test_axes.py

from collections import namedtuple
import datetime
from decimal import Decimal
import io
from itertools import product
import platform
from types import SimpleNamespace
try:
    from contextlib import nullcontext
except ImportError:
    from contextlib import ExitStack as nullcontext  # Py3.6.

import dateutil.tz

import numpy as np
from numpy import ma
from cycler import cycler
import pytest

import matplotlib
import matplotlib as mpl
from matplotlib.testing.decorators import (
    image_comparison, check_figures_equal, remove_ticks_and_titles)
import matplotlib.colors as mcolors
import matplotlib.dates as mdates
from matplotlib.figure import Figure
import matplotlib.font_manager as mfont_manager
import matplotlib.markers as mmarkers
import matplotlib.patches as mpatches
import matplotlib.pyplot as plt
import matplotlib.ticker as mticker
import matplotlib.transforms as mtransforms
from numpy.testing import (
    assert_allclose, assert_array_equal, assert_array_almost_equal)
from matplotlib import rc_context
from matplotlib.cbook import MatplotlibDeprecationWarning

# Note: Some test cases are run twice: once normally and once with labeled data
#       These two must be defined in the same test function or need to have
#       different baseline images to prevent race conditions when pytest runs
#       the tests with multiple threads.


def test_get_labels():
    fig, ax = plt.subplots()
    ax.set_xlabel('x label')
    ax.set_ylabel('y label')
    assert ax.get_xlabel() == 'x label'
    assert ax.get_ylabel() == 'y label'


@check_figures_equal()
def test_label_loc_vertical(fig_test, fig_ref):
    ax = fig_test.subplots()
    sc = ax.scatter([1, 2], [1, 2], c=[1, 2])
    ax.set_ylabel('Y Label', loc='top')
    ax.set_xlabel('X Label', loc='right')
    cbar = fig_test.colorbar(sc)
    cbar.set_label("Z Label", loc='top')

    ax = fig_ref.subplots()
    sc = ax.scatter([1, 2], [1, 2], c=[1, 2])
    ax.set_ylabel('Y Label', y=1, ha='right')
    ax.set_xlabel('X Label', x=1, ha='right')
    cbar = fig_ref.colorbar(sc)
    cbar.set_label("Z Label", y=1, ha='right')


@check_figures_equal()
def test_label_loc_horizontal(fig_test, fig_ref):
    ax = fig_test.subplots()
    sc = ax.scatter([1, 2], [1, 2], c=[1, 2])
    ax.set_ylabel('Y Label', loc='bottom')
    ax.set_xlabel('X Label', loc='left')
    cbar = fig_test.colorbar(sc, orientation='horizontal')
    cbar.set_label("Z Label", loc='left')

    ax = fig_ref.subplots()
    sc = ax.scatter([1, 2], [1, 2], c=[1, 2])
    ax.set_ylabel('Y Label', y=0, ha='left')
    ax.set_xlabel('X Label', x=0, ha='left')
    cbar = fig_ref.colorbar(sc, orientation='horizontal')
    cbar.set_label("Z Label", x=0, ha='left')


@check_figures_equal()
def test_label_loc_rc(fig_test, fig_ref):
    with matplotlib.rc_context({"xaxis.labellocation": "right",
                                "yaxis.labellocation": "top"}):
        ax = fig_test.subplots()
        sc = ax.scatter([1, 2], [1, 2], c=[1, 2])
        ax.set_ylabel('Y Label')
        ax.set_xlabel('X Label')
        cbar = fig_test.colorbar(sc, orientation='horizontal')
        cbar.set_label("Z Label")

    ax = fig_ref.subplots()
    sc = ax.scatter([1, 2], [1, 2], c=[1, 2])
    ax.set_ylabel('Y Label', y=1, ha='right')
    ax.set_xlabel('X Label', x=1, ha='right')
    cbar = fig_ref.colorbar(sc, orientation='horizontal')
    cbar.set_label("Z Label", x=1, ha='right')


@check_figures_equal(extensions=["png"])
def test_acorr(fig_test, fig_ref):
    np.random.seed(19680801)
    Nx = 512
    x = np.random.normal(0, 1, Nx).cumsum()
    maxlags = Nx-1

    ax_test = fig_test.subplots()
    ax_test.acorr(x, maxlags=maxlags)

    ax_ref = fig_ref.subplots()
    # Normalized autocorrelation
    norm_auto_corr = np.correlate(x, x, mode="full")/np.dot(x, x)
    lags = np.arange(-maxlags, maxlags+1)
    norm_auto_corr = norm_auto_corr[Nx-1-maxlags:Nx+maxlags]
    ax_ref.vlines(lags, [0], norm_auto_corr)
    ax_ref.axhline(y=0, xmin=0, xmax=1)


@check_figures_equal(extensions=["png"])
def test_spy(fig_test, fig_ref):
    np.random.seed(19680801)
    a = np.ones(32 * 32)
    a[:16 * 32] = 0
    np.random.shuffle(a)
    a = a.reshape((32, 32))

    axs_test = fig_test.subplots(2)
    axs_test[0].spy(a)
    axs_test[1].spy(a, marker=".", origin="lower")

    axs_ref = fig_ref.subplots(2)
    axs_ref[0].imshow(a, cmap="gray_r", interpolation="nearest")
    axs_ref[0].xaxis.tick_top()
    axs_ref[1].plot(*np.nonzero(a)[::-1], ".", markersize=10)
    axs_ref[1].set(
        aspect=1, xlim=axs_ref[0].get_xlim(), ylim=axs_ref[0].get_ylim()[::-1])
    for ax in axs_ref:
        ax.xaxis.set_ticks_position("both")


def test_spy_invalid_kwargs():
    fig, ax = plt.subplots()
    for unsupported_kw in [{'interpolation': 'nearest'},
                           {'marker': 'o', 'linestyle': 'solid'}]:
        with pytest.raises(TypeError):
            ax.spy(np.eye(3, 3), **unsupported_kw)


@check_figures_equal(extensions=["png"])
def test_matshow(fig_test, fig_ref):
    mpl.style.use("mpl20")
    a = np.random.rand(32, 32)
    fig_test.add_subplot().matshow(a)
    ax_ref = fig_ref.add_subplot()
    ax_ref.imshow(a)
    ax_ref.xaxis.tick_top()
    ax_ref.xaxis.set_ticks_position('both')


@image_comparison(['formatter_ticker_001',
                   'formatter_ticker_002',
                   'formatter_ticker_003',
                   'formatter_ticker_004',
                   'formatter_ticker_005',
                   ])
def test_formatter_ticker():
    import matplotlib.testing.jpl_units as units
    units.register()

    # This should affect the tick size.  (Tests issue #543)
    matplotlib.rcParams['lines.markeredgewidth'] = 30

    # This essentially test to see if user specified labels get overwritten
    # by the auto labeler functionality of the axes.
    xdata = [x*units.sec for x in range(10)]
    ydata1 = [(1.5*y - 0.5)*units.km for y in range(10)]
    ydata2 = [(1.75*y - 1.0)*units.km for y in range(10)]

    ax = plt.figure().subplots()
    ax.set_xlabel("x-label 001")

    ax = plt.figure().subplots()
    ax.set_xlabel("x-label 001")
    ax.plot(xdata, ydata1, color='blue', xunits="sec")

    ax = plt.figure().subplots()
    ax.set_xlabel("x-label 001")
    ax.plot(xdata, ydata1, color='blue', xunits="sec")
    ax.set_xlabel("x-label 003")

    ax = plt.figure().subplots()
    ax.plot(xdata, ydata1, color='blue', xunits="sec")
    ax.plot(xdata, ydata2, color='green', xunits="hour")
    ax.set_xlabel("x-label 004")

    # See SF bug 2846058
    # https://sourceforge.net/tracker/?func=detail&aid=2846058&group_id=80706&atid=560720
    ax = plt.figure().subplots()
    ax.plot(xdata, ydata1, color='blue', xunits="sec")
    ax.plot(xdata, ydata2, color='green', xunits="hour")
    ax.set_xlabel("x-label 005")
    ax.autoscale_view()


def test_funcformatter_auto_formatter():
    def _formfunc(x, pos):
        return ''

    ax = plt.figure().subplots()

    assert ax.xaxis.isDefault_majfmt
    assert ax.xaxis.isDefault_minfmt
    assert ax.yaxis.isDefault_majfmt
    assert ax.yaxis.isDefault_minfmt

    ax.xaxis.set_major_formatter(_formfunc)

    assert not ax.xaxis.isDefault_majfmt
    assert ax.xaxis.isDefault_minfmt
    assert ax.yaxis.isDefault_majfmt
    assert ax.yaxis.isDefault_minfmt

    targ_funcformatter = mticker.FuncFormatter(_formfunc)

    assert isinstance(ax.xaxis.get_major_formatter(),
                      mticker.FuncFormatter)

    assert ax.xaxis.get_major_formatter().func == targ_funcformatter.func


def test_strmethodformatter_auto_formatter():
    formstr = '{x}_{pos}'

    ax = plt.figure().subplots()

    assert ax.xaxis.isDefault_majfmt
    assert ax.xaxis.isDefault_minfmt
    assert ax.yaxis.isDefault_majfmt
    assert ax.yaxis.isDefault_minfmt

    ax.yaxis.set_minor_formatter(formstr)

    assert ax.xaxis.isDefault_majfmt
    assert ax.xaxis.isDefault_minfmt
    assert ax.yaxis.isDefault_majfmt
    assert not ax.yaxis.isDefault_minfmt

    targ_strformatter = mticker.StrMethodFormatter(formstr)

    assert isinstance(ax.yaxis.get_minor_formatter(),
                      mticker.StrMethodFormatter)

    assert ax.yaxis.get_minor_formatter().fmt == targ_strformatter.fmt


@image_comparison(["twin_axis_locators_formatters"])
def test_twin_axis_locators_formatters():
    vals = np.linspace(0, 1, num=5, endpoint=True)
    locs = np.sin(np.pi * vals / 2.0)

    majl = plt.FixedLocator(locs)
    minl = plt.FixedLocator([0.1, 0.2, 0.3])

    fig = plt.figure()
    ax1 = fig.add_subplot(1, 1, 1)
    ax1.plot([0.1, 100], [0, 1])
    ax1.yaxis.set_major_locator(majl)
    ax1.yaxis.set_minor_locator(minl)
    ax1.yaxis.set_major_formatter(plt.FormatStrFormatter('%08.2lf'))
    ax1.yaxis.set_minor_formatter(plt.FixedFormatter(['tricks', 'mind',
                                                      'jedi']))

    ax1.xaxis.set_major_locator(plt.LinearLocator())
    ax1.xaxis.set_minor_locator(plt.FixedLocator([15, 35, 55, 75]))
    ax1.xaxis.set_major_formatter(plt.FormatStrFormatter('%05.2lf'))
    ax1.xaxis.set_minor_formatter(plt.FixedFormatter(['c', '3', 'p', 'o']))
    ax1.twiny()
    ax1.twinx()


def test_twinx_cla():
    fig, ax = plt.subplots()
    ax2 = ax.twinx()
    ax3 = ax2.twiny()
    plt.draw()
    assert not ax2.xaxis.get_visible()
    assert not ax2.patch.get_visible()
    ax2.cla()
    ax3.cla()

    assert not ax2.xaxis.get_visible()
    assert not ax2.patch.get_visible()
    assert ax2.yaxis.get_visible()

    assert ax3.xaxis.get_visible()
    assert not ax3.patch.get_visible()
    assert not ax3.yaxis.get_visible()

    assert ax.xaxis.get_visible()
    assert ax.patch.get_visible()
    assert ax.yaxis.get_visible()


@pytest.mark.parametrize('twin', ('x', 'y'))
@check_figures_equal(extensions=['png'], tol=0.19)
def test_twin_logscale(fig_test, fig_ref, twin):
    twin_func = f'twin{twin}'  # test twinx or twiny
    set_scale = f'set_{twin}scale'
    x = np.arange(1, 100)

    # Change scale after twinning.
    ax_test = fig_test.add_subplot(2, 1, 1)
    ax_twin = getattr(ax_test, twin_func)()
    getattr(ax_test, set_scale)('log')
    ax_twin.plot(x, x)

    # Twin after changing scale.
    ax_test = fig_test.add_subplot(2, 1, 2)
    getattr(ax_test, set_scale)('log')
    ax_twin = getattr(ax_test, twin_func)()
    ax_twin.plot(x, x)

    for i in [1, 2]:
        ax_ref = fig_ref.add_subplot(2, 1, i)
        getattr(ax_ref, set_scale)('log')
        ax_ref.plot(x, x)

        # This is a hack because twinned Axes double-draw the frame.
        # Remove this when that is fixed.
        Path = matplotlib.path.Path
        fig_ref.add_artist(
            matplotlib.patches.PathPatch(
                Path([[0, 0], [0, 1],
                      [0, 1], [1, 1],
                      [1, 1], [1, 0],
                      [1, 0], [0, 0]],
                     [Path.MOVETO, Path.LINETO] * 4),
                transform=ax_ref.transAxes,
                facecolor='none',
                edgecolor=mpl.rcParams['axes.edgecolor'],
                linewidth=mpl.rcParams['axes.linewidth'],
                capstyle='projecting'))

    remove_ticks_and_titles(fig_test)
    remove_ticks_and_titles(fig_ref)


@image_comparison(['twin_autoscale.png'])
def test_twinx_axis_scales():
    x = np.array([0, 0.5, 1])
    y = 0.5 * x
    x2 = np.array([0, 1, 2])
    y2 = 2 * x2

    fig = plt.figure()
    ax = fig.add_axes((0, 0, 1, 1), autoscalex_on=False, autoscaley_on=False)
    ax.plot(x, y, color='blue', lw=10)

    ax2 = plt.twinx(ax)
    ax2.plot(x2, y2, 'r--', lw=5)

    ax.margins(0, 0)
    ax2.margins(0, 0)


def test_twin_inherit_autoscale_setting():
    fig, ax = plt.subplots()
    ax_x_on = ax.twinx()
    ax.set_autoscalex_on(False)
    ax_x_off = ax.twinx()

    assert ax_x_on.get_autoscalex_on()
    assert not ax_x_off.get_autoscalex_on()

    ax_y_on = ax.twiny()
    ax.set_autoscaley_on(False)
    ax_y_off = ax.twiny()

    assert ax_y_on.get_autoscaley_on()
    assert not ax_y_off.get_autoscaley_on()


def test_inverted_cla():
    # GitHub PR #5450. Setting autoscale should reset
    # axes to be non-inverted.
    # plotting an image, then 1d graph, axis is now down
    fig = plt.figure(0)
    ax = fig.gca()
    # 1. test that a new axis is not inverted per default
    assert not ax.xaxis_inverted()
    assert not ax.yaxis_inverted()
    img = np.random.random((100, 100))
    ax.imshow(img)
    # 2. test that a image axis is inverted
    assert not ax.xaxis_inverted()
    assert ax.yaxis_inverted()
    # 3. test that clearing and plotting a line, axes are
    # not inverted
    ax.cla()
    x = np.linspace(0, 2*np.pi, 100)
    ax.plot(x, np.cos(x))
    assert not ax.xaxis_inverted()
    assert not ax.yaxis_inverted()

    # 4. autoscaling should not bring back axes to normal
    ax.cla()
    ax.imshow(img)
    plt.autoscale()
    assert not ax.xaxis_inverted()
    assert ax.yaxis_inverted()

    # 5. two shared axes. Inverting the master axis should invert the shared
    # axes; clearing the master axis should bring axes in shared
    # axes back to normal.
    ax0 = plt.subplot(211)
    ax1 = plt.subplot(212, sharey=ax0)
    ax0.yaxis.set_inverted(True)
    assert ax1.yaxis_inverted()
    ax1.plot(x, np.cos(x))
    ax0.cla()
    assert not ax1.yaxis_inverted()
    ax1.cla()
    # 6. clearing the nonmaster should not touch limits
    ax0.imshow(img)
    ax1.plot(x, np.cos(x))
    ax1.cla()
    assert ax.yaxis_inverted()

    # clean up
    plt.close(fig)


@check_figures_equal(extensions=["png"])
def test_minorticks_on_rcParams_both(fig_test, fig_ref):
    with matplotlib.rc_context({"xtick.minor.visible": True,
                                "ytick.minor.visible": True}):
        ax_test = fig_test.subplots()
        ax_test.plot([0, 1], [0, 1])
    ax_ref = fig_ref.subplots()
    ax_ref.plot([0, 1], [0, 1])
    ax_ref.minorticks_on()


@image_comparison(["autoscale_tiny_range"], remove_text=True)
def test_autoscale_tiny_range():
    # github pull #904
    fig, axs = plt.subplots(2, 2)
    for i, ax in enumerate(axs.flat):
        y1 = 10**(-11 - i)
        ax.plot([0, 1], [1, 1 + y1])


@pytest.mark.style('default')
def test_autoscale_tight():
    fig, ax = plt.subplots(1, 1)
    ax.plot([1, 2, 3, 4])
    ax.autoscale(enable=True, axis='x', tight=False)
    ax.autoscale(enable=True, axis='y', tight=True)
    assert_allclose(ax.get_xlim(), (-0.15, 3.15))
    assert_allclose(ax.get_ylim(), (1.0, 4.0))


@pytest.mark.style('default')
def test_autoscale_log_shared():
    # related to github #7587
    # array starts at zero to trigger _minpos handling
    x = np.arange(100, dtype=float)
    fig, (ax1, ax2) = plt.subplots(2, 1, sharex=True)
    ax1.loglog(x, x)
    ax2.semilogx(x, x)
    ax1.autoscale(tight=True)
    ax2.autoscale(tight=True)
    plt.draw()
    lims = (x[1], x[-1])
    assert_allclose(ax1.get_xlim(), lims)
    assert_allclose(ax1.get_ylim(), lims)
    assert_allclose(ax2.get_xlim(), lims)
    assert_allclose(ax2.get_ylim(), (x[0], x[-1]))


