Add machinery to install module directory.

findoutlie/__init__.py

@@ -0,0 +1,4 @@
+""" Init for findoutlie module
+"""
+
+__version__ = '0.1a0'

findoutlie/metrics.py

@@ -0,0 +1,32 @@
+""" Scan outlier metrics
+"""
+
+# Any imports you need
+# +++your code here+++
+
+
+def dvars(img):
+    """ Calculate dvars metric on Nibabel image `img`
+
+    The dvars calculation between two volumes is defined as the square root of
+    (the mean of the (voxel differences squared)).
+
+    Parameters
+    ----------
+    img : nibabel image
+
+    Returns
+    -------
+    dvals : 1D array
+        One-dimensional array with n-1 elements, where n is the number of
+        volumes in `img`.
+    """
+    # Hint: remember 'axis='.  For example:
+    # In [2]: arr = np.array([[2, 3, 4], [5, 6, 7]])
+    # In [3]: np.mean(arr, axis=1)
+    # Out[2]: array([3., 6.])
+    #
+    # You may be be able to solve this in four lines, without a loop.
+    # But solve it any way you can.
+    # This is a placeholder, replace it to write your solution.
+    raise NotImplementedError('Code up this function')

findoutlie/tests/test_dvars.py

@@ -0,0 +1,35 @@
+""" Test dvars implementation
+
+You can run the tests from the root directory (containing ``README.md``) with::
+
+    python3 -m pytest .
+"""
+
+import numpy as np
+
+import nibabel as nib
+
+import nipraxis as npx
+
+from findoutlie.metrics import dvars
+
+
+TEST_FNAME = npx.fetch_file('ds114_sub009_t2r1.nii')
+
+
+def test_dvars():
+    img = nib.load(TEST_FNAME)
+    n_trs = img.shape[-1]
+    n_voxels = np.prod(img.shape[:-1])
+    dvals = dvars(img)
+    assert len(dvals) == n_trs - 1
+    # Calculate the values the long way round
+    data = img.get_fdata()
+    prev_vol = data[..., 0]
+    long_dvals = []
+    for i in range(1, n_trs):
+        this_vol = data[..., i]
+        d = this_vol - prev_vol
+        long_dvals.append(np.sqrt(np.sum(d ** 2) / n_voxels))
+        prev_vol = this_vol
+    assert np.allclose(dvals, long_dvals)

pyproject.toml

@@ -0,0 +1,20 @@
+[build-system]
+requires = ["flit_core >=2,<4"]
+build-backend = "flit_core.buildapi"
+
+[tool.flit.metadata]
+module = "findoutlie"
+author = "The team"
+author-email = "team@email"
+home-page = "https://github.com/nipraxis-fall-2022/diagnostics-example"
+classifiers = ["License :: OSI Approved :: BSD License",
+               "Programming Language :: Python :: 3"]
+description-file="README.md"
+# Check against requirements.txt
+requires = [
+    'nibabel',
+    'numpy',
+    'scipy',
+    'matplotlib',
+]
+requires-python=">=3.6"

pytest.ini

@@ -0,0 +1,2 @@
+[pytest]
+addopts = --ignore-glob=solutions/*

requirements.txt

@@ -0,0 +1,5 @@
+nibabel
+numpy
+scipy
+matplotlib
+nipraxis

solutions/.solutions.toml

@@ -1,5 +1,8 @@
 # Where the exercises etc go, given the solutions
 
+[solution.metrics]
+out_path = '{one_down}/findoutlie/metrics.py'
+
 [solution.validate_data]
 out_path = '{one_down}/scripts/validate_data.py'
 

solutions/.write_dvars.sh

@@ -0,0 +1,4 @@
+#!/bin/bash
+nxt-proc-solutions write-solutions
+pytest findoutlie
+nxt-proc-solutions write

solutions/metrics.py

@@ -0,0 +1,43 @@
+""" Scan outlier metrics
+"""
+
+# Any imports you need
+# LAB(begin solution)
+import numpy as np
+# LAB(replace solution)
+# +++your code here+++
+# LAB(end solution)
+
+
+def dvars(img):
+    """ Calculate dvars metric on Nibabel image `img`
+
+    The dvars calculation between two volumes is defined as the square root of
+    (the mean of the (voxel differences squared)).
+
+    Parameters
+    ----------
+    img : nibabel image
+
+    Returns
+    -------
+    dvals : 1D array
+        One-dimensional array with n-1 elements, where n is the number of
+        volumes in `img`.
+    """
+    # Hint: remember 'axis='.  For example:
+    # In [2]: arr = np.array([[2, 3, 4], [5, 6, 7]])
+    # In [3]: np.mean(arr, axis=1)
+    # Out[2]: array([3., 6.])
+    #
+    # You may be be able to solve this in four lines, without a loop.
+    # But solve it any way you can.
+    # LAB(begin solution)
+    data = img.get_fdata()
+    vx_by_time = np.reshape(data, (-1, data.shape[-1]))
+    time_diffs = np.diff(vx_by_time, axis=1)
+    return np.sqrt(np.mean(time_diffs ** 2, axis=0))
+    # LAB(replace solution)
+    # This is a placeholder, replace it to write your solution.
+    raise NotImplementedError('Code up this function')
+    # LAB(end solution)

test_requirements.txt

@@ -0,0 +1,3 @@
+# Test requirements
+-r requirements.txt
+pytest