Merge pull request #302 from rsagroup/conda-package

Conda packaging

.conda/meta.yaml

@@ -0,0 +1,61 @@
+{% set name = "rsatoolbox" %}
+{% set version = "0.1.2" %}
+
+package:
+  name: {{ name|lower }}
+  version: {{ version }}
+
+source:
+  url: https://pypi.io/packages/source/{{ name[0] }}/{{ name }}/rsatoolbox-{{ version }}.tar.gz
+  sha256: fee6e0134c345f4d7e7f2d7bdde20b13724e9d0d32f89e8b22b7fc17ed21ef9a
+
+build:
+  skip: true  # [py<37 or py>310 or py2k]
+  script: {{ PYTHON }} -m pip install . -vv
+  number: 0
+
+requirements:
+  build:
+    - {{ compiler('c') }}
+  host:
+    - python
+    - cython # >=3.0.0a11
+    - numpy >=1.21.2
+    - scipy
+    - setuptools
+    - setuptools-scm
+    - pip
+  run:
+    - python
+    - {{ pin_compatible('numpy') }}
+    - {{ pin_compatible('scipy') }}
+    - pandas
+    - scikit-learn
+    - scikit-image
+    - matplotlib-base
+    - h5py
+    - joblib
+    - tqdm
+    - coverage
+
+
+test:
+  imports:
+    - rsatoolbox
+  commands:
+    - pip check
+    - python -m unittest -v rsatoolbox.test
+  requires:
+    - pip
+
+about:
+  home: https://github.com/rsagroup/rsatoolbox
+  summary: Representational Similarity Analysis (RSA) in Python
+  license: MIT
+  license_file: LICENSE
+  doc_url: https://rsatoolbox.readthedocs.io/
+  dev_url: https://github.com/rsagroup/rsatoolbox
+
+extra:
+  recipe-maintainers:
+    - ilogue

.github/workflows/testbuild.yml

@@ -29,6 +29,8 @@ jobs:
       run: pip install .
     - name: Install test dependencies
       run: pip install -r tests/requirements.txt
+    - name: Skeleton tests
+      run: python -m unittest -v rsatoolbox.test
     - name: Unit tests
       run: pytest
 
@@ -64,6 +66,8 @@ jobs:
       if: matrix.os == 'windows-latest'
     - name: Install test dependencies
       run: pip install -r tests/requirements.txt
+    - name: Skeleton tests
+      run: python -m unittest -v rsatoolbox.test
     - name: Unit tests
       run: pytest
     - name: Check package compliance

README.md

@@ -20,6 +20,12 @@ To install the latest stable version of rsatoolbox with pip:
 pip install rsatoolbox
 ```
 
+or with conda:
+
+```sh
+conda install -c conda-forge rsatoolbox
+```
+
 
 here is a simple code sample:
 

docs/source/getting_started.rst

@@ -17,6 +17,13 @@ To get the bleeding edge or "pre-release" version:
     pip install --pre rsatoolbox
 
 
+Or if you're using a Conda environment:
+
+.. code-block:: sh
+
+    conda install -c conda-forge rsatoolbox
+
+
 To use rsatoolbox:
 
 .. code-block:: python

requirements.txt

@@ -1,10 +1,9 @@
-coverage
 numpy>=1.21.2
 scipy
 scikit-learn
 scikit-image
-tqdm
-h5py
+pandas
 matplotlib
+h5py
+tqdm
 joblib
-pandas

src/rsatoolbox/test.py

@@ -0,0 +1,102 @@
+"""Skeleton test module
+
+The full collection of unit and acceptance tests for rsatoolbox is kept
+in a separate package that is not part of our distributables. It can be run
+by checking out the rsatoolbox git repository.
+
+The tests in this module are a limited number of basic so-called skeleton
+tests, which check that the library and all its dependencies are installed
+correctly. It is not exhaustive and assumes that unittests have passed
+for other most package formats.
+If rsatoolbox is installed, the tests can be run with:
+
+`python -m unittest rsatoolbox.test`
+
+These tests have to:
+
+- not have any dependencies outside of direct rsatoolbox runtime dependencies
+- be fast (a few seconds)
+- test interfaces that depend on external packages
+- test compiled code
+
+In other words they have to check that all the moving parts are there,
+without looking at very specific calculation outcomes.
+"""
+# pylint: disable=import-outside-toplevel
+from unittest import TestCase
+
+
+class SkeletonTests(TestCase):
+    """Toolbox skeleton tests to ensure correct packaging and installation
+    """
+
+    def setUp(self):
+        """Create basic RDMs and Dataset objects for all tests
+        """
+        from numpy import asarray, ones
+        from numpy.testing import assert_almost_equal
+        from rsatoolbox.data.dataset import Dataset
+        from rsatoolbox.rdm.rdms import RDMs
+        self.data = Dataset(asarray([[0, 0], [1, 1], [2.0, 2.0]]))
+        self.rdms = RDMs(asarray([[1.0, 2, 3], [3, 4, 5]]))
+        self.larger_rdms = RDMs(ones([3, 10]))
+        self.array = asarray
+        self.arrayAlmostEqual = assert_almost_equal
+
+    def test_calc_compiled(self):
+        """Covers similarity calculation with compiled code
+        """
+        from rsatoolbox.rdm.calc_unbalanced import calc_rdm_unbalanced
+        rdms = calc_rdm_unbalanced(self.data)
+        self.arrayAlmostEqual(rdms.dissimilarities, self.array([[1, 4, 1]]))
+
+    def test_model_fit(self):
+        """Covers model fitting with scipy
+        """
+        from rsatoolbox.model.model import ModelWeighted
+        from rsatoolbox.model.fitter import fit_optimize
+        theta = fit_optimize(ModelWeighted('F', self.rdms), self.rdms)
+        self.arrayAlmostEqual(theta, [0.88, 0.47], decimal=2)
+
+    def test_plotting_with_mpl(self):
+        """Covers Matplotlib usage
+        """
+        from rsatoolbox.vis.rdm_plot import show_rdm
+        show_rdm(self.rdms)
+
+    def test_mds(self):
+        """Covers sklearn and Matplotlib usage
+        """
+        from rsatoolbox.vis.scatter_plot import show_MDS
+        show_MDS(self.rdms)
+
+    def test_evaluate(self):
+        """Covers tqdm usage and evaluate functionality
+        """
+        from rsatoolbox.inference import eval_fixed
+        from rsatoolbox.model import ModelFixed
+        model = ModelFixed('G', self.array(list(range(10))))
+        result = eval_fixed(model, self.larger_rdms)
+        self.assertAlmostEqual(result.test_zero()[0], 0)
+
+    def test_pandas_io(self):
+        """Covers pandas usage
+        """
+        df = self.rdms.to_df()
+        self.arrayAlmostEqual(
+            self.array(df.loc[:, 'dissimilarity'].values),
+            self.rdms.dissimilarities.ravel()
+        )
+
+    def test_hdf_io(self):
+        """Covers h5py library use
+        """
+        from io import BytesIO
+        from rsatoolbox.rdm.rdms import load_rdm
+        fhandle = BytesIO()
+        self.rdms.save(fhandle, file_type='hdf5')
+        reconstituted_rdms = load_rdm(fhandle, file_type='hdf5')
+        self.arrayAlmostEqual(
+            self.rdms.dissimilarities,
+            reconstituted_rdms.dissimilarities
+        )