@pytest.mark.style('default')
def test_use_sticky_edges():
    fig, ax = plt.subplots()
    ax.imshow([[0, 1], [2, 3]], origin='lower')
    assert_allclose(ax.get_xlim(), (-0.5, 1.5))
    assert_allclose(ax.get_ylim(), (-0.5, 1.5))
    ax.use_sticky_edges = False
    ax.autoscale()
    xlim = (-0.5 - 2 * ax._xmargin, 1.5 + 2 * ax._xmargin)
    ylim = (-0.5 - 2 * ax._ymargin, 1.5 + 2 * ax._ymargin)
    assert_allclose(ax.get_xlim(), xlim)
    assert_allclose(ax.get_ylim(), ylim)
    # Make sure it is reversible:
    ax.use_sticky_edges = True
    ax.autoscale()
    assert_allclose(ax.get_xlim(), (-0.5, 1.5))
    assert_allclose(ax.get_ylim(), (-0.5, 1.5))


@check_figures_equal(extensions=["png"])
def test_sticky_shared_axes(fig_test, fig_ref):
    # Check that sticky edges work whether they are set in an axes that is a
    # "master" in a share, or an axes that is a "follower".
    Z = np.arange(15).reshape(3, 5)

    ax0 = fig_test.add_subplot(211)
    ax1 = fig_test.add_subplot(212, sharex=ax0)
    ax1.pcolormesh(Z)

    ax0 = fig_ref.add_subplot(212)
    ax1 = fig_ref.add_subplot(211, sharex=ax0)
    ax0.pcolormesh(Z)


@image_comparison(['offset_points'], remove_text=True)
def test_basic_annotate():
    # Setup some data
    t = np.arange(0.0, 5.0, 0.01)
    s = np.cos(2.0*np.pi * t)

    # Offset Points

    fig = plt.figure()
    ax = fig.add_subplot(autoscale_on=False, xlim=(-1, 5), ylim=(-3, 5))
    line, = ax.plot(t, s, lw=3, color='purple')

    ax.annotate('local max', xy=(3, 1), xycoords='data',
                xytext=(3, 3), textcoords='offset points')


def test_annotate_parameter_warn():
    fig, ax = plt.subplots()
    with pytest.warns(MatplotlibDeprecationWarning,
                      match=r"The \'s\' parameter of annotate\(\) "
                             "has been renamed \'text\'"):
        ax.annotate(s='now named text', xy=(0, 1))


@image_comparison(['arrow_simple.png'], remove_text=True)
def test_arrow_simple():
    # Simple image test for ax.arrow
    # kwargs that take discrete values
    length_includes_head = (True, False)
    shape = ('full', 'left', 'right')
    head_starts_at_zero = (True, False)
    # Create outer product of values
    kwargs = product(length_includes_head, shape, head_starts_at_zero)

    fig, axs = plt.subplots(3, 4)
    for i, (ax, kwarg) in enumerate(zip(axs.flat, kwargs)):
        ax.set_xlim(-2, 2)
        ax.set_ylim(-2, 2)
        # Unpack kwargs
        (length_includes_head, shape, head_starts_at_zero) = kwarg
        theta = 2 * np.pi * i / 12
        # Draw arrow
        ax.arrow(0, 0, np.sin(theta), np.cos(theta),
                 width=theta/100,
                 length_includes_head=length_includes_head,
                 shape=shape,
                 head_starts_at_zero=head_starts_at_zero,
                 head_width=theta / 10,
                 head_length=theta / 10)


def test_arrow_empty():
    _, ax = plt.subplots()
    # Create an empty FancyArrow
    ax.arrow(0, 0, 0, 0, head_length=0)


def test_arrow_in_view():
    _, ax = plt.subplots()
    ax.arrow(1, 1, 1, 1)
    assert ax.get_xlim() == (0.8, 2.2)
    assert ax.get_ylim() == (0.8, 2.2)


def test_annotate_default_arrow():
    # Check that we can make an annotation arrow with only default properties.
    fig, ax = plt.subplots()
    ann = ax.annotate("foo", (0, 1), xytext=(2, 3))
    assert ann.arrow_patch is None
    ann = ax.annotate("foo", (0, 1), xytext=(2, 3), arrowprops={})
    assert ann.arrow_patch is not None


@image_comparison(['fill_units.png'], savefig_kwarg={'dpi': 60})
def test_fill_units():
    import matplotlib.testing.jpl_units as units
    units.register()

    # generate some data
    t = units.Epoch("ET", dt=datetime.datetime(2009, 4, 27))
    value = 10.0 * units.deg
    day = units.Duration("ET", 24.0 * 60.0 * 60.0)
    dt = np.arange('2009-04-27', '2009-04-29', dtype='datetime64[D]')
    dtn = mdates.date2num(dt)

    fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2)

    ax1.plot([t], [value], yunits='deg', color='red')
    ind = [0, 0, 1, 1]
    ax1.fill(dtn[ind], [0.0, 0.0, 90.0, 0.0], 'b')

    ax2.plot([t], [value], yunits='deg', color='red')
    ax2.fill([t, t, t + day, t + day],
             [0.0, 0.0, 90.0, 0.0], 'b')

    ax3.plot([t], [value], yunits='deg', color='red')
    ax3.fill(dtn[ind],
             [0 * units.deg, 0 * units.deg, 90 * units.deg, 0 * units.deg],
             'b')

    ax4.plot([t], [value], yunits='deg', color='red')
    ax4.fill([t, t, t + day, t + day],
             [0 * units.deg, 0 * units.deg, 90 * units.deg, 0 * units.deg],
             facecolor="blue")
    fig.autofmt_xdate()


def test_plot_format_kwarg_redundant():
    with pytest.warns(UserWarning, match="marker .* redundantly defined"):
        plt.plot([0], [0], 'o', marker='x')
    with pytest.warns(UserWarning, match="linestyle .* redundantly defined"):
        plt.plot([0], [0], '-', linestyle='--')
    with pytest.warns(UserWarning, match="color .* redundantly defined"):
        plt.plot([0], [0], 'r', color='blue')
    # smoke-test: should not warn
    plt.errorbar([0], [0], fmt='none', color='blue')


@image_comparison(['single_point', 'single_point'])
def test_single_point():
    # Issue #1796: don't let lines.marker affect the grid
    matplotlib.rcParams['lines.marker'] = 'o'
    matplotlib.rcParams['axes.grid'] = True

    fig, (ax1, ax2) = plt.subplots(2)
    ax1.plot([0], [0], 'o')
    ax2.plot([1], [1], 'o')

    # Reuse testcase from above for a labeled data test
    data = {'a': [0], 'b': [1]}

    fig, (ax1, ax2) = plt.subplots(2)
    ax1.plot('a', 'a', 'o', data=data)
    ax2.plot('b', 'b', 'o', data=data)


@image_comparison(['single_date.png'], style='mpl20')
def test_single_date():

    # use former defaults to match existing baseline image
    plt.rcParams['axes.formatter.limits'] = -7, 7
    dt = mdates.date2num(np.datetime64('0000-12-31'))

    time1 = [721964.0]
    data1 = [-65.54]

    fig, ax = plt.subplots(2, 1)
    ax[0].plot_date(time1 + dt, data1, 'o', color='r')
    ax[1].plot(time1, data1, 'o', color='r')


@check_figures_equal(extensions=["png"])
def test_shaped_data(fig_test, fig_ref):
    row = np.arange(10).reshape((1, -1))
    col = np.arange(0, 100, 10).reshape((-1, 1))

    axs = fig_test.subplots(2)
    axs[0].plot(row)  # Actually plots nothing (columns are single points).
    axs[1].plot(col)  # Same as plotting 1d.

    axs = fig_ref.subplots(2)
    # xlim from the implicit "x=0", ylim from the row datalim.
    axs[0].set(xlim=(-.06, .06), ylim=(0, 9))
    axs[1].plot(col.ravel())


def test_structured_data():
    # support for structured data
    pts = np.array([(1, 1), (2, 2)], dtype=[("ones", float), ("twos", float)])

    # this should not read second name as a format and raise ValueError
    axs = plt.figure().subplots(2)
    axs[0].plot("ones", "twos", data=pts)
    axs[1].plot("ones", "twos", "r", data=pts)


@image_comparison(['aitoff_proj'], extensions=["png"],
                  remove_text=True, style='mpl20')
def test_aitoff_proj():
    """
    Test aitoff projection ref.:
    https://github.com/matplotlib/matplotlib/pull/14451
    """
    x = np.linspace(-np.pi, np.pi, 20)
    y = np.linspace(-np.pi / 2, np.pi / 2, 20)
    X, Y = np.meshgrid(x, y)

    fig, ax = plt.subplots(figsize=(8, 4.2),
                           subplot_kw=dict(projection="aitoff"))
    ax.grid()
    ax.plot(X.flat, Y.flat, 'o', markersize=4)


@image_comparison(['axvspan_epoch'])
def test_axvspan_epoch():
    import matplotlib.testing.jpl_units as units
    units.register()

    # generate some data
    t0 = units.Epoch("ET", dt=datetime.datetime(2009, 1, 20))
    tf = units.Epoch("ET", dt=datetime.datetime(2009, 1, 21))
    dt = units.Duration("ET", units.day.convert("sec"))

    ax = plt.gca()
    ax.axvspan(t0, tf, facecolor="blue", alpha=0.25)
    ax.set_xlim(t0 - 5.0*dt, tf + 5.0*dt)


@image_comparison(['axhspan_epoch'], tol=0.02)
def test_axhspan_epoch():
    import matplotlib.testing.jpl_units as units
    units.register()

    # generate some data
    t0 = units.Epoch("ET", dt=datetime.datetime(2009, 1, 20))
    tf = units.Epoch("ET", dt=datetime.datetime(2009, 1, 21))
    dt = units.Duration("ET", units.day.convert("sec"))

    ax = plt.gca()
    ax.axhspan(t0, tf, facecolor="blue", alpha=0.25)
    ax.set_ylim(t0 - 5.0*dt, tf + 5.0*dt)


@image_comparison(['hexbin_extent.png', 'hexbin_extent.png'], remove_text=True)
def test_hexbin_extent():
    # this test exposes sf bug 2856228
    fig, ax = plt.subplots()
    data = (np.arange(2000) / 2000).reshape((2, 1000))
    x, y = data

    ax.hexbin(x, y, extent=[.1, .3, .6, .7])

    # Reuse testcase from above for a labeled data test
    data = {"x": x, "y": y}

    fig, ax = plt.subplots()
    ax.hexbin("x", "y", extent=[.1, .3, .6, .7], data=data)


@image_comparison(['hexbin_empty.png'], remove_text=True)
def test_hexbin_empty():
    # From #3886: creating hexbin from empty dataset raises ValueError
    ax = plt.gca()
    ax.hexbin([], [])


def test_hexbin_pickable():
    # From #1973: Test that picking a hexbin collection works
    fig, ax = plt.subplots()
    data = (np.arange(200) / 200).reshape((2, 100))
    x, y = data
    hb = ax.hexbin(x, y, extent=[.1, .3, .6, .7], picker=-1)
    mouse_event = SimpleNamespace(x=400, y=300)
    assert hb.contains(mouse_event)[0]


@image_comparison(['hexbin_log.png'], style='mpl20')
def test_hexbin_log():
    # Issue #1636 (and also test log scaled colorbar)

    # Remove this line when this test image is regenerated.
    plt.rcParams['pcolormesh.snap'] = False

    np.random.seed(19680801)
    n = 100000
    x = np.random.standard_normal(n)
    y = 2.0 + 3.0 * x + 4.0 * np.random.standard_normal(n)
    y = np.power(2, y * 0.5)

    fig, ax = plt.subplots()
    h = ax.hexbin(x, y, yscale='log', bins='log')
    plt.colorbar(h)


def test_inverted_limits():
    # Test gh:1553
    # Calling invert_xaxis prior to plotting should not disable autoscaling
    # while still maintaining the inverted direction
    fig, ax = plt.subplots()
    ax.invert_xaxis()
    ax.plot([-5, -3, 2, 4], [1, 2, -3, 5])

    assert ax.get_xlim() == (4, -5)
    assert ax.get_ylim() == (-3, 5)
    plt.close()

    fig, ax = plt.subplots()
    ax.invert_yaxis()
    ax.plot([-5, -3, 2, 4], [1, 2, -3, 5])

    assert ax.get_xlim() == (-5, 4)
    assert ax.get_ylim() == (5, -3)

    # Test inverting nonlinear axes.
    fig, ax = plt.subplots()
    ax.set_yscale("log")
    ax.set_ylim(10, 1)
    assert ax.get_ylim() == (10, 1)


@image_comparison(['nonfinite_limits'])
def test_nonfinite_limits():
    x = np.arange(0., np.e, 0.01)
    # silence divide by zero warning from log(0)
    with np.errstate(divide='ignore'):
        y = np.log(x)
    x[len(x)//2] = np.nan
    fig, ax = plt.subplots()
    ax.plot(x, y)


@pytest.mark.style('default')
@pytest.mark.parametrize('plot_fun',
                         ['scatter', 'plot', 'fill_between'])
@check_figures_equal(extensions=["png"])
def test_limits_empty_data(plot_fun, fig_test, fig_ref):
    # Check that plotting empty data doesn't change autoscaling of dates
    x = np.arange("2010-01-01", "2011-01-01", dtype="datetime64[D]")

    ax_test = fig_test.subplots()
    ax_ref = fig_ref.subplots()

    getattr(ax_test, plot_fun)([], [])

    for ax in [ax_test, ax_ref]:
        getattr(ax, plot_fun)(x, range(len(x)), color='C0')


@image_comparison(['imshow', 'imshow'], remove_text=True, style='mpl20')
def test_imshow():
    # use former defaults to match existing baseline image
    matplotlib.rcParams['image.interpolation'] = 'nearest'
    # Create a NxN image
    N = 100
    (x, y) = np.indices((N, N))
    x -= N//2
    y -= N//2
    r = np.sqrt(x**2+y**2-x*y)

    # Create a contour plot at N/4 and extract both the clip path and transform
    fig, ax = plt.subplots()
    ax.imshow(r)

    # Reuse testcase from above for a labeled data test
    data = {"r": r}
    fig, ax = plt.subplots()
    ax.imshow("r", data=data)


@image_comparison(['imshow_clip'], style='mpl20')
def test_imshow_clip():
    # As originally reported by Gellule Xg <gellule.xg@free.fr>
    # use former defaults to match existing baseline image
    matplotlib.rcParams['image.interpolation'] = 'nearest'

    # Create a NxN image
    N = 100
    (x, y) = np.indices((N, N))
    x -= N//2
    y -= N//2
    r = np.sqrt(x**2+y**2-x*y)

    # Create a contour plot at N/4 and extract both the clip path and transform
    fig, ax = plt.subplots()

    c = ax.contour(r, [N/4])
    x = c.collections[0]
    clip_path = x.get_paths()[0]
    clip_transform = x.get_transform()

    clip_path = mtransforms.TransformedPath(clip_path, clip_transform)

    # Plot the image clipped by the contour
    ax.imshow(r, clip_path=clip_path)


@check_figures_equal(extensions=["png"])
def test_imshow_norm_vminvmax(fig_test, fig_ref):
    """Parameters vmin, vmax should be ignored if norm is given."""
    a = [[1, 2], [3, 4]]
    ax = fig_ref.subplots()
    ax.imshow(a, vmin=0, vmax=5)
    ax = fig_test.subplots()
    with pytest.warns(MatplotlibDeprecationWarning,
                      match="Passing parameters norm and vmin/vmax "
                            "simultaneously is deprecated."):
        ax.imshow(a, norm=mcolors.Normalize(-10, 10), vmin=0, vmax=5)


@image_comparison(['polycollection_joinstyle'], remove_text=True)
def test_polycollection_joinstyle():
    # Bug #2890979 reported by Matthew West
    fig, ax = plt.subplots()
    verts = np.array([[1, 1], [1, 2], [2, 2], [2, 1]])
    c = mpl.collections.PolyCollection([verts], linewidths=40)
    ax.add_collection(c)
    ax.set_xbound(0, 3)
    ax.set_ybound(0, 3)


@pytest.mark.parametrize(
    'x, y1, y2', [
        (np.zeros((2, 2)), 3, 3),
        (np.arange(0.0, 2, 0.02), np.zeros((2, 2)), 3),
        (np.arange(0.0, 2, 0.02), 3, np.zeros((2, 2)))
    ], ids=[
        '2d_x_input',
        '2d_y1_input',
        '2d_y2_input'
    ]
)
def test_fill_between_input(x, y1, y2):
    fig, ax = plt.subplots()
    with pytest.raises(ValueError):
        ax.fill_between(x, y1, y2)


@pytest.mark.parametrize(
    'y, x1, x2', [
        (np.zeros((2, 2)), 3, 3),
        (np.arange(0.0, 2, 0.02), np.zeros((2, 2)), 3),
        (np.arange(0.0, 2, 0.02), 3, np.zeros((2, 2)))
    ], ids=[
        '2d_y_input',
        '2d_x1_input',
        '2d_x2_input'
    ]
)
def test_fill_betweenx_input(y, x1, x2):
    fig, ax = plt.subplots()
    with pytest.raises(ValueError):
        ax.fill_betweenx(y, x1, x2)


@image_comparison(['fill_between_interpolate'], remove_text=True)
def test_fill_between_interpolate():
    x = np.arange(0.0, 2, 0.02)
    y1 = np.sin(2*np.pi*x)
    y2 = 1.2*np.sin(4*np.pi*x)

    fig, (ax1, ax2) = plt.subplots(2, 1, sharex=True)
    ax1.plot(x, y1, x, y2, color='black')
    ax1.fill_between(x, y1, y2, where=y2 >= y1, facecolor='white', hatch='/',
                     interpolate=True)
    ax1.fill_between(x, y1, y2, where=y2 <= y1, facecolor='red',
                     interpolate=True)

