DOC: whats_new + example for stepped fill_between

.travis.yml

@@ -67,7 +67,7 @@ install:
         pip install $PRE python-dateutil $NUMPY pyparsing pillow sphinx!=1.3.0;
     fi
   # Always install from pypi
-  - pip install $PRE nose pep8
+  - pip install $PRE nose pep8 cycler
 
   # Install mock on python 2. Python 2.6 requires mock 1.0.1
   # Since later versions have dropped support

doc/users/whats_new/2015-05_filledstep.rst

@@ -0,0 +1,12 @@
+Add step kwmargs to fill_between
+--------------------------------
+
+Added ``step`` kwarg to `Axes.fill_between` to allow to fill between
+lines drawn using the 'step' draw style.  The values of ``step`` match
+those of the ``where`` kwarg of `Axes.step`.  The asymmetry of of the
+kwargs names is not ideal, but `Axes.fill_between` already has a
+``where`` kwarg.
+
+This is particularly useful for plotting pre-binned histograms.
+
+.. plot:: mpl_examples/api/filled_step.py

examples/api/filled_step.py

@@ -0,0 +1,206 @@
+import itertools
+from functools import partial
+
+import numpy as np
+import matplotlib.pyplot as plt
+from cycler import cycler
+from six.moves import zip
+
+
+def filled_hist(ax, edges, values, bottoms=None, orientation='v',
+                **kwargs):
+    """
+    Draw a histogram as a stepped patch.
+
+    Extra kwargs are passed through to `fill_between`
+
+    Parameters
+    ----------
+    ax : Axes
+        The axes to plot to
+
+    edges : array
+        A an n+1 array giving the left edges of each bin and the
+        right edge of the last bin.
+
+    values : array
+        A length n array of bin
+
+    bottoms : scalar or array, optional
+        A length n array of the bottom of the bars.  If None, zero is used.
+
+    orientation : {'v', 'h'}
+       Orientation of the histogram.  'v' is default value an has
+       the bars increasing in the positive y-direction.
+
+    Returns
+    -------
+    ret : PolyCollection
+        Added artist
+    """
+    print(orientation)
+    if orientation not in set('hv'):
+        raise ValueError("orientation must be in {'h', 'v'} "
+                         "not {o}".format(o=orientation))
+
+    kwargs.setdefault('step', 'post')
+    edges = np.asarray(edges)
+    values = np.asarray(values)
+    if len(edges) - 1 != len(values):
+        raise ValueError('Must provide one more bin edge than value not: '
+                         'len(edges): {lb} len(values): {lv}'.format(
+                             lb=len(edges), lv=len(values)))
+
+    if bottoms is None:
+        bottoms = np.zeros_like(values)
+    if np.isscalar(bottoms):
+        bottoms = np.ones_like(values) * bottoms
+
+    values = np.r_[values, values[-1]]
+    bottoms = np.r_[bottoms, bottoms[-1]]
+    if orientation == 'h':
+        return ax.fill_betweenx(edges, values, bottoms, **kwargs)
+    elif orientation == 'v':
+        return ax.fill_between(edges, values, bottoms, **kwargs)
+    else:
+        raise AssertionError("you should never be here")
+
+
+def stack_hist(ax, stacked_data, sty_cycle, bottoms=None,
+               hist_func=None, labels=None,
+               plot_func=None, plot_kwargs=None):
+    """
+    ax : axes.Axes
+        The axes to add artists too
+
+    stacked_data : array or Mapping
+        A (N, M) shaped array.  The first dimension will be iterated over to
+        compute histograms row-wise
+
+    sty_cycle : Cycler or operable of dict
+        Style to apply to each set
+
+    bottoms : array, optional
+        The initial positions of the bottoms, defaults to 0
+
+    hist_func : callable, optional
+        Must have signature `bin_vals, bin_edges = f(data)`.
+        `bin_edges` expected to be one longer than `bin_vals`
+
+    labels : list of str, optional
