Youpi ! RPCA, EM_sampler, docs et mocker

Author: Gsaes

docs/EM_sampler.rst

@@ -210,16 +210,7 @@ The resulting imputed data is the final imputed array, obtained at convergence.
 Multivariate time series
 ************************
 
-To explicitely take into account the temporal aspect of the data 
-(temporal correlations), we construct an extended matrix :math:`\mathbf{X}^{ext}` 
-by considering the shifted columns, i.e.
-:math:`\mathbf{X}^{ext} := [\mathbf{X}, \mathbf{X}^{s+1}, \mathbf{X}^{s-1}]` where
-:math:`\mathbf{X}^{s+1}` (resp. :math:`\mathbf{X}^{s-1}`) is the :math:`\mathbf{X}` matrix 
-where all columns are shifted +1 for one step backward in time (resp. -1 for one step forward in time).
-The covariance matrix :math:`\mathbf{\Sigma}^{ext}` is therefore richer in information since the presence of additional 
-(temporal) correlations.
-
-.. image:: images/extended_matrix.png
+The class VAR1EM takes into account the temporal structure of the data by using likelihood based on a VAR(1) assumption. Stay tuned for more details!
 
 
 

docs/RPCA.rst

@@ -77,7 +77,8 @@ Some algorithms are implemented:
    & \text{s.t.} \quad \mathbf{D} = \mathbf{X} + \mathbf{A}
    \end{align*}
 
-* :class:`GraphRPCA` class (based on this `paper <https://arxiv.org/abs/1507.08173>`__). The optimisation problem is the following
+..
+   * :class:`GraphRPCA` class (based on this `paper <https://arxiv.org/abs/1507.08173>`__). The optimisation problem is the following
 
 .. math:: 
 
@@ -95,7 +96,8 @@ By defining :math:`\Vert \mathbf{XH_k} \Vert_p` is either :math:`\Vert \mathbf{X
    \text{minimise} \quad \Vert P_{\Omega}(\mathbf{X}+\mathbf{A}-\mathbf{D}) \Vert_F^2 + \lambda_1 \Vert \mathbf{X} \Vert_* + \lambda_2 \Vert \mathbf{A} \Vert_1 + \sum_{k=1}^K \eta_k \Vert \mathbf{XH_k} \Vert_p
 
 
-* :class:`OnlineTemporalRPCA` class. This class implements the online version of the above problem using stochastic optimisation (based on this `paper <https://www.hindawi.com/journals/jat/2018/7191549/>`__  and this `paper <https://dl.acm.org/doi/10.5555/2999611.2999657>`__). This allows to deal with large ammount of data or data that arrives continuously and does not assume a stable subspace.
+..
+   * :class:`OnlineTemporalRPCA` class. This class implements the online version of the above problem using stochastic optimisation (based on this `paper <https://www.hindawi.com/journals/jat/2018/7191549/>`__  and this `paper <https://dl.acm.org/doi/10.5555/2999611.2999657>`__). This allows to deal with large ammount of data or data that arrives continuously and does not assume a stable subspace.
 
 
 The operator :math:`P_{\Omega}` is the projection operator such that 

qolmat/utils/plot.py

@@ -157,7 +157,7 @@ def plot_images(
 
 def make_ellipses(
     X: np.ndarray,
-    ax: Any,
+    ax: mpl.axes.Axes,
     color: Union[str, Tuple[float, float, float]],
 ):
     """Draw ellipses on a figure
@@ -223,9 +223,9 @@ def compare_covariances(
 
 def multibar(
     df: pd.DataFrame,
-    ax: Any = None,
+    ax: Optional[mpl.axes.Axes] = None,
     orientation: str = "vertical",
-    colors=None,
+    colors: Any = None,
     decimals: float = 0,
 ):
     """Create a multi-bar graph to represent the values of the different dataframe columns.

tests/utils/test_plot.py

@@ -6,6 +6,7 @@
 import pytest
 import scipy.sparse
 from qolmat.utils import plot
+from pytest_mock.plugin import MockerFixture
 
 plt.switch_backend("Agg")
 