    # Test support for masked arrays.
    y2 = np.ma.masked_greater(y2, 1.0)
    # Test that plotting works for masked arrays with the first element masked
    y2[0] = np.ma.masked
    ax2.plot(x, y1, x, y2, color='black')
    ax2.fill_between(x, y1, y2, where=y2 >= y1, facecolor='green',
                     interpolate=True)
    ax2.fill_between(x, y1, y2, where=y2 <= y1, facecolor='red',
                     interpolate=True)


@image_comparison(['fill_between_interpolate_decreasing'],
                  style='mpl20', remove_text=True)
def test_fill_between_interpolate_decreasing():
    p = np.array([724.3, 700, 655])
    t = np.array([9.4, 7, 2.2])
    prof = np.array([7.9, 6.6, 3.8])

    fig, ax = plt.subplots(figsize=(9, 9))

    ax.plot(t, p, 'tab:red')
    ax.plot(prof, p, 'k')

    ax.fill_betweenx(p, t, prof, where=prof < t,
                     facecolor='blue', interpolate=True, alpha=0.4)
    ax.fill_betweenx(p, t, prof, where=prof > t,
                     facecolor='red', interpolate=True, alpha=0.4)

    ax.set_xlim(0, 30)
    ax.set_ylim(800, 600)


# test_symlog and test_symlog2 used to have baseline images in all three
# formats, but the png and svg baselines got invalidated by the removal of
# minor tick overstriking.
@image_comparison(['symlog.pdf'])
def test_symlog():
    x = np.array([0, 1, 2, 4, 6, 9, 12, 24])
    y = np.array([1000000, 500000, 100000, 100, 5, 0, 0, 0])

    fig, ax = plt.subplots()
    ax.plot(x, y)
    ax.set_yscale('symlog')
    ax.set_xscale('linear')
    ax.set_ylim(-1, 10000000)


@image_comparison(['symlog2.pdf'], remove_text=True)
def test_symlog2():
    # Numbers from -50 to 50, with 0.1 as step
    x = np.arange(-50, 50, 0.001)

    fig, axs = plt.subplots(5, 1)
    for ax, linthresh in zip(axs, [20., 2., 1., 0.1, 0.01]):
        ax.plot(x, x)
        ax.set_xscale('symlog', linthresh=linthresh)
        ax.grid(True)
    axs[-1].set_ylim(-0.1, 0.1)


def test_pcolorargs_5205():
    # Smoketest to catch issue found in gh:5205
    x = [-1.5, -1.0, -0.5, 0.0, 0.5, 1.0, 1.5]
    y = [-1.5, -1.25, -1.0, -0.75, -0.5, -0.25, 0,
         0.25, 0.5, 0.75, 1.0, 1.25, 1.5]
    X, Y = np.meshgrid(x, y)
    Z = np.hypot(X, Y)

    plt.pcolor(Z)
    plt.pcolor(list(Z))
    plt.pcolor(x, y, Z[:-1, :-1])
    plt.pcolor(X, Y, list(Z[:-1, :-1]))


@image_comparison(['pcolormesh'], remove_text=True)
def test_pcolormesh():
    # Remove this line when this test image is regenerated.
    plt.rcParams['pcolormesh.snap'] = False

    n = 12
    x = np.linspace(-1.5, 1.5, n)
    y = np.linspace(-1.5, 1.5, n*2)
    X, Y = np.meshgrid(x, y)
    Qx = np.cos(Y) - np.cos(X)
    Qz = np.sin(Y) + np.sin(X)
    Qx = (Qx + 1.1)
    Z = np.hypot(X, Y) / 5
    Z = (Z - Z.min()) / Z.ptp()

    # The color array can include masked values:
    Zm = ma.masked_where(np.abs(Qz) < 0.5 * np.max(Qz), Z)

    fig, (ax1, ax2, ax3) = plt.subplots(1, 3)
    ax1.pcolormesh(Qx, Qz, Z[:-1, :-1], lw=0.5, edgecolors='k')
    ax2.pcolormesh(Qx, Qz, Z[:-1, :-1], lw=2, edgecolors=['b', 'w'])
    ax3.pcolormesh(Qx, Qz, Z, shading="gouraud")


@image_comparison(['pcolormesh_alpha'], extensions=["png", "pdf"],
                  remove_text=True)
def test_pcolormesh_alpha():
    # Remove this line when this test image is regenerated.
    plt.rcParams['pcolormesh.snap'] = False

    n = 12
    X, Y = np.meshgrid(
        np.linspace(-1.5, 1.5, n),
        np.linspace(-1.5, 1.5, n*2)
    )
    Qx = X
    Qy = Y + np.sin(X)
    Z = np.hypot(X, Y) / 5
    Z = (Z - Z.min()) / Z.ptp()
    vir = plt.get_cmap("viridis", 16)
    # make another colormap with varying alpha
    colors = vir(np.arange(16))
    colors[:, 3] = 0.5 + 0.5*np.sin(np.arange(16))
    cmap = mcolors.ListedColormap(colors)

    fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2)
    for ax in ax1, ax2, ax3, ax4:
        ax.add_patch(mpatches.Rectangle(
            (0, -1.5), 1.5, 3, facecolor=[.7, .1, .1, .5], zorder=0
        ))
    # ax1, ax2: constant alpha
    ax1.pcolormesh(Qx, Qy, Z[:-1, :-1], cmap=vir, alpha=0.4,
                   shading='flat', zorder=1)
    ax2.pcolormesh(Qx, Qy, Z, cmap=vir, alpha=0.4, shading='gouraud', zorder=1)
    # ax3, ax4: alpha from colormap
    ax3.pcolormesh(Qx, Qy, Z[:-1, :-1], cmap=cmap, shading='flat', zorder=1)
    ax4.pcolormesh(Qx, Qy, Z, cmap=cmap, shading='gouraud', zorder=1)


@image_comparison(['pcolormesh_datetime_axis.png'],
                  remove_text=False, style='mpl20')
def test_pcolormesh_datetime_axis():
    # Remove this line when this test image is regenerated.
    plt.rcParams['pcolormesh.snap'] = False

    fig = plt.figure()
    fig.subplots_adjust(hspace=0.4, top=0.98, bottom=.15)
    base = datetime.datetime(2013, 1, 1)
    x = np.array([base + datetime.timedelta(days=d) for d in range(21)])
    y = np.arange(21)
    z1, z2 = np.meshgrid(np.arange(20), np.arange(20))
    z = z1 * z2
    plt.subplot(221)
    plt.pcolormesh(x[:-1], y[:-1], z[:-1, :-1])
    plt.subplot(222)
    plt.pcolormesh(x, y, z)
    x = np.repeat(x[np.newaxis], 21, axis=0)
    y = np.repeat(y[:, np.newaxis], 21, axis=1)
    plt.subplot(223)
    plt.pcolormesh(x[:-1, :-1], y[:-1, :-1], z[:-1, :-1])
    plt.subplot(224)
    plt.pcolormesh(x, y, z)
    for ax in fig.get_axes():
        for label in ax.get_xticklabels():
            label.set_ha('right')
            label.set_rotation(30)


@image_comparison(['pcolor_datetime_axis.png'],
                  remove_text=False, style='mpl20')
def test_pcolor_datetime_axis():
    fig = plt.figure()
    fig.subplots_adjust(hspace=0.4, top=0.98, bottom=.15)
    base = datetime.datetime(2013, 1, 1)
    x = np.array([base + datetime.timedelta(days=d) for d in range(21)])
    y = np.arange(21)
    z1, z2 = np.meshgrid(np.arange(20), np.arange(20))
    z = z1 * z2
    plt.subplot(221)
    plt.pcolor(x[:-1], y[:-1], z[:-1, :-1])
    plt.subplot(222)
    plt.pcolor(x, y, z)
    x = np.repeat(x[np.newaxis], 21, axis=0)
    y = np.repeat(y[:, np.newaxis], 21, axis=1)
    plt.subplot(223)
    plt.pcolor(x[:-1, :-1], y[:-1, :-1], z[:-1, :-1])
    plt.subplot(224)
    plt.pcolor(x, y, z)
    for ax in fig.get_axes():
        for label in ax.get_xticklabels():
            label.set_ha('right')
            label.set_rotation(30)


def test_pcolorargs():
    n = 12
    x = np.linspace(-1.5, 1.5, n)
    y = np.linspace(-1.5, 1.5, n*2)
    X, Y = np.meshgrid(x, y)
    Z = np.hypot(X, Y) / 5

    _, ax = plt.subplots()
    with pytest.raises(TypeError):
        ax.pcolormesh(y, x, Z)
    with pytest.raises(TypeError):
        ax.pcolormesh(X, Y, Z.T)
    with pytest.raises(TypeError):
        ax.pcolormesh(x, y, Z[:-1, :-1], shading="gouraud")
    with pytest.raises(TypeError):
        ax.pcolormesh(X, Y, Z[:-1, :-1], shading="gouraud")
    x[0] = np.NaN
    with pytest.raises(ValueError):
        ax.pcolormesh(x, y, Z[:-1, :-1])
    with np.errstate(invalid='ignore'):
        x = np.ma.array(x, mask=(x < 0))
    with pytest.raises(ValueError):
        ax.pcolormesh(x, y, Z[:-1, :-1])
    # Expect a warning with non-increasing coordinates
    x = [359, 0, 1]
    y = [-10, 10]
    X, Y = np.meshgrid(x, y)
    Z = np.zeros(X.shape)
    with pytest.warns(UserWarning,
                      match='are not monotonically increasing or decreasing'):
        ax.pcolormesh(X, Y, Z, shading='auto')


@check_figures_equal(extensions=["png"])
def test_pcolornearest(fig_test, fig_ref):
    ax = fig_test.subplots()
    x = np.arange(0, 10)
    y = np.arange(0, 3)
    np.random.seed(19680801)
    Z = np.random.randn(2, 9)
    ax.pcolormesh(x, y, Z, shading='flat')

    ax = fig_ref.subplots()
    # specify the centers
    x2 = x[:-1] + np.diff(x) / 2
    y2 = y[:-1] + np.diff(y) / 2
    ax.pcolormesh(x2, y2, Z, shading='nearest')


@check_figures_equal(extensions=["png"])
def test_pcolornearestunits(fig_test, fig_ref):
    ax = fig_test.subplots()
    x = [datetime.datetime.fromtimestamp(x * 3600) for x in range(10)]
    y = np.arange(0, 3)
    np.random.seed(19680801)
    Z = np.random.randn(2, 9)
    ax.pcolormesh(x, y, Z, shading='flat')

    ax = fig_ref.subplots()
    # specify the centers
    x2 = [datetime.datetime.fromtimestamp((x + 0.5) * 3600) for x in range(9)]
    y2 = y[:-1] + np.diff(y) / 2
    ax.pcolormesh(x2, y2, Z, shading='nearest')


@check_figures_equal(extensions=["png"])
def test_pcolordropdata(fig_test, fig_ref):
    ax = fig_test.subplots()
    x = np.arange(0, 10)
    y = np.arange(0, 4)
    np.random.seed(19680801)
    Z = np.random.randn(3, 9)
    # fake dropping the data
    ax.pcolormesh(x[:-1], y[:-1], Z[:-1, :-1], shading='flat')

    ax = fig_ref.subplots()
    # test dropping the data...
    x2 = x[:-1]
    y2 = y[:-1]
    with pytest.warns(MatplotlibDeprecationWarning):
        ax.pcolormesh(x2, y2, Z, shading='flat')


@check_figures_equal(extensions=["png"])
def test_pcolorauto(fig_test, fig_ref):
    ax = fig_test.subplots()
    x = np.arange(0, 10)
    y = np.arange(0, 4)
    np.random.seed(19680801)
    Z = np.random.randn(3, 9)
    ax.pcolormesh(x, y, Z, shading='auto')

    ax = fig_ref.subplots()
    # specify the centers
    x2 = x[:-1] + np.diff(x) / 2
    y2 = y[:-1] + np.diff(y) / 2
    ax.pcolormesh(x2, y2, Z, shading='auto')


@image_comparison(['canonical'])
def test_canonical():
    fig, ax = plt.subplots()
    ax.plot([1, 2, 3])


@image_comparison(['arc_angles.png'], remove_text=True, style='default')
def test_arc_angles():
    # Ellipse parameters
    w = 2
    h = 1
    centre = (0.2, 0.5)
    scale = 2

    fig, axs = plt.subplots(3, 3)
    for i, ax in enumerate(axs.flat):
        theta2 = i * 360 / 9
        theta1 = theta2 - 45

        ax.add_patch(mpatches.Ellipse(centre, w, h, alpha=0.3))
        ax.add_patch(mpatches.Arc(centre, w, h, theta1=theta1, theta2=theta2))
        # Straight lines intersecting start and end of arc
        ax.plot([scale * np.cos(np.deg2rad(theta1)) + centre[0],
                 centre[0],
                 scale * np.cos(np.deg2rad(theta2)) + centre[0]],
                [scale * np.sin(np.deg2rad(theta1)) + centre[1],
                 centre[1],
                 scale * np.sin(np.deg2rad(theta2)) + centre[1]])

        ax.set_xlim(-scale, scale)
        ax.set_ylim(-scale, scale)

        # This looks the same, but it triggers a different code path when it
        # gets large enough.
        w *= 10
        h *= 10
        centre = (centre[0] * 10, centre[1] * 10)
        scale *= 10


@image_comparison(['arc_ellipse'], remove_text=True)
def test_arc_ellipse():
    xcenter, ycenter = 0.38, 0.52
    width, height = 1e-1, 3e-1
    angle = -30

    theta = np.deg2rad(np.arange(360))
    x = width / 2. * np.cos(theta)
    y = height / 2. * np.sin(theta)

    rtheta = np.deg2rad(angle)
    R = np.array([
        [np.cos(rtheta), -np.sin(rtheta)],
        [np.sin(rtheta), np.cos(rtheta)]])

    x, y = np.dot(R, np.array([x, y]))
    x += xcenter
    y += ycenter

    fig = plt.figure()
    ax = fig.add_subplot(211, aspect='auto')
    ax.fill(x, y, alpha=0.2, facecolor='yellow', edgecolor='yellow',
            linewidth=1, zorder=1)

    e1 = mpatches.Arc((xcenter, ycenter), width, height,
                      angle=angle, linewidth=2, fill=False, zorder=2)

    ax.add_patch(e1)

    ax = fig.add_subplot(212, aspect='equal')
    ax.fill(x, y, alpha=0.2, facecolor='green', edgecolor='green', zorder=1)
    e2 = mpatches.Arc((xcenter, ycenter), width, height,
                      angle=angle, linewidth=2, fill=False, zorder=2)

    ax.add_patch(e2)


def test_marker_as_markerstyle():
    fix, ax = plt.subplots()
    m = mmarkers.MarkerStyle('o')
    ax.plot([1, 2, 3], [3, 2, 1], marker=m)
    ax.scatter([1, 2, 3], [4, 3, 2], marker=m)
    ax.errorbar([1, 2, 3], [5, 4, 3], marker=m)


@image_comparison(['markevery'], remove_text=True)
def test_markevery():
    x = np.linspace(0, 10, 100)
    y = np.sin(x) * np.sqrt(x/10 + 0.5)

    # check marker only plot
    fig, ax = plt.subplots()
    ax.plot(x, y, 'o', label='default')
    ax.plot(x, y, 'd', markevery=None, label='mark all')
    ax.plot(x, y, 's', markevery=10, label='mark every 10')
    ax.plot(x, y, '+', markevery=(5, 20), label='mark every 5 starting at 10')
    ax.legend()


@image_comparison(['markevery_line'], remove_text=True)
def test_markevery_line():
    x = np.linspace(0, 10, 100)
    y = np.sin(x) * np.sqrt(x/10 + 0.5)

    # check line/marker combos
    fig, ax = plt.subplots()
    ax.plot(x, y, '-o', label='default')
    ax.plot(x, y, '-d', markevery=None, label='mark all')
    ax.plot(x, y, '-s', markevery=10, label='mark every 10')
    ax.plot(x, y, '-+', markevery=(5, 20), label='mark every 5 starting at 10')
    ax.legend()


@image_comparison(['markevery_linear_scales'], remove_text=True)
def test_markevery_linear_scales():
    cases = [None,
             8,
             (30, 8),
             [16, 24, 30], [0, -1],
             slice(100, 200, 3),
             0.1, 0.3, 1.5,
             (0.0, 0.1), (0.45, 0.1)]

    cols = 3
    gs = matplotlib.gridspec.GridSpec(len(cases) // cols + 1, cols)

    delta = 0.11
    x = np.linspace(0, 10 - 2 * delta, 200) + delta
    y = np.sin(x) + 1.0 + delta

    for i, case in enumerate(cases):
        row = (i // cols)
        col = i % cols
        plt.subplot(gs[row, col])
        plt.title('markevery=%s' % str(case))
        plt.plot(x, y, 'o', ls='-', ms=4,  markevery=case)


@image_comparison(['markevery_linear_scales_zoomed'], remove_text=True)
def test_markevery_linear_scales_zoomed():
    cases = [None,
             8,
             (30, 8),
             [16, 24, 30], [0, -1],
             slice(100, 200, 3),
             0.1, 0.3, 1.5,
             (0.0, 0.1), (0.45, 0.1)]

    cols = 3
    gs = matplotlib.gridspec.GridSpec(len(cases) // cols + 1, cols)

    delta = 0.11
    x = np.linspace(0, 10 - 2 * delta, 200) + delta
    y = np.sin(x) + 1.0 + delta

    for i, case in enumerate(cases):
        row = (i // cols)
        col = i % cols
        plt.subplot(gs[row, col])
        plt.title('markevery=%s' % str(case))
        plt.plot(x, y, 'o', ls='-', ms=4,  markevery=case)
        plt.xlim((6, 6.7))
        plt.ylim((1.1, 1.7))


