Network plots

deeptime/covariance/util/_moments.py

@@ -50,7 +50,7 @@
       using the dot product on sliced arrays (option 2). Exceptions are when the
       data is extremely sparse, such that only a few columns are selected.
     * Copying and subselecting columns (option 3) is only faster than the full
-      dot product (option 1), if 50% or less columns are selected. This observation
+      dot product (option 1), if 50% or fewer columns are selected. This observation
       is roughly independent of N.
     * The observations above are valid for  matrices (T x N) that are sufficiently
       large. We assume that "sufficiently large" means that they don't fully fit
@@ -66,7 +66,7 @@
        number of constant column candidates drops below the minimum number, to
        avoid wasting time on the decision.
     3. Subselect the desired columns and copy the data to a new array X0 (Y0).
-    4. Run operation on the new array X0 (Y0), including in-place substraction
+    4. Run operation on the new array X0 (Y0), including in-place subtraction
        of the mean if needed.
 
 """

deeptime/decomposition/__init__.py

@@ -66,6 +66,7 @@
     vamp_score_cv
 
     cvsplit_trajs
+    blocksplit_trajs
 
     deep.TVAEEncoder
     deep.koopman_matrix
@@ -76,7 +77,7 @@
     deep.kvad_score
 """
 
-from ._score import vamp_score, vamp_score_data, vamp_score_cv, cvsplit_trajs
+from ._score import vamp_score, vamp_score_data, vamp_score_cv, cvsplit_trajs, blocksplit_trajs
 from ._tica import TICA
 from ._vamp import VAMP
 from ._koopman import TransferOperatorModel, CovarianceKoopmanModel

deeptime/markov/msm/_markov_state_model.py

@@ -771,7 +771,7 @@ def correlation(self, a, b=None, maxtime=None, k=None, ncv=None):
         >>> times, acf = M.correlation(a)
         >>>
         >>> import matplotlib.pylab as plt # doctest: +SKIP
-        >>> plt.plot(times, acf)  # doctest: +SKIP
+        >>> plt.plot(times,acf)  # doctest: +SKIP
         """
         # input checking is done in low-level API
         # compute number of tau steps

deeptime/plots/__init__.py

@@ -9,8 +9,13 @@
     plot_ck_test
     plot_energy2d
     Energy2dPlot
+
+    plot_adjacency
+    plot_markov_model
+    Network
 """
 
 from .implied_timescales import plot_implied_timescales
 from .chapman_kolmogorov import plot_ck_test
 from .energy import plot_energy2d, Energy2dPlot
+from .network import Network, plot_adjacency, plot_markov_model

deeptime/plots/chapman_kolmogorov.py

@@ -93,6 +93,7 @@ def n_tests(self):
         return len(self._tests)
 
 
+@plotting_function()
 def plot_ck_test(data: ChapmanKolmogorovTest, height=2.5, aspect=1.,
                  conf: float = 0.95, color=None, grid: CKTestGrid = None, legend=True,
                  xlabel='lagtime (steps)', ylabel='probability', y01=True, sharey=True, **plot_kwargs):

deeptime/plots/energy.py

@@ -3,7 +3,11 @@
 
 
 class Energy2dPlot:
-    r""" The result of a :meth:`plot_energy2d` call.
+    r""" The result of a :meth:`plot_energy2d` call. Instances of this class can be unpacked like a tuple:
+
+    >>> import numpy as np
+    >>> from deeptime.util import energy2d
+    >>> ax, contour, cbar = plot_energy2d(energy2d(*np.random.uniform(size=(100 ,2)).T))
 
     Parameters
     ----------
@@ -25,6 +29,14 @@ def __init__(self, ax, contour, colorbar=None):
         self.contour = contour
         self.colorbar = colorbar
 
+    # makes object unpackable
+    def __len__(self):
+        return 3
+
+    # makes object unpackable
+    def __iter__(self):
+        return iter((self.ax, self.contour, self.colorbar))
+
 
 @plotting_function()
 def plot_energy2d(energies: EnergyLandscape2d, ax=None, levels=100, contourf_kws=None, cbar=True,