mion_response_function(t_r, oversampling)\n        - mion_response_function(t_r, oversampling, do)\n    ) / do\n\n    return drf"
       ]
     },
     {

dev/_downloads/f8c984d44c420ec0ae630f5186f526d2/plot_surf_stat_map_experimental.ipynb

@@ -40,7 +40,7 @@
       },
       "outputs": [],
       "source": [
-        "# Load resting state time series from nilearn\ntimeseries = nki_dataset[0].data.parts[hemi].T\n\n# Extract seed region via label\npcc_region = \"G_cingul-Post-dorsal\"\n\nimport numpy as np\n\npcc_labels = np.where(parcellation == labels.index(pcc_region))[0]\n\n# Extract time series from seed region\nseed_timeseries = np.mean(timeseries[pcc_labels], axis=0)"
+        "# Load resting state time series from nilearn\ntimeseries = nki_dataset[0].data.parts[hemi].T\n\n# Coercing to float is required to avoid errors withj scipy >= 0.14.0\ntimeseries = timeseries.astype(float)\n\n# Extract seed region via label\npcc_region = \"G_cingul-Post-dorsal\"\n\nimport numpy as np\n\npcc_labels = np.where(parcellation == labels.index(pcc_region))[0]\n\n# Extract time series from seed region\nseed_timeseries = np.mean(timeseries[pcc_labels], axis=0)"
       ]
     },
     {

dev/_sources/auto_examples/00_tutorials/plot_3d_and_4d_niimg.rst.txt

@@ -126,7 +126,7 @@ statistical map:
  .. code-block:: none
 
 
-    <nilearn.plotting.displays._slicers.OrthoSlicer object at 0x7f5a637e73e0>
+    <nilearn.plotting.displays._slicers.OrthoSlicer object at 0x7f5faed782f0>
 
 
 
@@ -155,7 +155,7 @@ Visualizing works better with a threshold
  .. code-block:: none
 
 
-    <nilearn.plotting.displays._slicers.OrthoSlicer object at 0x7f5a563dda90>
+    <nilearn.plotting.displays._slicers.OrthoSlicer object at 0x7f5fa75a2c90>
 
 
 
@@ -183,7 +183,7 @@ correspondence between rest and task
  .. code-block:: none
 
     Downloading data from https://www.fmrib.ox.ac.uk/datasets/brainmap+rsns/PNAS_Smith09_rsn10.nii.gz ...
-    Downloaded 1482752 of 7565016 bytes (19.6%,    4.3s remaining) ...done. (2 seconds, 0 min)
+     ...done. (1 seconds, 0 min)
 
     '/home/runner/work/nilearn/nilearn/nilearn_data/smith_2009/PNAS_Smith09_rsn10.nii.gz'
 
@@ -266,7 +266,7 @@ We can then plot it
  .. code-block:: none
 
 
-    <nilearn.plotting.displays._slicers.OrthoSlicer object at 0x7f5a648f2b70>
+    <nilearn.plotting.displays._slicers.OrthoSlicer object at 0x7f5fa6be9fa0>
 
 
 
@@ -477,7 +477,7 @@ image.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 18.751 seconds)
+   **Total running time of the script:** (0 minutes 18.741 seconds)
 
 **Estimated memory usage:**  212 MB