@image_comparison(['markevery_log_scales'], remove_text=True)
def test_markevery_log_scales():
    cases = [None,
             8,
             (30, 8),
             [16, 24, 30], [0, -1],
             slice(100, 200, 3),
             0.1, 0.3, 1.5,
             (0.0, 0.1), (0.45, 0.1)]

    cols = 3
    gs = matplotlib.gridspec.GridSpec(len(cases) // cols + 1, cols)

    delta = 0.11
    x = np.linspace(0, 10 - 2 * delta, 200) + delta
    y = np.sin(x) + 1.0 + delta

    for i, case in enumerate(cases):
        row = (i // cols)
        col = i % cols
        plt.subplot(gs[row, col])
        plt.title('markevery=%s' % str(case))
        plt.xscale('log')
        plt.yscale('log')
        plt.plot(x, y, 'o', ls='-', ms=4,  markevery=case)


@image_comparison(['markevery_polar'], style='default', remove_text=True)
def test_markevery_polar():
    cases = [None,
             8,
             (30, 8),
             [16, 24, 30], [0, -1],
             slice(100, 200, 3),
             0.1, 0.3, 1.5,
             (0.0, 0.1), (0.45, 0.1)]

    cols = 3
    gs = matplotlib.gridspec.GridSpec(len(cases) // cols + 1, cols)

    r = np.linspace(0, 3.0, 200)
    theta = 2 * np.pi * r

    for i, case in enumerate(cases):
        row = (i // cols)
        col = i % cols
        plt.subplot(gs[row, col], polar=True)
        plt.title('markevery=%s' % str(case))
        plt.plot(theta, r, 'o', ls='-', ms=4,  markevery=case)


@image_comparison(['marker_edges'], remove_text=True)
def test_marker_edges():
    x = np.linspace(0, 1, 10)
    fig, ax = plt.subplots()
    ax.plot(x, np.sin(x), 'y.', ms=30.0, mew=0, mec='r')
    ax.plot(x+0.1, np.sin(x), 'y.', ms=30.0, mew=1, mec='r')
    ax.plot(x+0.2, np.sin(x), 'y.', ms=30.0, mew=2, mec='b')


@image_comparison(['bar_tick_label_single.png', 'bar_tick_label_single.png'])
def test_bar_tick_label_single():
    # From 2516: plot bar with array of string labels for x axis
    ax = plt.gca()
    ax.bar(0, 1, align='edge', tick_label='0')

    # Reuse testcase from above for a labeled data test
    data = {"a": 0, "b": 1}
    fig, ax = plt.subplots()
    ax = plt.gca()
    ax.bar("a", "b", align='edge', tick_label='0', data=data)


def test_nan_bar_values():
    fig, ax = plt.subplots()
    ax.bar([0, 1], [np.nan, 4])


def test_bar_ticklabel_fail():
    fig, ax = plt.subplots()
    ax.bar([], [])


@image_comparison(['bar_tick_label_multiple.png'])
def test_bar_tick_label_multiple():
    # From 2516: plot bar with array of string labels for x axis
    ax = plt.gca()
    ax.bar([1, 2.5], [1, 2], width=[0.2, 0.5], tick_label=['a', 'b'],
           align='center')


@image_comparison(['bar_tick_label_multiple_old_label_alignment.png'])
def test_bar_tick_label_multiple_old_alignment():
    # Test that the alignment for class is backward compatible
    matplotlib.rcParams["ytick.alignment"] = "center"
    ax = plt.gca()
    ax.bar([1, 2.5], [1, 2], width=[0.2, 0.5], tick_label=['a', 'b'],
           align='center')


@check_figures_equal(extensions=["png"])
def test_bar_decimal_center(fig_test, fig_ref):
    ax = fig_test.subplots()
    x0 = [1.5, 8.4, 5.3, 4.2]
    y0 = [1.1, 2.2, 3.3, 4.4]
    x = [Decimal(x) for x in x0]
    y = [Decimal(y) for y in y0]
    # Test image - vertical, align-center bar chart with Decimal() input
    ax.bar(x, y, align='center')
    # Reference image
    ax = fig_ref.subplots()
    ax.bar(x0, y0, align='center')


@check_figures_equal(extensions=["png"])
def test_barh_decimal_center(fig_test, fig_ref):
    ax = fig_test.subplots()
    x0 = [1.5, 8.4, 5.3, 4.2]
    y0 = [1.1, 2.2, 3.3, 4.4]
    x = [Decimal(x) for x in x0]
    y = [Decimal(y) for y in y0]
    # Test image - horizontal, align-center bar chart with Decimal() input
    ax.barh(x, y, height=[0.5, 0.5, 1, 1], align='center')
    # Reference image
    ax = fig_ref.subplots()
    ax.barh(x0, y0, height=[0.5, 0.5, 1, 1], align='center')


@check_figures_equal(extensions=["png"])
def test_bar_decimal_width(fig_test, fig_ref):
    x = [1.5, 8.4, 5.3, 4.2]
    y = [1.1, 2.2, 3.3, 4.4]
    w0 = [0.7, 1.45, 1, 2]
    w = [Decimal(i) for i in w0]
    # Test image - vertical bar chart with Decimal() width
    ax = fig_test.subplots()
    ax.bar(x, y, width=w, align='center')
    # Reference image
    ax = fig_ref.subplots()
    ax.bar(x, y, width=w0, align='center')


@check_figures_equal(extensions=["png"])
def test_barh_decimal_height(fig_test, fig_ref):
    x = [1.5, 8.4, 5.3, 4.2]
    y = [1.1, 2.2, 3.3, 4.4]
    h0 = [0.7, 1.45, 1, 2]
    h = [Decimal(i) for i in h0]
    # Test image - horizontal bar chart with Decimal() height
    ax = fig_test.subplots()
    ax.barh(x, y, height=h, align='center')
    # Reference image
    ax = fig_ref.subplots()
    ax.barh(x, y, height=h0, align='center')


def test_bar_color_none_alpha():
    ax = plt.gca()
    rects = ax.bar([1, 2], [2, 4], alpha=0.3, color='none', edgecolor='r')
    for rect in rects:
        assert rect.get_facecolor() == (0, 0, 0, 0)
        assert rect.get_edgecolor() == (1, 0, 0, 0.3)


def test_bar_edgecolor_none_alpha():
    ax = plt.gca()
    rects = ax.bar([1, 2], [2, 4], alpha=0.3, color='r', edgecolor='none')
    for rect in rects:
        assert rect.get_facecolor() == (1, 0, 0, 0.3)
        assert rect.get_edgecolor() == (0, 0, 0, 0)


@image_comparison(['barh_tick_label.png'])
def test_barh_tick_label():
    # From 2516: plot barh with array of string labels for y axis
    ax = plt.gca()
    ax.barh([1, 2.5], [1, 2], height=[0.2, 0.5], tick_label=['a', 'b'],
            align='center')


def test_bar_timedelta():
    """Smoketest that bar can handle width and height in delta units."""
    fig, ax = plt.subplots()
    ax.bar(datetime.datetime(2018, 1, 1), 1.,
           width=datetime.timedelta(hours=3))
    ax.bar(datetime.datetime(2018, 1, 1), 1.,
           xerr=datetime.timedelta(hours=2),
           width=datetime.timedelta(hours=3))
    fig, ax = plt.subplots()
    ax.barh(datetime.datetime(2018, 1, 1), 1,
            height=datetime.timedelta(hours=3))
    ax.barh(datetime.datetime(2018, 1, 1), 1,
            height=datetime.timedelta(hours=3),
            yerr=datetime.timedelta(hours=2))
    fig, ax = plt.subplots()
    ax.barh([datetime.datetime(2018, 1, 1), datetime.datetime(2018, 1, 1)],
            np.array([1, 1.5]),
            height=datetime.timedelta(hours=3))
    ax.barh([datetime.datetime(2018, 1, 1), datetime.datetime(2018, 1, 1)],
            np.array([1, 1.5]),
            height=[datetime.timedelta(hours=t) for t in [1, 2]])
    ax.broken_barh([(datetime.datetime(2018, 1, 1),
                     datetime.timedelta(hours=1))],
                   (10, 20))


def test_boxplot_dates_pandas(pd):
    # smoke test for boxplot and dates in pandas
    data = np.random.rand(5, 2)
    years = pd.date_range('1/1/2000',
                          periods=2, freq=pd.DateOffset(years=1)).year
    plt.figure()
    plt.boxplot(data, positions=years)


def test_bar_pandas(pd):
    # Smoke test for pandas
    df = pd.DataFrame(
        {'year': [2018, 2018, 2018],
         'month': [1, 1, 1],
         'day': [1, 2, 3],
         'value': [1, 2, 3]})
    df['date'] = pd.to_datetime(df[['year', 'month', 'day']])

    monthly = df[['date', 'value']].groupby(['date']).sum()
    dates = monthly.index
    forecast = monthly['value']
    baseline = monthly['value']

    fig, ax = plt.subplots()
    ax.bar(dates, forecast, width=10, align='center')
    ax.plot(dates, baseline, color='orange', lw=4)


def test_bar_pandas_indexed(pd):
    # Smoke test for indexed pandas
    df = pd.DataFrame({"x": [1., 2., 3.], "width": [.2, .4, .6]},
                      index=[1, 2, 3])
    fig, ax = plt.subplots()
    ax.bar(df.x, 1., width=df.width)


@check_figures_equal()
@pytest.mark.style('default')
def test_bar_hatches(fig_test, fig_ref):
    ax_test = fig_test.subplots()
    ax_ref = fig_ref.subplots()

    x = [1, 2]
    y = [2, 3]
    hatches = ['x', 'o']
    for i in range(2):
        ax_ref.bar(x[i], y[i], color='C0', hatch=hatches[i])

    ax_test.bar(x, y, hatch=hatches)


def test_pandas_minimal_plot(pd):
    # smoke test that series and index objcets do not warn
    x = pd.Series([1, 2], dtype="float64")
    plt.plot(x, x)
    plt.plot(x.index, x)
    plt.plot(x)
    plt.plot(x.index)


@image_comparison(['hist_log'], remove_text=True)
def test_hist_log():
    data0 = np.linspace(0, 1, 200)**3
    data = np.concatenate([1 - data0, 1 + data0])
    fig, ax = plt.subplots()
    ax.hist(data, fill=False, log=True)


@check_figures_equal(extensions=["png"])
def test_hist_log_2(fig_test, fig_ref):
    axs_test = fig_test.subplots(2, 3)
    axs_ref = fig_ref.subplots(2, 3)
    for i, histtype in enumerate(["bar", "step", "stepfilled"]):
        # Set log scale, then call hist().
        axs_test[0, i].set_yscale("log")
        axs_test[0, i].hist(1, 1, histtype=histtype)
        # Call hist(), then set log scale.
        axs_test[1, i].hist(1, 1, histtype=histtype)
        axs_test[1, i].set_yscale("log")
        # Use hist(..., log=True).
        for ax in axs_ref[:, i]:
            ax.hist(1, 1, log=True, histtype=histtype)


def test_hist_log_barstacked():
    fig, axs = plt.subplots(2)
    axs[0].hist([[0], [0, 1]], 2, histtype="barstacked")
    axs[0].set_yscale("log")
    axs[1].hist([0, 0, 1], 2, histtype="barstacked")
    axs[1].set_yscale("log")
    fig.canvas.draw()
    assert axs[0].get_ylim() == axs[1].get_ylim()


@image_comparison(['hist_bar_empty.png'], remove_text=True)
def test_hist_bar_empty():
    # From #3886: creating hist from empty dataset raises ValueError
    ax = plt.gca()
    ax.hist([], histtype='bar')


@image_comparison(['hist_step_empty.png'], remove_text=True)
def test_hist_step_empty():
    # From #3886: creating hist from empty dataset raises ValueError
    ax = plt.gca()
    ax.hist([], histtype='step')


@image_comparison(['hist_step_filled.png'], remove_text=True)
def test_hist_step_filled():
    np.random.seed(0)
    x = np.random.randn(1000, 3)
    n_bins = 10

    kwargs = [{'fill': True}, {'fill': False}, {'fill': None}, {}]*2
    types = ['step']*4+['stepfilled']*4
    fig, axs = plt.subplots(nrows=2, ncols=4)

    for kg, _type, ax in zip(kwargs, types, axs.flat):
        ax.hist(x, n_bins, histtype=_type, stacked=True, **kg)
        ax.set_title('%s/%s' % (kg, _type))
        ax.set_ylim(bottom=-50)

    patches = axs[0, 0].patches
    assert all(p.get_facecolor() == p.get_edgecolor() for p in patches)


@image_comparison(['hist_density.png'])
def test_hist_density():
    np.random.seed(19680801)
    data = np.random.standard_normal(2000)
    fig, ax = plt.subplots()
    ax.hist(data, density=True)


def test_hist_unequal_bins_density():
    # Test correct behavior of normalized histogram with unequal bins
    # https://github.com/matplotlib/matplotlib/issues/9557
    rng = np.random.RandomState(57483)
    t = rng.randn(100)
    bins = [-3, -1, -0.5, 0, 1, 5]
    mpl_heights, _, _ = plt.hist(t, bins=bins, density=True)
    np_heights, _ = np.histogram(t, bins=bins, density=True)
    assert_allclose(mpl_heights, np_heights)


def test_hist_datetime_datasets():
    data = [[datetime.datetime(2017, 1, 1), datetime.datetime(2017, 1, 1)],
            [datetime.datetime(2017, 1, 1), datetime.datetime(2017, 1, 2)]]
    fig, ax = plt.subplots()
    ax.hist(data, stacked=True)
    ax.hist(data, stacked=False)


@pytest.mark.parametrize("bins_preprocess",
                         [mpl.dates.date2num,
                          lambda bins: bins,
                          lambda bins: np.asarray(bins).astype('datetime64')],
                         ids=['date2num', 'datetime.datetime',
                              'np.datetime64'])
def test_hist_datetime_datasets_bins(bins_preprocess):
    data = [[datetime.datetime(2019, 1, 5), datetime.datetime(2019, 1, 11),
             datetime.datetime(2019, 2, 1), datetime.datetime(2019, 3, 1)],
            [datetime.datetime(2019, 1, 11), datetime.datetime(2019, 2, 5),
             datetime.datetime(2019, 2, 18), datetime.datetime(2019, 3, 1)]]

    date_edges = [datetime.datetime(2019, 1, 1), datetime.datetime(2019, 2, 1),
                  datetime.datetime(2019, 3, 1)]

    fig, ax = plt.subplots()
    _, bins, _ = ax.hist(data, bins=bins_preprocess(date_edges), stacked=True)
    np.testing.assert_allclose(bins, mpl.dates.date2num(date_edges))

    _, bins, _ = ax.hist(data, bins=bins_preprocess(date_edges), stacked=False)
    np.testing.assert_allclose(bins, mpl.dates.date2num(date_edges))


@pytest.mark.parametrize('data, expected_number_of_hists',
                         [([], 1),
                          ([[]], 1),
                          ([[], []], 2)])
def test_hist_with_empty_input(data, expected_number_of_hists):
    hists, _, _ = plt.hist(data)
    hists = np.asarray(hists)

    if hists.ndim == 1:
        assert 1 == expected_number_of_hists
    else:
        assert hists.shape[0] == expected_number_of_hists


@pytest.mark.parametrize("histtype, zorder",
                         [("bar", mpl.patches.Patch.zorder),
                          ("step", mpl.lines.Line2D.zorder),
                          ("stepfilled", mpl.patches.Patch.zorder)])
def test_hist_zorder(histtype, zorder):
    ax = plt.figure().add_subplot()
    ax.hist([1, 2], histtype=histtype)
    assert ax.patches
    for patch in ax.patches:
        assert patch.get_zorder() == zorder


@check_figures_equal(extensions=['png'])
def test_stairs(fig_test, fig_ref):
    import matplotlib.lines as mlines
    y = np.array([6, 14, 32, 37, 48, 32, 21,  4])  # hist
    x = np.array([1., 2., 3., 4., 5., 6., 7., 8., 9.])  # bins

    test_axes = fig_test.subplots(3, 2).flatten()
    test_axes[0].stairs(y, x, baseline=None)
    test_axes[1].stairs(y, x, baseline=None, orientation='horizontal')
    test_axes[2].stairs(y, x)
    test_axes[3].stairs(y, x, orientation='horizontal')
    test_axes[4].stairs(y, x)
    test_axes[4].semilogy()
    test_axes[5].stairs(y, x, orientation='horizontal')
    test_axes[5].semilogx()

    # defaults of `PathPatch` to be used for all following Line2D
    style = {'solid_joinstyle': 'miter', 'solid_capstyle': 'butt'}

    ref_axes = fig_ref.subplots(3, 2).flatten()
    ref_axes[0].plot(x, np.append(y, y[-1]), drawstyle='steps-post', **style)
    ref_axes[1].plot(np.append(y[0], y), x, drawstyle='steps-post', **style)

    ref_axes[2].plot(x, np.append(y, y[-1]), drawstyle='steps-post', **style)
    ref_axes[2].add_line(mlines.Line2D([x[0], x[0]], [0, y[0]], **style))
    ref_axes[2].add_line(mlines.Line2D([x[-1], x[-1]], [0, y[-1]], **style))
    ref_axes[2].set_ylim(0, None)

    ref_axes[3].plot(np.append(y[0], y), x, drawstyle='steps-post', **style)
    ref_axes[3].add_line(mlines.Line2D([0, y[0]], [x[0], x[0]], **style))
    ref_axes[3].add_line(mlines.Line2D([0, y[-1]], [x[-1], x[-1]], **style))
    ref_axes[3].set_xlim(0, None)

    ref_axes[4].plot(x, np.append(y, y[-1]), drawstyle='steps-post', **style)
    ref_axes[4].add_line(mlines.Line2D([x[0], x[0]], [0, y[0]], **style))
    ref_axes[4].add_line(mlines.Line2D([x[-1], x[-1]], [0, y[-1]], **style))
    ref_axes[4].semilogy()

    ref_axes[5].plot(np.append(y[0], y), x, drawstyle='steps-post', **style)
    ref_axes[5].add_line(mlines.Line2D([0, y[0]], [x[0], x[0]], **style))
    ref_axes[5].add_line(mlines.Line2D([0, y[-1]], [x[-1], x[-1]], **style))
    ref_axes[5].semilogx()