+        The label for each set.
+
+        If not given and stacked data is an array defaults to 'default set {n}'
+
+        If stacked_data is a mapping, and labels is None, default to the keys
+        (which may come out in a random order).
+
+        If stacked_data is a mapping and labels is given then only
+        the columns listed by be plotted.
+
+    plot_func : callable, optional
+        Function to call to draw the histogram must have signature:
+
+          ret = plot_func(ax, edges, top, bottoms=bottoms,
+                          label=label, **kwargs)
+
+    plot_kwargs : dict, optional
+        Any extra kwargs to pass through to the plotting function.  This
+        will be the same for all calls to the plotting function and will
+        over-ride the values in cycle.
+
+    Returns
+    -------
+    arts : dict
+        Dictionary of artists keyed on their labels
+    """
+    # deal with default binning function
+    if hist_func is None:
+        hist_func = np.histogram
+
+    # deal with default plotting function
+    if plot_func is None:
+        plot_func = filled_hist
+
+    # deal with default
+    if plot_kwargs is None:
+        plot_kwargs = {}
+    print(plot_kwargs)
+    try:
+        l_keys = stacked_data.keys()
+        label_data = True
+        if labels is None:
+            labels = l_keys
+
+    except AttributeError:
+        label_data = False
+        if labels is None:
+            labels = itertools.repeat(None)
+
+    if label_data:
+        loop_iter = enumerate((stacked_data[lab], lab, s) for lab, s in
+                              zip(labels, sty_cycle))
+    else:
+        loop_iter = enumerate(zip(stacked_data, labels, sty_cycle))
+
+    arts = {}
+    for j, (data, label, sty) in loop_iter:
+        if label is None:
+            label = 'default set {n}'.format(n=j)
+        label = sty.pop('label', label)
+        vals, edges = hist_func(data)
+        if bottoms is None:
+            bottoms = np.zeros_like(vals)
+        top = bottoms + vals
+        print(sty)
+        sty.update(plot_kwargs)
+        print(sty)
+        ret = plot_func(ax, edges, top, bottoms=bottoms,
+                        label=label, **sty)
+        bottoms = top
+        arts[label] = ret
+    ax.legend()
+    return arts
+
+
+# set up histogram function to fixed bins
+edges = np.linspace(-3, 3, 20, endpoint=True)
+hist_func = partial(np.histogram, bins=edges)
+
+# set up style cycles
+color_cycle = cycler('facecolor', 'rgbm')
+label_cycle = cycler('label', ['set {n}'.format(n=n) for n in range(4)])
+hatch_cycle = cycler('hatch', ['/', '*', '+', '|'])
+
+# make some synthetic data
+stack_data = np.random.randn(4, 12250)
+dict_data = {lab: d for lab, d in zip(list(c['label'] for c in label_cycle),
+                                      stack_data)}
+
+# work with plain arrays
+fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(12, 6), tight_layout=True)
+arts = stack_hist(ax1, stack_data, color_cycle + label_cycle + hatch_cycle,
+                  hist_func=hist_func)
+
+arts = stack_hist(ax2, stack_data, color_cycle,
+                  hist_func=hist_func,
+                  plot_kwargs=dict(edgecolor='w', orientation='h'))
+ax1.set_ylabel('counts')
+ax1.set_xlabel('x')
+ax2.set_xlabel('counts')
+ax2.set_ylabel('x')
+
+# work with labeled data
+
+fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(12, 6),
+                               tight_layout=True, sharey=True)
+
+arts = stack_hist(ax1, dict_data, color_cycle + hatch_cycle,
+                  hist_func=hist_func)
+
+arts = stack_hist(ax2, dict_data, color_cycle + hatch_cycle,
+                  hist_func=hist_func, labels=['set 0', 'set 3'])
+
+ax1.set_ylabel('counts')
+ax1.set_xlabel('x')
+ax2.set_xlabel('x')