@@ -33,48 +34,57 @@
 }
 
 
-def show_fake():
-    pass
-
-
-@pytest.fixture(autouse=True)
-def patch_matplotlib_show(monkeypatch):
-    monkeypatch.setattr(plt, "show", show_fake)
-
-
 @pytest.mark.parametrize("list_matrices", [list_matrices])
-def test_utils_plot_plot_matrices(list_matrices: List[np.ndarray]) -> None:
+def test_utils_plot_plot_matrices(list_matrices: List[np.ndarray], mocker: MockerFixture) -> None:
+    mocker.patch("matplotlib.pyplot.savefig")
+    mocker.patch("matplotlib.pyplot.show")
     plot.plot_matrices(list_matrices=list_matrices, title="title")
     assert len(plt.gcf().get_axes()) > 0
+    assert plt.savefig.call_count == 1
     plt.close("all")
 
 
 @pytest.mark.parametrize("list_signals", [list_signals])
-def test_utils_plot_plot_signal(list_signals: List[np.ndarray]) -> None:
+def test_utils_plot_plot_signal(list_signals: List[np.ndarray], mocker: MockerFixture) -> None:
+    mocker.patch("matplotlib.pyplot.savefig")
+    mocker.patch("matplotlib.pyplot.show")
     plot.plot_signal(list_signals=list_signals, ylabel="ylabel", title="title")
     assert len(plt.gcf().get_axes()) > 0
+    assert plt.savefig.call_count == 1
     plt.close("all")
 
 
 @pytest.mark.parametrize("M, A, E, index_array, dims", [(M, A, E, [0, 1, 2], (10, 10))])
 def test__utils_plot_plot_images(
-    M: np.ndarray, A: np.ndarray, E: np.ndarray, index_array: List[int], dims: Tuple[int, int]
+    M: np.ndarray,
+    A: np.ndarray,
+    E: np.ndarray,
+    index_array: List[int],
+    dims: Tuple[int, int],
+    mocker: MockerFixture,
 ):
+    mocker.patch("matplotlib.pyplot.savefig")
+    mocker.patch("matplotlib.pyplot.show")
     plot.plot_images(M, A, E, index_array, dims, filename="filename")
     assert len(plt.gcf().get_axes()) > 0
+    assert plt.savefig.call_count == 1
     plt.close("all")
 
 
 @pytest.mark.parametrize("X", [X])
-def test_utils_plot_make_ellipses(X: np.ndarray):
+def test_utils_plot_make_ellipses(X: np.ndarray, mocker: MockerFixture):
+    mocker.patch("matplotlib.pyplot.show")
     ax = plt.gca()
     plot.make_ellipses(X, ax, color="blue")
     assert len(plt.gcf().get_axes()) > 0
     plt.close("all")
 
 
 @pytest.mark.parametrize("df1,df2", [(df1, df2)])
-def test_utils_plot_compare_covariances(df1: pd.DataFrame, df2: pd.DataFrame):
+def test_utils_plot_compare_covariances(
+    df1: pd.DataFrame, df2: pd.DataFrame, mocker: MockerFixture
+):
+    mocker.patch("matplotlib.pyplot.show")
     ax = plt.gca()
     plot.compare_covariances(df1, df2, "x", "y", ax)
     assert len(plt.gcf().get_axes()) > 0
@@ -83,14 +93,16 @@ def test_utils_plot_compare_covariances(df1: pd.DataFrame, df2: pd.DataFrame):
 
 @pytest.mark.parametrize("df", [df])
 @pytest.mark.parametrize("orientation", ["horizontal", "vertical"])
-def test_utils_plot_multibar(df: pd.DataFrame, orientation: str):
+def test_utils_plot_multibar(df: pd.DataFrame, orientation: str, mocker: MockerFixture):
+    mocker.patch("matplotlib.pyplot.show")
     plot.multibar(df, orientation=orientation)
     assert len(plt.gcf().get_axes()) > 0
     plt.close("all")
 
 
 @pytest.mark.parametrize("df", [df])
-def test_utils_plot_plot_imputations(df: pd.DataFrame):
+def test_utils_plot_plot_imputations(df: pd.DataFrame, mocker: MockerFixture):
+    mocker.patch("matplotlib.pyplot.show")
     plot.plot_imputations(df, dict_df_imputed)
     assert len(plt.gcf().get_axes()) > 0
     plt.close("all")