@check_figures_equal(extensions=['png'])
def test_stairs_fill(fig_test, fig_ref):
    h, bins = [1, 2, 3, 4, 2], [0, 1, 2, 3, 4, 5]
    bs = -2
    # Test
    test_axes = fig_test.subplots(2, 2).flatten()
    test_axes[0].stairs(h, bins, fill=True)
    test_axes[1].stairs(h, bins, orientation='horizontal', fill=True)
    test_axes[2].stairs(h, bins, baseline=bs, fill=True)
    test_axes[3].stairs(h, bins, baseline=bs, orientation='horizontal',
                        fill=True)

    # # Ref
    ref_axes = fig_ref.subplots(2, 2).flatten()
    ref_axes[0].fill_between(bins, np.append(h, h[-1]), step='post', lw=0)
    ref_axes[0].set_ylim(0, None)
    ref_axes[1].fill_betweenx(bins, np.append(h, h[-1]), step='post', lw=0)
    ref_axes[1].set_xlim(0, None)
    ref_axes[2].fill_between(bins, np.append(h, h[-1]),
                             np.ones(len(h)+1)*bs, step='post', lw=0)
    ref_axes[2].set_ylim(bs, None)
    ref_axes[3].fill_betweenx(bins, np.append(h, h[-1]),
                              np.ones(len(h)+1)*bs, step='post', lw=0)
    ref_axes[3].set_xlim(bs, None)


@check_figures_equal(extensions=['png'])
def test_stairs_update(fig_test, fig_ref):
    # fixed ylim because stairs() does autoscale, but updating data does not
    ylim = -3, 4
    # Test
    test_ax = fig_test.add_subplot()
    h = test_ax.stairs([1, 2, 3])
    test_ax.set_ylim(ylim)
    h.set_data([3, 2, 1])
    h.set_data(edges=np.arange(4)+2)
    h.set_data([1, 2, 1], np.arange(4)/2)
    h.set_data([1, 2, 3])
    h.set_data(None, np.arange(4))
    assert np.allclose(h.get_data()[0], np.arange(1, 4))
    assert np.allclose(h.get_data()[1], np.arange(4))
    h.set_data(baseline=-2)
    assert h.get_data().baseline == -2

    # Ref
    ref_ax = fig_ref.add_subplot()
    h = ref_ax.stairs([1, 2, 3], baseline=-2)
    ref_ax.set_ylim(ylim)


@check_figures_equal(extensions=['png'])
def test_stairs_baseline_0(fig_test, fig_ref):
    # Test
    test_ax = fig_test.add_subplot()
    test_ax.stairs([5, 6, 7], baseline=None)

    # Ref
    ref_ax = fig_ref.add_subplot()
    style = {'solid_joinstyle': 'miter', 'solid_capstyle': 'butt'}
    ref_ax.plot(range(4), [5, 6, 7, 7], drawstyle='steps-post', **style)
    ref_ax.set_ylim(0, None)


def test_stairs_invalid_nan():
    with pytest.raises(ValueError, match='Nan values in "edges"'):
        plt.stairs([1, 2], [0, np.nan, 1])


def test_stairs_invalid_mismatch():
    with pytest.raises(ValueError, match='Size mismatch'):
        plt.stairs([1, 2], [0, 1])


def test_stairs_invalid_update():
    h = plt.stairs([1, 2], [0, 1, 2])
    with pytest.raises(ValueError, match='Nan values in "edges"'):
        h.set_data(edges=[1, np.nan, 2])


def test_stairs_invalid_update2():
    h = plt.stairs([1, 2], [0, 1, 2])
    with pytest.raises(ValueError, match='Size mismatch'):
        h.set_data(edges=np.arange(5))


@image_comparison(['test_stairs_options.png'], remove_text=True)
def test_stairs_options():
    x, y = np.array([1, 2, 3, 4, 5]), np.array([1, 2, 3, 4]).astype(float)
    yn = y.copy()
    yn[1] = np.nan

    fig, ax = plt.subplots()
    ax.stairs(y*3, x, color='green', fill=True, label="A")
    ax.stairs(y, x*3-3, color='red', fill=True,
              orientation='horizontal', label="B")
    ax.stairs(yn, x, color='orange', ls='--', lw=2, label="C")
    ax.stairs(yn/3, x*3-2, ls='--', lw=2, baseline=0.5,
              orientation='horizontal', label="D")
    ax.stairs(y[::-1]*3+13, x-1, color='red', ls='--', lw=2, baseline=None,
              label="E")
    ax.stairs(y[::-1]*3+14, x, baseline=26,
              color='purple', ls='--', lw=2, label="F")
    ax.stairs(yn[::-1]*3+15, x+1, baseline=np.linspace(27, 25, len(y)),
              color='blue', ls='--', lw=2, label="G", fill=True)
    ax.stairs(y[:-1][::-1]*2+11, x[:-1]+0.5, color='black', ls='--', lw=2,
              baseline=12, hatch='//', label="H")
    ax.legend(loc=0)


@image_comparison(['test_stairs_datetime.png'])
def test_stairs_datetime():
    f, ax = plt.subplots(constrained_layout=True)
    ax.stairs(np.arange(36),
              np.arange(np.datetime64('2001-12-27'),
                        np.datetime64('2002-02-02')))
    plt.xticks(rotation=30)


def contour_dat():
    x = np.linspace(-3, 5, 150)
    y = np.linspace(-3, 5, 120)
    z = np.cos(x) + np.sin(y[:, np.newaxis])
    return x, y, z


@image_comparison(['contour_hatching'], remove_text=True, style='mpl20')
def test_contour_hatching():
    x, y, z = contour_dat()
    fig, ax = plt.subplots()
    ax.contourf(x, y, z, 7, hatches=['/', '\\', '//', '-'],
                cmap=plt.get_cmap('gray'),
                extend='both', alpha=0.5)


@image_comparison(['contour_colorbar'], style='mpl20')
def test_contour_colorbar():
    # Remove this line when this test image is regenerated.
    plt.rcParams['pcolormesh.snap'] = False

    x, y, z = contour_dat()

    fig, ax = plt.subplots()
    cs = ax.contourf(x, y, z, levels=np.arange(-1.8, 1.801, 0.2),
                     cmap=plt.get_cmap('RdBu'),
                     vmin=-0.6,
                     vmax=0.6,
                     extend='both')
    cs1 = ax.contour(x, y, z, levels=np.arange(-2.2, -0.599, 0.2),
                     colors=['y'],
                     linestyles='solid',
                     linewidths=2)
    cs2 = ax.contour(x, y, z, levels=np.arange(0.6, 2.2, 0.2),
                     colors=['c'],
                     linewidths=2)
    cbar = fig.colorbar(cs, ax=ax)
    cbar.add_lines(cs1)
    cbar.add_lines(cs2, erase=False)


@image_comparison(['hist2d', 'hist2d'], remove_text=True, style='mpl20')
def test_hist2d():
    # Remove this line when this test image is regenerated.
    plt.rcParams['pcolormesh.snap'] = False

    np.random.seed(0)
    # make it not symmetric in case we switch x and y axis
    x = np.random.randn(100)*2+5
    y = np.random.randn(100)-2
    fig, ax = plt.subplots()
    ax.hist2d(x, y, bins=10, rasterized=True)

    # Reuse testcase from above for a labeled data test
    data = {"x": x, "y": y}
    fig, ax = plt.subplots()
    ax.hist2d("x", "y", bins=10, data=data, rasterized=True)


@image_comparison(['hist2d_transpose'], remove_text=True, style='mpl20')
def test_hist2d_transpose():
    # Remove this line when this test image is regenerated.
    plt.rcParams['pcolormesh.snap'] = False

    np.random.seed(0)
    # make sure the output from np.histogram is transposed before
    # passing to pcolorfast
    x = np.array([5]*100)
    y = np.random.randn(100)-2
    fig, ax = plt.subplots()
    ax.hist2d(x, y, bins=10, rasterized=True)


def test_hist2d_density():
    x, y = np.random.random((2, 100))
    ax = plt.figure().subplots()
    for obj in [ax, plt]:
        obj.hist2d(x, y, density=True)


class TestScatter:
    @image_comparison(['scatter'], style='mpl20', remove_text=True)
    def test_scatter_plot(self):
        data = {"x": np.array([3, 4, 2, 6]), "y": np.array([2, 5, 2, 3]),
                "c": ['r', 'y', 'b', 'lime'], "s": [24, 15, 19, 29],
                "c2": ['0.5', '0.6', '0.7', '0.8']}

        fig, ax = plt.subplots()
        ax.scatter(data["x"] - 1., data["y"] - 1., c=data["c"], s=data["s"])
        ax.scatter(data["x"] + 1., data["y"] + 1., c=data["c2"], s=data["s"])
        ax.scatter("x", "y", c="c", s="s", data=data)

    @image_comparison(['scatter_marker.png'], remove_text=True)
    def test_scatter_marker(self):
        fig, (ax0, ax1, ax2) = plt.subplots(ncols=3)
        ax0.scatter([3, 4, 2, 6], [2, 5, 2, 3],
                    c=[(1, 0, 0), 'y', 'b', 'lime'],
                    s=[60, 50, 40, 30],
                    edgecolors=['k', 'r', 'g', 'b'],
                    marker='s')
        ax1.scatter([3, 4, 2, 6], [2, 5, 2, 3],
                    c=[(1, 0, 0), 'y', 'b', 'lime'],
                    s=[60, 50, 40, 30],
                    edgecolors=['k', 'r', 'g', 'b'],
                    marker=mmarkers.MarkerStyle('o', fillstyle='top'))
        # unit area ellipse
        rx, ry = 3, 1
        area = rx * ry * np.pi
        theta = np.linspace(0, 2 * np.pi, 21)
        verts = np.column_stack([np.cos(theta) * rx / area,
                                 np.sin(theta) * ry / area])
        ax2.scatter([3, 4, 2, 6], [2, 5, 2, 3],
                    c=[(1, 0, 0), 'y', 'b', 'lime'],
                    s=[60, 50, 40, 30],
                    edgecolors=['k', 'r', 'g', 'b'],
                    marker=verts)

    @image_comparison(['scatter_2D'], remove_text=True, extensions=['png'])
    def test_scatter_2D(self):
        x = np.arange(3)
        y = np.arange(2)
        x, y = np.meshgrid(x, y)
        z = x + y
        fig, ax = plt.subplots()
        ax.scatter(x, y, c=z, s=200, edgecolors='face')

    @check_figures_equal(extensions=["png"])
    def test_scatter_decimal(self, fig_test, fig_ref):
        x0 = np.array([1.5, 8.4, 5.3, 4.2])
        y0 = np.array([1.1, 2.2, 3.3, 4.4])
        x = np.array([Decimal(i) for i in x0])
        y = np.array([Decimal(i) for i in y0])
        c = ['r', 'y', 'b', 'lime']
        s = [24, 15, 19, 29]
        # Test image - scatter plot with Decimal() input
        ax = fig_test.subplots()
        ax.scatter(x, y, c=c, s=s)
        # Reference image
        ax = fig_ref.subplots()
        ax.scatter(x0, y0, c=c, s=s)

    def test_scatter_color(self):
        # Try to catch cases where 'c' kwarg should have been used.
        with pytest.raises(ValueError):
            plt.scatter([1, 2], [1, 2], color=[0.1, 0.2])
        with pytest.raises(ValueError):
            plt.scatter([1, 2, 3], [1, 2, 3], color=[1, 2, 3])

    def test_scatter_unfilled(self):
        coll = plt.scatter([0, 1, 2], [1, 3, 2], c=['0.1', '0.3', '0.5'],
                           marker=mmarkers.MarkerStyle('o', fillstyle='none'),
                           linewidths=[1.1, 1.2, 1.3])
        assert coll.get_facecolors().shape == (0, 4)  # no facecolors
        assert_array_equal(coll.get_edgecolors(), [[0.1, 0.1, 0.1, 1],
                                                   [0.3, 0.3, 0.3, 1],
                                                   [0.5, 0.5, 0.5, 1]])
        assert_array_equal(coll.get_linewidths(), [1.1, 1.2, 1.3])

    def test_scatter_size_arg_size(self):
        x = np.arange(4)
        with pytest.raises(ValueError, match='same size as x and y'):
            plt.scatter(x, x, x[1:])
        with pytest.raises(ValueError, match='same size as x and y'):
            plt.scatter(x[1:], x[1:], x)
        with pytest.raises(ValueError, match='float array-like'):
            plt.scatter(x, x, 'foo')

    @check_figures_equal(extensions=["png"])
    def test_scatter_invalid_color(self, fig_test, fig_ref):
        ax = fig_test.subplots()
        cmap = plt.get_cmap("viridis", 16)
        cmap.set_bad("k", 1)
        # Set a nonuniform size to prevent the last call to `scatter` (plotting
        # the invalid points separately in fig_ref) from using the marker
        # stamping fast path, which would result in slightly offset markers.
        ax.scatter(range(4), range(4),
                   c=[1, np.nan, 2, np.nan], s=[1, 2, 3, 4],
                   cmap=cmap, plotnonfinite=True)
        ax = fig_ref.subplots()
        cmap = plt.get_cmap("viridis", 16)
        ax.scatter([0, 2], [0, 2], c=[1, 2], s=[1, 3], cmap=cmap)
        ax.scatter([1, 3], [1, 3], s=[2, 4], color="k")

    @check_figures_equal(extensions=["png"])
    def test_scatter_no_invalid_color(self, fig_test, fig_ref):
        # With plotninfinite=False we plot only 2 points.
        ax = fig_test.subplots()
        cmap = plt.get_cmap("viridis", 16)
        cmap.set_bad("k", 1)
        ax.scatter(range(4), range(4),
                   c=[1, np.nan, 2, np.nan], s=[1, 2, 3, 4],
                   cmap=cmap, plotnonfinite=False)
        ax = fig_ref.subplots()
        ax.scatter([0, 2], [0, 2], c=[1, 2], s=[1, 3], cmap=cmap)

    @check_figures_equal(extensions=["png"])
    def test_scatter_norm_vminvmax(self, fig_test, fig_ref):
        """Parameters vmin, vmax should be ignored if norm is given."""
        x = [1, 2, 3]
        ax = fig_ref.subplots()
        ax.scatter(x, x, c=x, vmin=0, vmax=5)
        ax = fig_test.subplots()
        with pytest.warns(MatplotlibDeprecationWarning,
                          match="Passing parameters norm and vmin/vmax "
                                "simultaneously is deprecated."):
            ax.scatter(x, x, c=x, norm=mcolors.Normalize(-10, 10),
                       vmin=0, vmax=5)

    @check_figures_equal(extensions=["png"])
    def test_scatter_single_point(self, fig_test, fig_ref):
        ax = fig_test.subplots()
        ax.scatter(1, 1, c=1)
        ax = fig_ref.subplots()
        ax.scatter([1], [1], c=[1])

    @check_figures_equal(extensions=["png"])
    def test_scatter_different_shapes(self, fig_test, fig_ref):
        x = np.arange(10)
        ax = fig_test.subplots()
        ax.scatter(x, x.reshape(2, 5), c=x.reshape(5, 2))
        ax = fig_ref.subplots()
        ax.scatter(x.reshape(5, 2), x, c=x.reshape(2, 5))

    # Parameters for *test_scatter_c*. NB: assuming that the
    # scatter plot will have 4 elements. The tuple scheme is:
    # (*c* parameter case, exception regexp key or None if no exception)
    params_test_scatter_c = [
        # single string:
        ('0.5', None),
        # Single letter-sequences
        (["rgby"], "conversion"),
        # Special cases
        ("red", None),
        ("none", None),
        (None, None),
        (["r", "g", "b", "none"], None),
        # Non-valid color spec (FWIW, 'jaune' means yellow in French)
        ("jaune", "conversion"),
        (["jaune"], "conversion"),  # wrong type before wrong size
        (["jaune"]*4, "conversion"),
        # Value-mapping like
        ([0.5]*3, None),  # should emit a warning for user's eyes though
        ([0.5]*4, None),  # NB: no warning as matching size allows mapping
        ([0.5]*5, "shape"),
        # list of strings:
        (['0.5', '0.4', '0.6', '0.7'], None),
        (['0.5', 'red', '0.6', 'C5'], None),
        (['0.5', 0.5, '0.6', 'C5'], "conversion"),
        # RGB values
        ([[1, 0, 0]], None),
        ([[1, 0, 0]]*3, "shape"),
        ([[1, 0, 0]]*4, None),
        ([[1, 0, 0]]*5, "shape"),
        # RGBA values
        ([[1, 0, 0, 0.5]], None),
        ([[1, 0, 0, 0.5]]*3, "shape"),
        ([[1, 0, 0, 0.5]]*4, None),
        ([[1, 0, 0, 0.5]]*5, "shape"),
        # Mix of valid color specs
        ([[1, 0, 0, 0.5]]*3 + [[1, 0, 0]], None),
        ([[1, 0, 0, 0.5], "red", "0.0"], "shape"),
        ([[1, 0, 0, 0.5], "red", "0.0", "C5"], None),
        ([[1, 0, 0, 0.5], "red", "0.0", "C5", [0, 1, 0]], "shape"),
        # Mix of valid and non valid color specs
        ([[1, 0, 0, 0.5], "red", "jaune"], "conversion"),
        ([[1, 0, 0, 0.5], "red", "0.0", "jaune"], "conversion"),
        ([[1, 0, 0, 0.5], "red", "0.0", "C5", "jaune"], "conversion"),
    ]

    @pytest.mark.parametrize('c_case, re_key', params_test_scatter_c)
    def test_scatter_c(self, c_case, re_key):
        def get_next_color():
            return 'blue'  # currently unused

        xsize = 4
        # Additional checking of *c* (introduced in #11383).
        REGEXP = {
            "shape": "^'c' argument has [0-9]+ elements",  # shape mismatch
            "conversion": "^'c' argument must be a color",  # bad vals
            }

        if re_key is None:
            mpl.axes.Axes._parse_scatter_color_args(
                c=c_case, edgecolors="black", kwargs={}, xsize=xsize,
                get_next_color_func=get_next_color)
        else:
            with pytest.raises(ValueError, match=REGEXP[re_key]):
                mpl.axes.Axes._parse_scatter_color_args(
                    c=c_case, edgecolors="black", kwargs={}, xsize=xsize,
                    get_next_color_func=get_next_color)

    @pytest.mark.style('default')
    @check_figures_equal(extensions=["png"])
    def test_scatter_single_color_c(self, fig_test, fig_ref):
        rgb = [[1, 0.5, 0.05]]
        rgba = [[1, 0.5, 0.05, .5]]

        # set via color kwarg
        ax_ref = fig_ref.subplots()
        ax_ref.scatter(np.ones(3), range(3), color=rgb)
        ax_ref.scatter(np.ones(4)*2, range(4), color=rgba)

        # set via broadcasting via c
        ax_test = fig_test.subplots()
        ax_test.scatter(np.ones(3), range(3), c=rgb)
        ax_test.scatter(np.ones(4)*2, range(4), c=rgba)

    def test_scatter_linewidths(self):
        x = np.arange(5)

        fig, ax = plt.subplots()
        for i in range(3):
            pc = ax.scatter(x, np.full(5, i), c=f'C{i}', marker='x', s=100,
                            linewidths=i + 1)
            assert pc.get_linewidths() == i + 1

        pc = ax.scatter(x, np.full(5, 3), c='C3', marker='x', s=100,
                        linewidths=[*range(1, 5), None])
        assert_array_equal(pc.get_linewidths(),
                           [*range(1, 5), mpl.rcParams['lines.linewidth']])


def _params(c=None, xsize=2, *, edgecolors=None, **kwargs):
    return (c, edgecolors, kwargs if kwargs is not None else {}, xsize)
_result = namedtuple('_result', 'c, colors')


@pytest.mark.parametrize(
    'params, expected_result',
    [(_params(),
      _result(c='b', colors=np.array([[0, 0, 1, 1]]))),
     (_params(c='r'),
      _result(c='r', colors=np.array([[1, 0, 0, 1]]))),
     (_params(c='r', colors='b'),
      _result(c='r', colors=np.array([[1, 0, 0, 1]]))),
     # color
     (_params(color='b'),
      _result(c='b', colors=np.array([[0, 0, 1, 1]]))),
     (_params(color=['b', 'g']),
      _result(c=['b', 'g'], colors=np.array([[0, 0, 1, 1], [0, .5, 0, 1]]))),
     ])
def test_parse_scatter_color_args(params, expected_result):
    def get_next_color():
        return 'blue'  # currently unused

    c, colors, _edgecolors = mpl.axes.Axes._parse_scatter_color_args(
        *params, get_next_color_func=get_next_color)
    assert c == expected_result.c
    assert_allclose(colors, expected_result.colors)

del _params
del _result


@pytest.mark.parametrize(
    'kwargs, expected_edgecolors',
    [(dict(), None),
     (dict(c='b'), None),
     (dict(edgecolors='r'), 'r'),
     (dict(edgecolors=['r', 'g']), ['r', 'g']),
     (dict(edgecolor='r'), 'r'),
     (dict(edgecolors='face'), 'face'),
     (dict(edgecolors='none'), 'none'),
     (dict(edgecolor='r', edgecolors='g'), 'r'),
     (dict(c='b', edgecolor='r', edgecolors='g'), 'r'),
     (dict(color='r'), 'r'),
     (dict(color='r', edgecolor='g'), 'g'),
     ])
def test_parse_scatter_color_args_edgecolors(kwargs, expected_edgecolors):
    def get_next_color():
        return 'blue'  # currently unused

    c = kwargs.pop('c', None)
    edgecolors = kwargs.pop('edgecolors', None)
    _, _, result_edgecolors = \
        mpl.axes.Axes._parse_scatter_color_args(
            c, edgecolors, kwargs, xsize=2, get_next_color_func=get_next_color)
    assert result_edgecolors == expected_edgecolors


def test_parse_scatter_color_args_error():
    def get_next_color():
        return 'blue'  # currently unused

    with pytest.raises(ValueError,
                       match="RGBA values should be within 0-1 range"):
        c = np.array([[0.1, 0.2, 0.7], [0.2, 0.4, 1.4]])  # value > 1
        mpl.axes.Axes._parse_scatter_color_args(
            c, None, kwargs={}, xsize=2, get_next_color_func=get_next_color)


def test_as_mpl_axes_api():
    # tests the _as_mpl_axes api
    from matplotlib.projections.polar import PolarAxes

    class Polar:
        def __init__(self):
            self.theta_offset = 0

        def _as_mpl_axes(self):
            # implement the matplotlib axes interface
            return PolarAxes, {'theta_offset': self.theta_offset}

    prj = Polar()
    prj2 = Polar()
    prj2.theta_offset = np.pi
    prj3 = Polar()

    # testing axes creation with plt.axes
    ax = plt.axes([0, 0, 1, 1], projection=prj)
    assert type(ax) == PolarAxes
    with pytest.warns(
            MatplotlibDeprecationWarning,
            match=r'Calling gca\(\) with keyword arguments was deprecated'):
        ax_via_gca = plt.gca(projection=prj)
    assert ax_via_gca is ax
    plt.close()

    # testing axes creation with gca
    with pytest.warns(
            MatplotlibDeprecationWarning,
            match=r'Calling gca\(\) with keyword arguments was deprecated'):
        ax = plt.gca(projection=prj)
    assert type(ax) == mpl.axes._subplots.subplot_class_factory(PolarAxes)
    with pytest.warns(
            MatplotlibDeprecationWarning,
            match=r'Calling gca\(\) with keyword arguments was deprecated'):
        ax_via_gca = plt.gca(projection=prj)
    assert ax_via_gca is ax
    # try getting the axes given a different polar projection
    with pytest.warns(
            MatplotlibDeprecationWarning,
            match=r'Calling gca\(\) with keyword arguments was deprecated'):
        ax_via_gca = plt.gca(projection=prj2)
    assert ax_via_gca is ax
    assert ax.get_theta_offset() == 0
    # try getting the axes given an == (not is) polar projection
    with pytest.warns(
            MatplotlibDeprecationWarning,
            match=r'Calling gca\(\) with keyword arguments was deprecated'):
        ax_via_gca = plt.gca(projection=prj3)
    assert ax_via_gca is ax
    plt.close()

    # testing axes creation with subplot
    ax = plt.subplot(121, projection=prj)
    assert type(ax) == mpl.axes._subplots.subplot_class_factory(PolarAxes)
    plt.close()


def test_pyplot_axes():
    # test focusing of Axes in other Figure
    fig1, ax1 = plt.subplots()
    fig2, ax2 = plt.subplots()
    plt.sca(ax1)
    assert ax1 is plt.gca()
    assert fig1 is plt.gcf()
    plt.close(fig1)
    plt.close(fig2)


@image_comparison(['log_scales'])
def test_log_scales():
    fig, ax = plt.subplots()
    ax.plot(np.log(np.linspace(0.1, 100)))
    ax.set_yscale('log', base=5.5)
    ax.invert_yaxis()
    ax.set_xscale('log', base=9.0)


def test_log_scales_no_data():
    _, ax = plt.subplots()
    ax.set(xscale="log", yscale="log")
    ax.xaxis.set_major_locator(mticker.MultipleLocator(1))
    assert ax.get_xlim() == ax.get_ylim() == (1, 10)


def test_log_scales_invalid():
    fig, ax = plt.subplots()
    ax.set_xscale('log')
    with pytest.warns(UserWarning, match='Attempted to set non-positive'):
        ax.set_xlim(-1, 10)
    ax.set_yscale('log')
    with pytest.warns(UserWarning, match='Attempted to set non-positive'):
        ax.set_ylim(-1, 10)


@image_comparison(['stackplot_test_image', 'stackplot_test_image'])
def test_stackplot():
    fig = plt.figure()
    x = np.linspace(0, 10, 10)
    y1 = 1.0 * x
    y2 = 2.0 * x + 1
    y3 = 3.0 * x + 2
    ax = fig.add_subplot(1, 1, 1)
    ax.stackplot(x, y1, y2, y3)
    ax.set_xlim((0, 10))
    ax.set_ylim((0, 70))

    # Reuse testcase from above for a labeled data test
    data = {"x": x, "y1": y1, "y2": y2, "y3": y3}
    fig, ax = plt.subplots()
    ax.stackplot("x", "y1", "y2", "y3", data=data)
    ax.set_xlim((0, 10))
    ax.set_ylim((0, 70))


@image_comparison(['stackplot_test_baseline'], remove_text=True)
def test_stackplot_baseline():
    np.random.seed(0)

    def layers(n, m):
        a = np.zeros((m, n))
        for i in range(n):
            for j in range(5):
                x = 1 / (.1 + np.random.random())
                y = 2 * np.random.random() - .5
                z = 10 / (.1 + np.random.random())
                a[:, i] += x * np.exp(-((np.arange(m) / m - y) * z) ** 2)
        return a

    d = layers(3, 100)
    d[50, :] = 0  # test for fixed weighted wiggle (issue #6313)

    fig, axs = plt.subplots(2, 2)

    axs[0, 0].stackplot(range(100), d.T, baseline='zero')
    axs[0, 1].stackplot(range(100), d.T, baseline='sym')
    axs[1, 0].stackplot(range(100), d.T, baseline='wiggle')
    axs[1, 1].stackplot(range(100), d.T, baseline='weighted_wiggle')


def _bxp_test_helper(
        stats_kwargs={}, transform_stats=lambda s: s, bxp_kwargs={}):
    np.random.seed(937)
    logstats = mpl.cbook.boxplot_stats(
        np.random.lognormal(mean=1.25, sigma=1., size=(37, 4)), **stats_kwargs)
    fig, ax = plt.subplots()
    if bxp_kwargs.get('vert', True):
        ax.set_yscale('log')
    else:
        ax.set_xscale('log')
    # Work around baseline images generate back when bxp did not respect the
    # boxplot.boxprops.linewidth rcParam when patch_artist is False.
    if not bxp_kwargs.get('patch_artist', False):
        mpl.rcParams['boxplot.boxprops.linewidth'] = \
            mpl.rcParams['lines.linewidth']
    ax.bxp(transform_stats(logstats), **bxp_kwargs)


@image_comparison(['bxp_baseline.png'],
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_baseline():
    _bxp_test_helper()


@image_comparison(['bxp_rangewhis.png'],
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_rangewhis():
    _bxp_test_helper(stats_kwargs=dict(whis=[0, 100]))


@image_comparison(['bxp_percentilewhis.png'],
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_percentilewhis():
    _bxp_test_helper(stats_kwargs=dict(whis=[5, 95]))


@image_comparison(['bxp_with_xlabels.png'],
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_with_xlabels():
    def transform(stats):
        for s, label in zip(stats, list('ABCD')):
            s['label'] = label
        return stats

    _bxp_test_helper(transform_stats=transform)


@image_comparison(['bxp_horizontal.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default',
                  tol=0.1)
def test_bxp_horizontal():
    _bxp_test_helper(bxp_kwargs=dict(vert=False))


@image_comparison(['bxp_with_ylabels.png'],
                  savefig_kwarg={'dpi': 40},
                  style='default',
                  tol=0.1)
def test_bxp_with_ylabels():
    def transform(stats):
        for s, label in zip(stats, list('ABCD')):
            s['label'] = label
        return stats

    _bxp_test_helper(transform_stats=transform, bxp_kwargs=dict(vert=False))


@image_comparison(['bxp_patchartist.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_patchartist():
    _bxp_test_helper(bxp_kwargs=dict(patch_artist=True))


@image_comparison(['bxp_custompatchartist.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 100},
                  style='default')
def test_bxp_custompatchartist():
    _bxp_test_helper(bxp_kwargs=dict(
        patch_artist=True,
        boxprops=dict(facecolor='yellow', edgecolor='green', ls=':')))


@image_comparison(['bxp_customoutlier.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_customoutlier():
    _bxp_test_helper(bxp_kwargs=dict(
        flierprops=dict(linestyle='none', marker='d', mfc='g')))


@image_comparison(['bxp_withmean_custompoint.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_showcustommean():
    _bxp_test_helper(bxp_kwargs=dict(
        showmeans=True,
        meanprops=dict(linestyle='none', marker='d', mfc='green'),
    ))


@image_comparison(['bxp_custombox.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_custombox():
    _bxp_test_helper(bxp_kwargs=dict(
        boxprops=dict(linestyle='--', color='b', lw=3)))


@image_comparison(['bxp_custommedian.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_custommedian():
    _bxp_test_helper(bxp_kwargs=dict(
        medianprops=dict(linestyle='--', color='b', lw=3)))


@image_comparison(['bxp_customcap.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_customcap():
    _bxp_test_helper(bxp_kwargs=dict(
        capprops=dict(linestyle='--', color='g', lw=3)))


@image_comparison(['bxp_customwhisker.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_customwhisker():
    _bxp_test_helper(bxp_kwargs=dict(
        whiskerprops=dict(linestyle='-', color='m', lw=3)))


@image_comparison(['bxp_withnotch.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_shownotches():
    _bxp_test_helper(bxp_kwargs=dict(shownotches=True))


@image_comparison(['bxp_nocaps.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_nocaps():
    _bxp_test_helper(bxp_kwargs=dict(showcaps=False))


@image_comparison(['bxp_nobox.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_nobox():
    _bxp_test_helper(bxp_kwargs=dict(showbox=False))


@image_comparison(['bxp_no_flier_stats.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_no_flier_stats():
    def transform(stats):
        for s in stats:
            s.pop('fliers', None)
        return stats

    _bxp_test_helper(transform_stats=transform,
                     bxp_kwargs=dict(showfliers=False))


@image_comparison(['bxp_withmean_point.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_showmean():
    _bxp_test_helper(bxp_kwargs=dict(showmeans=True, meanline=False))


@image_comparison(['bxp_withmean_line.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_showmeanasline():
    _bxp_test_helper(bxp_kwargs=dict(showmeans=True, meanline=True))


@image_comparison(['bxp_scalarwidth.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_scalarwidth():
    _bxp_test_helper(bxp_kwargs=dict(widths=.25))


@image_comparison(['bxp_customwidths.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_customwidths():
    _bxp_test_helper(bxp_kwargs=dict(widths=[0.10, 0.25, 0.65, 0.85]))


@image_comparison(['bxp_custompositions.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_bxp_custompositions():
    _bxp_test_helper(bxp_kwargs=dict(positions=[1, 5, 6, 7]))


def test_bxp_bad_widths():
    with pytest.raises(ValueError):
        _bxp_test_helper(bxp_kwargs=dict(widths=[1]))


def test_bxp_bad_positions():
    with pytest.raises(ValueError):
        _bxp_test_helper(bxp_kwargs=dict(positions=[2, 3]))


@image_comparison(['boxplot', 'boxplot'], tol=1.28, style='default')
def test_boxplot():
    # Randomness used for bootstrapping.
    np.random.seed(937)

    x = np.linspace(-7, 7, 140)
    x = np.hstack([-25, x, 25])
    fig, ax = plt.subplots()

    ax.boxplot([x, x], bootstrap=10000, notch=1)
    ax.set_ylim((-30, 30))

    # Reuse testcase from above for a labeled data test
    data = {"x": [x, x]}
    fig, ax = plt.subplots()
    ax.boxplot("x", bootstrap=10000, notch=1, data=data)
    ax.set_ylim((-30, 30))


@image_comparison(['boxplot_sym2.png'], remove_text=True, style='default')
def test_boxplot_sym2():
    # Randomness used for bootstrapping.
    np.random.seed(937)

    x = np.linspace(-7, 7, 140)
    x = np.hstack([-25, x, 25])
    fig, [ax1, ax2] = plt.subplots(1, 2)

    ax1.boxplot([x, x], bootstrap=10000, sym='^')
    ax1.set_ylim((-30, 30))

    ax2.boxplot([x, x], bootstrap=10000, sym='g')
    ax2.set_ylim((-30, 30))


@image_comparison(['boxplot_sym.png'],
                  remove_text=True,
                  savefig_kwarg={'dpi': 40},
                  style='default')
def test_boxplot_sym():
    x = np.linspace(-7, 7, 140)
    x = np.hstack([-25, x, 25])
    fig, ax = plt.subplots()

    ax.boxplot([x, x], sym='gs')
    ax.set_ylim((-30, 30))


@image_comparison(['boxplot_autorange_false_whiskers.png',
                   'boxplot_autorange_true_whiskers.png'],
                  style='default')
def test_boxplot_autorange_whiskers():
    # Randomness used for bootstrapping.
    np.random.seed(937)

    x = np.ones(140)
    x = np.hstack([0, x, 2])

    fig1, ax1 = plt.subplots()
    ax1.boxplot([x, x], bootstrap=10000, notch=1)
    ax1.set_ylim((-5, 5))

    fig2, ax2 = plt.subplots()
    ax2.boxplot([x, x], bootstrap=10000, notch=1, autorange=True)
    ax2.set_ylim((-5, 5))


def _rc_test_bxp_helper(ax, rc_dict):
    x = np.linspace(-7, 7, 140)
    x = np.hstack([-25, x, 25])
    with matplotlib.rc_context(rc_dict):
        ax.boxplot([x, x])
    return ax


@image_comparison(['boxplot_rc_parameters'],
                  savefig_kwarg={'dpi': 100}, remove_text=True,
                  tol=1, style='default')
def test_boxplot_rc_parameters():
    # Randomness used for bootstrapping.
    np.random.seed(937)

    fig, ax = plt.subplots(3)

    rc_axis0 = {
        'boxplot.notch': True,
        'boxplot.whiskers': [5, 95],
        'boxplot.bootstrap': 10000,

        'boxplot.flierprops.color': 'b',
        'boxplot.flierprops.marker': 'o',
        'boxplot.flierprops.markerfacecolor': 'g',
        'boxplot.flierprops.markeredgecolor': 'b',
        'boxplot.flierprops.markersize': 5,
        'boxplot.flierprops.linestyle': '--',
        'boxplot.flierprops.linewidth': 2.0,

        'boxplot.boxprops.color': 'r',
        'boxplot.boxprops.linewidth': 2.0,
        'boxplot.boxprops.linestyle': '--',

        'boxplot.capprops.color': 'c',
        'boxplot.capprops.linewidth': 2.0,
        'boxplot.capprops.linestyle': '--',

        'boxplot.medianprops.color': 'k',
        'boxplot.medianprops.linewidth': 2.0,
        'boxplot.medianprops.linestyle': '--',
    }

    rc_axis1 = {
        'boxplot.vertical': False,
        'boxplot.whiskers': [0, 100],
        'boxplot.patchartist': True,
    }

    rc_axis2 = {
        'boxplot.whiskers': 2.0,
        'boxplot.showcaps': False,
        'boxplot.showbox': False,
        'boxplot.showfliers': False,
        'boxplot.showmeans': True,
        'boxplot.meanline': True,

        'boxplot.meanprops.color': 'c',
        'boxplot.meanprops.linewidth': 2.0,
        'boxplot.meanprops.linestyle': '--',

        'boxplot.whiskerprops.color': 'r',
        'boxplot.whiskerprops.linewidth': 2.0,
        'boxplot.whiskerprops.linestyle': '-.',
    }
    dict_list = [rc_axis0, rc_axis1, rc_axis2]
    for axis, rc_axis in zip(ax, dict_list):
        _rc_test_bxp_helper(axis, rc_axis)

    assert (matplotlib.patches.PathPatch in
            [type(t) for t in ax[1].get_children()])


@image_comparison(['boxplot_with_CIarray.png'],
                  remove_text=True, savefig_kwarg={'dpi': 40}, style='default')
def test_boxplot_with_CIarray():
    # Randomness used for bootstrapping.
    np.random.seed(937)

    x = np.linspace(-7, 7, 140)
    x = np.hstack([-25, x, 25])
    fig, ax = plt.subplots()
    CIs = np.array([[-1.5, 3.], [-1., 3.5]])

    # show a boxplot with Matplotlib medians and confidence intervals, and
    # another with manual values
    ax.boxplot([x, x], bootstrap=10000, usermedians=[None, 1.0],
               conf_intervals=CIs, notch=1)
    ax.set_ylim((-30, 30))


@image_comparison(['boxplot_no_inverted_whisker.png'],
                  remove_text=True, savefig_kwarg={'dpi': 40}, style='default')
def test_boxplot_no_weird_whisker():
    x = np.array([3, 9000, 150, 88, 350, 200000, 1400, 960],
                 dtype=np.float64)
    ax1 = plt.axes()
    ax1.boxplot(x)
    ax1.set_yscale('log')
    ax1.yaxis.grid(False, which='minor')
    ax1.xaxis.grid(False)


def test_boxplot_bad_medians():
    x = np.linspace(-7, 7, 140)
    x = np.hstack([-25, x, 25])
    fig, ax = plt.subplots()
    with pytest.raises(ValueError):
        ax.boxplot(x, usermedians=[1, 2])
    with pytest.raises(ValueError):
        ax.boxplot([x, x], usermedians=[[1, 2], [1, 2]])


def test_boxplot_bad_ci():
    x = np.linspace(-7, 7, 140)
    x = np.hstack([-25, x, 25])
    fig, ax = plt.subplots()
    with pytest.raises(ValueError):
        ax.boxplot([x, x], conf_intervals=[[1, 2]])
    with pytest.raises(ValueError):
        ax.boxplot([x, x], conf_intervals=[[1, 2], [1]])


def test_boxplot_zorder():
    x = np.arange(10)
    fix, ax = plt.subplots()
    assert ax.boxplot(x)['boxes'][0].get_zorder() == 2
    assert ax.boxplot(x, zorder=10)['boxes'][0].get_zorder() == 10


def test_boxplot_marker_behavior():
    plt.rcParams['lines.marker'] = 's'
    plt.rcParams['boxplot.flierprops.marker'] = 'o'
    plt.rcParams['boxplot.meanprops.marker'] = '^'
    fig, ax = plt.subplots()
    test_data = np.arange(100)
    test_data[-1] = 150  # a flier point
    bxp_handle = ax.boxplot(test_data, showmeans=True)
    for bxp_lines in ['whiskers', 'caps', 'boxes', 'medians']:
        for each_line in bxp_handle[bxp_lines]:
            # Ensure that the rcParams['lines.marker'] is overridden by ''
            assert each_line.get_marker() == ''

    # Ensure that markers for fliers and means aren't overridden with ''
    assert bxp_handle['fliers'][0].get_marker() == 'o'
    assert bxp_handle['means'][0].get_marker() == '^'


@image_comparison(['boxplot_mod_artists_after_plotting.png'],
                  remove_text=True, savefig_kwarg={'dpi': 40}, style='default')
def test_boxplot_mod_artist_after_plotting():
    x = [0.15, 0.11, 0.06, 0.06, 0.12, 0.56, -0.56]
    fig, ax = plt.subplots()
    bp = ax.boxplot(x, sym="o")
    for key in bp:
        for obj in bp[key]:
            obj.set_color('green')


@image_comparison(['violinplot_vert_baseline.png',
                   'violinplot_vert_baseline.png'])
def test_vert_violinplot_baseline():
    # First 9 digits of frac(sqrt(2))
    np.random.seed(414213562)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax = plt.axes()
    ax.violinplot(data, positions=range(4), showmeans=0, showextrema=0,
                  showmedians=0)

    # Reuse testcase from above for a labeled data test
    data = {"d": data}
    fig, ax = plt.subplots()
    ax.violinplot("d", positions=range(4), showmeans=0, showextrema=0,
                  showmedians=0, data=data)


@image_comparison(['violinplot_vert_showmeans.png'])
def test_vert_violinplot_showmeans():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(3))
    np.random.seed(732050807)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), showmeans=1, showextrema=0,
                  showmedians=0)


@image_comparison(['violinplot_vert_showextrema.png'])
def test_vert_violinplot_showextrema():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(5))
    np.random.seed(236067977)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), showmeans=0, showextrema=1,
                  showmedians=0)


@image_comparison(['violinplot_vert_showmedians.png'])
def test_vert_violinplot_showmedians():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(7))
    np.random.seed(645751311)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), showmeans=0, showextrema=0,
                  showmedians=1)


@image_comparison(['violinplot_vert_showall.png'])
def test_vert_violinplot_showall():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(11))
    np.random.seed(316624790)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), showmeans=1, showextrema=1,
                  showmedians=1,
                  quantiles=[[0.1, 0.9], [0.2, 0.8], [0.3, 0.7], [0.4, 0.6]])


@image_comparison(['violinplot_vert_custompoints_10.png'])
def test_vert_violinplot_custompoints_10():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(13))
    np.random.seed(605551275)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), showmeans=0, showextrema=0,
                  showmedians=0, points=10)


@image_comparison(['violinplot_vert_custompoints_200.png'])
def test_vert_violinplot_custompoints_200():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(17))
    np.random.seed(123105625)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), showmeans=0, showextrema=0,
                  showmedians=0, points=200)


@image_comparison(['violinplot_horiz_baseline.png'])
def test_horiz_violinplot_baseline():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(19))
    np.random.seed(358898943)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), vert=False, showmeans=0,
                  showextrema=0, showmedians=0)


@image_comparison(['violinplot_horiz_showmedians.png'])
def test_horiz_violinplot_showmedians():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(23))
    np.random.seed(795831523)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), vert=False, showmeans=0,
                  showextrema=0, showmedians=1)


@image_comparison(['violinplot_horiz_showmeans.png'])
def test_horiz_violinplot_showmeans():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(29))
    np.random.seed(385164807)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), vert=False, showmeans=1,
                  showextrema=0, showmedians=0)


@image_comparison(['violinplot_horiz_showextrema.png'])
def test_horiz_violinplot_showextrema():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(31))
    np.random.seed(567764362)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), vert=False, showmeans=0,
                  showextrema=1, showmedians=0)


@image_comparison(['violinplot_horiz_showall.png'])
def test_horiz_violinplot_showall():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(37))
    np.random.seed(82762530)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), vert=False, showmeans=1,
                  showextrema=1, showmedians=1,
                  quantiles=[[0.1, 0.9], [0.2, 0.8], [0.3, 0.7], [0.4, 0.6]])


@image_comparison(['violinplot_horiz_custompoints_10.png'])
def test_horiz_violinplot_custompoints_10():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(41))
    np.random.seed(403124237)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), vert=False, showmeans=0,
                  showextrema=0, showmedians=0, points=10)


@image_comparison(['violinplot_horiz_custompoints_200.png'])
def test_horiz_violinplot_custompoints_200():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(43))
    np.random.seed(557438524)
    data = [np.random.normal(size=100) for _ in range(4)]
    ax.violinplot(data, positions=range(4), vert=False, showmeans=0,
                  showextrema=0, showmedians=0, points=200)


def test_violinplot_bad_positions():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(47))
    np.random.seed(855654600)
    data = [np.random.normal(size=100) for _ in range(4)]
    with pytest.raises(ValueError):
        ax.violinplot(data, positions=range(5))


def test_violinplot_bad_widths():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(53))
    np.random.seed(280109889)
    data = [np.random.normal(size=100) for _ in range(4)]
    with pytest.raises(ValueError):
        ax.violinplot(data, positions=range(4), widths=[1, 2, 3])


def test_violinplot_bad_quantiles():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(73))
    np.random.seed(544003745)
    data = [np.random.normal(size=100)]

    # Different size quantile list and plots
    with pytest.raises(ValueError):
        ax.violinplot(data, quantiles=[[0.1, 0.2], [0.5, 0.7]])


def test_violinplot_outofrange_quantiles():
    ax = plt.axes()
    # First 9 digits of frac(sqrt(79))
    np.random.seed(888194417)
    data = [np.random.normal(size=100)]

    # Quantile value above 100
    with pytest.raises(ValueError):
        ax.violinplot(data, quantiles=[[0.1, 0.2, 0.3, 1.05]])

    # Quantile value below 0
    with pytest.raises(ValueError):
        ax.violinplot(data, quantiles=[[-0.05, 0.2, 0.3, 0.75]])


@check_figures_equal(extensions=["png"])
def test_violinplot_single_list_quantiles(fig_test, fig_ref):
    # Ensures quantile list for 1D can be passed in as single list
    # First 9 digits of frac(sqrt(83))
    np.random.seed(110433579)
    data = [np.random.normal(size=100)]

    # Test image
    ax = fig_test.subplots()
    ax.violinplot(data, quantiles=[0.1, 0.3, 0.9])

    # Reference image
    ax = fig_ref.subplots()
    ax.violinplot(data, quantiles=[[0.1, 0.3, 0.9]])


@check_figures_equal(extensions=["png"])
def test_violinplot_pandas_series(fig_test, fig_ref, pd):
    np.random.seed(110433579)
    s1 = pd.Series(np.random.normal(size=7), index=[9, 8, 7, 6, 5, 4, 3])
    s2 = pd.Series(np.random.normal(size=9), index=list('ABCDEFGHI'))
    s3 = pd.Series(np.random.normal(size=11))
    fig_test.subplots().violinplot([s1, s2, s3])
    fig_ref.subplots().violinplot([s1.values, s2.values, s3.values])


def test_manage_xticks():
    _, ax = plt.subplots()
    ax.set_xlim(0, 4)
    old_xlim = ax.get_xlim()
    np.random.seed(0)
    y1 = np.random.normal(10, 3, 20)
    y2 = np.random.normal(3, 1, 20)
    ax.boxplot([y1, y2], positions=[1, 2], manage_ticks=False)
    new_xlim = ax.get_xlim()
    assert_array_equal(old_xlim, new_xlim)


def test_boxplot_not_single():
    fig, ax = plt.subplots()
    ax.boxplot(np.random.rand(100), positions=[3])
    ax.boxplot(np.random.rand(100), positions=[5])
    fig.canvas.draw()
    assert ax.get_xlim() == (2.5, 5.5)
    assert list(ax.get_xticks()) == [3, 5]
    assert [t.get_text() for t in ax.get_xticklabels()] == ["3", "5"]


def test_tick_space_size_0():
    # allow font size to be zero, which affects ticks when there is
    # no other text in the figure.
    plt.plot([0, 1], [0, 1])
    matplotlib.rcParams.update({'font.size': 0})
    b = io.BytesIO()
    plt.savefig(b, dpi=80, format='raw')


@image_comparison(['errorbar_basic', 'errorbar_mixed', 'errorbar_basic'])
def test_errorbar():
    x = np.arange(0.1, 4, 0.5)
    y = np.exp(-x)

    yerr = 0.1 + 0.2*np.sqrt(x)
    xerr = 0.1 + yerr

    # First illustrate basic pyplot interface, using defaults where possible.
    fig = plt.figure()
    ax = fig.gca()
    ax.errorbar(x, y, xerr=0.2, yerr=0.4)
    ax.set_title("Simplest errorbars, 0.2 in x, 0.4 in y")

    # Now switch to a more OO interface to exercise more features.
    fig, axs = plt.subplots(nrows=2, ncols=2, sharex=True)
    ax = axs[0, 0]
    ax.errorbar(x, y, yerr=yerr, fmt='o')
    ax.set_title('Vert. symmetric')

    # With 4 subplots, reduce the number of axis ticks to avoid crowding.
    ax.locator_params(nbins=4)

    ax = axs[0, 1]
    ax.errorbar(x, y, xerr=xerr, fmt='o', alpha=0.4)
    ax.set_title('Hor. symmetric w/ alpha')

    ax = axs[1, 0]
    ax.errorbar(x, y, yerr=[yerr, 2*yerr], xerr=[xerr, 2*xerr], fmt='--o')
    ax.set_title('H, V asymmetric')

    ax = axs[1, 1]
    ax.set_yscale('log')
    # Here we have to be careful to keep all y values positive:
    ylower = np.maximum(1e-2, y - yerr)
    yerr_lower = y - ylower

    ax.errorbar(x, y, yerr=[yerr_lower, 2*yerr], xerr=xerr,
                fmt='o', ecolor='g', capthick=2)
    ax.set_title('Mixed sym., log y')

    fig.suptitle('Variable errorbars')

    # Reuse the first testcase from above for a labeled data test
    data = {"x": x, "y": y}
    fig = plt.figure()
    ax = fig.gca()
    ax.errorbar("x", "y", xerr=0.2, yerr=0.4, data=data)
    ax.set_title("Simplest errorbars, 0.2 in x, 0.4 in y")


def test_errorbar_colorcycle():

    f, ax = plt.subplots()
    x = np.arange(10)
    y = 2*x

    e1, _, _ = ax.errorbar(x, y, c=None)
    e2, _, _ = ax.errorbar(x, 2*y, c=None)
    ln1, = ax.plot(x, 4*y)

    assert mcolors.to_rgba(e1.get_color()) == mcolors.to_rgba('C0')
    assert mcolors.to_rgba(e2.get_color()) == mcolors.to_rgba('C1')
    assert mcolors.to_rgba(ln1.get_color()) == mcolors.to_rgba('C2')


@check_figures_equal()
def test_errorbar_cycle_ecolor(fig_test, fig_ref):
    x = np.arange(0.1, 4, 0.5)
    y = [np.exp(-x+n) for n in range(4)]

    axt = fig_test.subplots()
    axr = fig_ref.subplots()

    for yi, color in zip(y, ['C0', 'C1', 'C2', 'C3']):
        axt.errorbar(x, yi, yerr=(yi * 0.25), linestyle='-',
                     marker='o', ecolor='black')
        axr.errorbar(x, yi, yerr=(yi * 0.25), linestyle='-',
                     marker='o', color=color, ecolor='black')


def test_errorbar_shape():
    fig = plt.figure()
    ax = fig.gca()

    x = np.arange(0.1, 4, 0.5)
    y = np.exp(-x)
    yerr1 = 0.1 + 0.2*np.sqrt(x)
    yerr = np.vstack((yerr1, 2*yerr1)).T
    xerr = 0.1 + yerr

    with pytest.raises(ValueError):
        ax.errorbar(x, y, yerr=yerr, fmt='o')
    with pytest.raises(ValueError):
        ax.errorbar(x, y, xerr=xerr, fmt='o')
    with pytest.raises(ValueError):
        ax.errorbar(x, y, yerr=yerr, xerr=xerr, fmt='o')


@image_comparison(['errorbar_limits'])
def test_errorbar_limits():
    x = np.arange(0.5, 5.5, 0.5)
    y = np.exp(-x)
    xerr = 0.1
    yerr = 0.2
    ls = 'dotted'

    fig, ax = plt.subplots()

    # standard error bars
    ax.errorbar(x, y, xerr=xerr, yerr=yerr, ls=ls, color='blue')

    # including upper limits
    uplims = np.zeros_like(x)
    uplims[[1, 5, 9]] = True
    ax.errorbar(x, y+0.5, xerr=xerr, yerr=yerr, uplims=uplims, ls=ls,
                color='green')

    # including lower limits
    lolims = np.zeros_like(x)
    lolims[[2, 4, 8]] = True
    ax.errorbar(x, y+1.0, xerr=xerr, yerr=yerr, lolims=lolims, ls=ls,
                color='red')

    # including upper and lower limits
    ax.errorbar(x, y+1.5, marker='o', ms=8, xerr=xerr, yerr=yerr,
                lolims=lolims, uplims=uplims, ls=ls, color='magenta')

    # including xlower and xupper limits
    xerr = 0.2
    yerr = np.full_like(x, 0.2)
    yerr[[3, 6]] = 0.3
    xlolims = lolims
    xuplims = uplims
    lolims = np.zeros_like(x)
    uplims = np.zeros_like(x)
    lolims[[6]] = True
    uplims[[3]] = True
    ax.errorbar(x, y+2.1, marker='o', ms=8, xerr=xerr, yerr=yerr,
                xlolims=xlolims, xuplims=xuplims, uplims=uplims,
                lolims=lolims, ls='none', mec='blue', capsize=0,
                color='cyan')
    ax.set_xlim((0, 5.5))
    ax.set_title('Errorbar upper and lower limits')


def test_errobar_nonefmt():
    # Check that passing 'none' as a format still plots errorbars
    x = np.arange(5)
    y = np.arange(5)

    plotline, _, barlines = plt.errorbar(x, y, xerr=1, yerr=1, fmt='none')
    assert plotline is None
    for errbar in barlines:
        assert np.all(errbar.get_color() == mcolors.to_rgba('C0'))


@check_figures_equal(extensions=['png'])
def test_errorbar_with_prop_cycle(fig_test, fig_ref):
    ax = fig_ref.subplots()
    ax.errorbar(x=[2, 4, 10], y=[0, 1, 2], yerr=0.5,
                ls='--', marker='s', mfc='k')
    ax.errorbar(x=[2, 4, 10], y=[2, 3, 4], yerr=0.5, color='tab:green',
                ls=':', marker='s', mfc='y')
    ax.errorbar(x=[2, 4, 10], y=[4, 5, 6], yerr=0.5, fmt='tab:blue',
                ls='-.', marker='o', mfc='c')
    ax.set_xlim(1, 11)

    _cycle = cycler(ls=['--', ':', '-.'], marker=['s', 's', 'o'],
                    mfc=['k', 'y', 'c'], color=['b', 'g', 'r'])
    plt.rc("axes", prop_cycle=_cycle)
    ax = fig_test.subplots()
    ax.errorbar(x=[2, 4, 10], y=[0, 1, 2], yerr=0.5)
    ax.errorbar(x=[2, 4, 10], y=[2, 3, 4], yerr=0.5, color='tab:green')
    ax.errorbar(x=[2, 4, 10], y=[4, 5, 6], yerr=0.5, fmt='tab:blue')
    ax.set_xlim(1, 11)


def test_errorbar_every_invalid():
    x = np.linspace(0, 1, 15)
    y = x * (1-x)
    yerr = y/6

    ax = plt.figure().subplots()

    with pytest.raises(ValueError, match='not a tuple of two integers'):
        ax.errorbar(x, y, yerr, errorevery=(1, 2, 3))
    with pytest.raises(ValueError, match='not a tuple of two integers'):
        ax.errorbar(x, y, yerr, errorevery=(1.3, 3))
    with pytest.raises(ValueError, match='not a valid NumPy fancy index'):
        ax.errorbar(x, y, yerr, errorevery=[False, True])
    with pytest.raises(ValueError, match='not a recognized value'):
        ax.errorbar(x, y, yerr, errorevery='foobar')


@check_figures_equal()
def test_errorbar_every(fig_test, fig_ref):
    x = np.linspace(0, 1, 15)
    y = x * (1-x)
    yerr = y/6

    ax_ref = fig_ref.subplots()
    ax_test = fig_test.subplots()

    for color, shift in zip('rgbk', [0, 0, 2, 7]):
        y += .02

        # Check errorevery using an explicit offset and step.
        ax_test.errorbar(x, y, yerr, errorevery=(shift, 4),
                         capsize=4, c=color)

        # Using manual errorbars
        # n.b. errorbar draws the main plot at z=2.1 by default
        ax_ref.plot(x, y, c=color, zorder=2.1)
        ax_ref.errorbar(x[shift::4], y[shift::4], yerr[shift::4],
                        capsize=4, c=color, fmt='none')

    # Check that markevery is propagated to line, without affecting errorbars.
    ax_test.errorbar(x, y + 0.1, yerr, markevery=(1, 4), capsize=4, fmt='o')
    ax_ref.plot(x[1::4], y[1::4] + 0.1, 'o', zorder=2.1)
    ax_ref.errorbar(x, y + 0.1, yerr, capsize=4, fmt='none')

    # Check that passing a slice to markevery/errorevery works.
    ax_test.errorbar(x, y + 0.2, yerr, errorevery=slice(2, None, 3),
                     markevery=slice(2, None, 3),
                     capsize=4, c='C0', fmt='o')
    ax_ref.plot(x[2::3], y[2::3] + 0.2, 'o', c='C0', zorder=2.1)
    ax_ref.errorbar(x[2::3], y[2::3] + 0.2, yerr[2::3],
                    capsize=4, c='C0', fmt='none')

    # Check that passing an iterable to markevery/errorevery works.
    ax_test.errorbar(x, y + 0.2, yerr, errorevery=[False, True, False] * 5,
                     markevery=[False, True, False] * 5,
                     capsize=4, c='C1', fmt='o')
    ax_ref.plot(x[1::3], y[1::3] + 0.2, 'o', c='C1', zorder=2.1)
    ax_ref.errorbar(x[1::3], y[1::3] + 0.2, yerr[1::3],
                    capsize=4, c='C1', fmt='none')


@image_comparison(['hist_stacked_stepfilled', 'hist_stacked_stepfilled'])
def test_hist_stacked_stepfilled():
    # make some data
    d1 = np.linspace(1, 3, 20)
    d2 = np.linspace(0, 10, 50)
    fig, ax = plt.subplots()
    ax.hist((d1, d2), histtype="stepfilled", stacked=True)

    # Reuse testcase from above for a labeled data test
    data = {"x": (d1, d2)}
    fig, ax = plt.subplots()
    ax.hist("x", histtype="stepfilled", stacked=True, data=data)


@image_comparison(['hist_offset'])
def test_hist_offset():
    # make some data
    d1 = np.linspace(0, 10, 50)
    d2 = np.linspace(1, 3, 20)
    fig, ax = plt.subplots()
    ax.hist(d1, bottom=5)
    ax.hist(d2, bottom=15)


@image_comparison(['hist_step.png'], remove_text=True)
def test_hist_step():
    # make some data
    d1 = np.linspace(1, 3, 20)
    fig, ax = plt.subplots()
    ax.hist(d1, histtype="step")
    ax.set_ylim(0, 10)
    ax.set_xlim(-1, 5)


@image_comparison(['hist_step_horiz.png'])
def test_hist_step_horiz():
    # make some data
    d1 = np.linspace(0, 10, 50)
    d2 = np.linspace(1, 3, 20)
    fig, ax = plt.subplots()
    ax.hist((d1, d2), histtype="step", orientation="horizontal")


@image_comparison(['hist_stacked_weights'])
def test_hist_stacked_weighted():
    # make some data
    d1 = np.linspace(0, 10, 50)
    d2 = np.linspace(1, 3, 20)
    w1 = np.linspace(0.01, 3.5, 50)
    w2 = np.linspace(0.05, 2., 20)
    fig, ax = plt.subplots()
    ax.hist((d1, d2), weights=(w1, w2), histtype="stepfilled", stacked=True)


@pytest.mark.parametrize("use_line_collection", [True, False],
                         ids=['w/ line collection', 'w/o line collection'])
@image_comparison(['stem.png'], style='mpl20', remove_text=True)
def test_stem(use_line_collection):
    x = np.linspace(0.1, 2 * np.pi, 100)

    fig, ax = plt.subplots()
    # Label is a single space to force a legend to be drawn, but to avoid any
    # text being drawn
    ax.stem(x, np.cos(x),
            linefmt='C2-.', markerfmt='k+', basefmt='C1-.', label=' ',
            use_line_collection=use_line_collection)
    ax.legend()


def test_stem_args():
    fig, ax = plt.subplots()

    x = list(range(10))
    y = list(range(10))

    # Test the call signatures
    ax.stem(y)
    ax.stem(x, y)
    ax.stem(x, y, 'r--')
    ax.stem(x, y, 'r--', basefmt='b--')


def test_stem_dates():
    fig, ax = plt.subplots(1, 1)
    xs = [dateutil.parser.parse("2013-9-28 11:00:00"),
          dateutil.parser.parse("2013-9-28 12:00:00")]
    ys = [100, 200]
    ax.stem(xs, ys, "*-")


@pytest.mark.parametrize("use_line_collection", [True, False],
                         ids=['w/ line collection', 'w/o line collection'])
@image_comparison(['stem_orientation.png'], style='mpl20', remove_text=True)
def test_stem_orientation(use_line_collection):
    x = np.linspace(0.1, 2*np.pi, 50)

    fig, ax = plt.subplots()
    ax.stem(x, np.cos(x),
            linefmt='C2-.', markerfmt='kx', basefmt='C1-.',
            use_line_collection=use_line_collection, orientation='horizontal')


@image_comparison(['hist_stacked_stepfilled_alpha'])
def test_hist_stacked_stepfilled_alpha():
    # make some data
    d1 = np.linspace(1, 3, 20)
    d2 = np.linspace(0, 10, 50)
    fig, ax = plt.subplots()
    ax.hist((d1, d2), histtype="stepfilled", stacked=True, alpha=0.5)


@image_comparison(['hist_stacked_step'])
def test_hist_stacked_step():
    # make some data
    d1 = np.linspace(1, 3, 20)
    d2 = np.linspace(0, 10, 50)
    fig, ax = plt.subplots()
    ax.hist((d1, d2), histtype="step", stacked=True)


@image_comparison(['hist_stacked_normed'])
def test_hist_stacked_density():
    # make some data
    d1 = np.linspace(1, 3, 20)
    d2 = np.linspace(0, 10, 50)
    fig, ax = plt.subplots()
    ax.hist((d1, d2), stacked=True, density=True)


@image_comparison(['hist_step_bottom.png'], remove_text=True)
def test_hist_step_bottom():
    # make some data
    d1 = np.linspace(1, 3, 20)
    fig, ax = plt.subplots()
    ax.hist(d1, bottom=np.arange(10), histtype="stepfilled")


def test_hist_stepfilled_geometry():
    bins = [0, 1, 2, 3]
    data = [0, 0, 1, 1, 1, 2]
    _, _, (polygon, ) = plt.hist(data,
                                 bins=bins,
                                 histtype='stepfilled')
    xy = [[0, 0], [0, 2], [1, 2], [1, 3], [2, 3], [2, 1], [3, 1],
          [3, 0], [2, 0], [2, 0], [1, 0], [1, 0], [0, 0]]
    assert_array_equal(polygon.get_xy(), xy)


def test_hist_step_geometry():
    bins = [0, 1, 2, 3]
    data = [0, 0, 1, 1, 1, 2]
    _, _, (polygon, ) = plt.hist(data,
                                 bins=bins,
                                 histtype='step')
    xy = [[0, 0], [0, 2], [1, 2], [1, 3], [2, 3], [2, 1], [3, 1], [3, 0]]
    assert_array_equal(polygon.get_xy(), xy)


def test_hist_stepfilled_bottom_geometry():
    bins = [0, 1, 2, 3]
    data = [0, 0, 1, 1, 1, 2]
    _, _, (polygon, ) = plt.hist(data,
                                 bins=bins,
                                 bottom=[1, 2, 1.5],
                                 histtype='stepfilled')
    xy = [[0, 1], [0, 3], [1, 3], [1, 5], [2, 5], [2, 2.5], [3, 2.5],
          [3, 1.5], [2, 1.5], [2, 2], [1, 2], [1, 1], [0, 1]]
    assert_array_equal(polygon.get_xy(), xy)


def test_hist_step_bottom_geometry():
    bins = [0, 1, 2, 3]
    data = [0, 0, 1, 1, 1, 2]
    _, _, (polygon, ) = plt.hist(data,
                                 bins=bins,
                                 bottom=[1, 2, 1.5],
                                 histtype='step')
    xy = [[0, 1], [0, 3], [1, 3], [1, 5], [2, 5], [2, 2.5], [3, 2.5], [3, 1.5]]
    assert_array_equal(polygon.get_xy(), xy)


def test_hist_stacked_stepfilled_geometry():
    bins = [0, 1, 2, 3]
    data_1 = [0, 0, 1, 1, 1, 2]
    data_2 = [0, 1, 2]
    _, _, patches = plt.hist([data_1, data_2],
                             bins=bins,
                             stacked=True,
                             histtype='stepfilled')

    assert len(patches) == 2

    polygon,  = patches[0]
    xy = [[0, 0], [0, 2], [1, 2], [1, 3], [2, 3], [2, 1], [3, 1],
          [3, 0], [2, 0], [2, 0], [1, 0], [1, 0], [0, 0]]
    assert_array_equal(polygon.get_xy(), xy)

    polygon,  = patches[1]
    xy = [[0, 2], [0, 3], [1, 3], [1, 4], [2, 4], [2, 2], [3, 2],
          [3, 1], [2, 1], [2, 3], [1, 3], [1, 2], [0, 2]]
    assert_array_equal(polygon.get_xy(), xy)


def test_hist_stacked_step_geometry():
    bins = [0, 1, 2, 3]
    data_1 = [0, 0, 1, 1, 1, 2]
    data_2 = [0, 1, 2]
    _, _, patches = plt.hist([data_1, data_2],
                             bins=bins,
                             stacked=True,
                             histtype='step')

    assert len(patches) == 2

    polygon,  = patches[0]
    xy = [[0, 0], [0, 2], [1, 2], [1, 3], [2, 3], [2, 1], [3, 1], [3, 0]]
    assert_array_equal(polygon.get_xy(), xy)

    polygon,  = patches[1]
    xy = [[0, 2], [0, 3], [1, 3], [1, 4], [2, 4], [2, 2], [3, 2], [3, 1]]
    assert_array_equal(polygon.get_xy(), xy)


def test_hist_stacked_stepfilled_bottom_geometry():
    bins = [0, 1, 2, 3]
    data_1 = [0, 0, 1, 1, 1, 2]
    data_2 = [0, 1, 2]
    _, _, patches = plt.hist([data_1, data_2],
                             bins=bins,
                             stacked=True,
                             bottom=[1, 2, 1.5],
                             histtype='stepfilled')

    assert len(patches) == 2

    polygon,  = patches[0]
    xy = [[0, 1], [0, 3], [1, 3], [1, 5], [2, 5], [2, 2.5], [3, 2.5],
          [3, 1.5], [2, 1.5], [2, 2], [1, 2], [1, 1], [0, 1]]
    assert_array_equal(polygon.get_xy(), xy)

    polygon,  = patches[1]
    xy = [[0, 3], [0, 4], [1, 4], [1, 6], [2, 6], [2, 3.5], [3, 3.5],
          [3, 2.5], [2, 2.5], [2, 5], [1, 5], [1, 3], [0, 3]]
    assert_array_equal(polygon.get_xy(), xy)


def test_hist_stacked_step_bottom_geometry():
    bins = [0, 1, 2, 3]
    data_1 = [0, 0, 1, 1, 1, 2]
    data_2 = [0, 1, 2]
    _, _, patches = plt.hist([data_1, data_2],
                             bins=bins,
                             stacked=True,
                             bottom=[1, 2, 1.5],
                             histtype='step')

    assert len(patches) == 2

    polygon,  = patches[0]
    xy = [[0, 1], [0, 3], [1, 3], [1, 5], [2, 5], [2, 2.5], [3, 2.5], [3, 1.5]]
    assert_array_equal(polygon.get_xy(), xy)

    polygon,  = patches[1]
    xy = [[0, 3], [0, 4], [1, 4], [1, 6], [2, 6], [2, 3.5], [3, 3.5], [3, 2.5]]
    assert_array_equal(polygon.get_xy(), xy)


@image_comparison(['hist_stacked_bar'])
def test_hist_stacked_bar():
    # make some data
    d = [[100, 100, 100, 100, 200, 320, 450, 80, 20, 600, 310, 800],
         [20, 23, 50, 11, 100, 420], [120, 120, 120, 140, 140, 150, 180],
         [60, 60, 60, 60, 300, 300, 5, 5, 5, 5, 10, 300],
         [555, 555, 555, 30, 30, 30, 30, 30, 100, 100, 100, 100, 30, 30],
         [30, 30, 30, 30, 400, 400, 400, 400, 400, 400, 400, 400]]
    colors = [(0.5759849696758961, 1.0, 0.0), (0.0, 1.0, 0.350624650815206),
              (0.0, 1.0, 0.6549834156005998), (0.0, 0.6569064625276622, 1.0),
              (0.28302699607823545, 0.0, 1.0), (0.6849123462299822, 0.0, 1.0)]
    labels = ['green', 'orange', ' yellow', 'magenta', 'black']
    fig, ax = plt.subplots()
    ax.hist(d, bins=10, histtype='barstacked', align='mid', color=colors,
            label=labels)
    ax.legend(loc='upper right', bbox_to_anchor=(1.0, 1.0), ncol=1)


def test_hist_barstacked_bottom_unchanged():
    b = np.array([10, 20])
    plt.hist([[0, 1], [0, 1]], 2, histtype="barstacked", bottom=b)
    assert b.tolist() == [10, 20]


def test_hist_emptydata():
    fig, ax = plt.subplots()
    ax.hist([[], range(10), range(10)], histtype="step")


def test_hist_labels():
    # test singleton labels OK
    fig, ax = plt.subplots()
    _, _, bars = ax.hist([0, 1], label=0)
    assert bars[0].get_label() == '0'
    _, _, bars = ax.hist([0, 1], label=[0])
    assert bars[0].get_label() == '0'
    _, _, bars = ax.hist([0, 1], label=None)
    assert bars[0].get_label() == '_nolegend_'
    _, _, bars = ax.hist([0, 1], label='0')
    assert bars[0].get_label() == '0'
    _, _, bars = ax.hist([0, 1], label='00')
    assert bars[0].get_label() == '00'


@image_comparison(['transparent_markers'], remove_text=True)
def test_transparent_markers():
    np.random.seed(0)
    data = np.random.random(50)

    fig, ax = plt.subplots()
    ax.plot(data, 'D', mfc='none', markersize=100)


@image_comparison(['rgba_markers'], remove_text=True)
def test_rgba_markers():
    fig, axs = plt.subplots(ncols=2)
    rcolors = [(1, 0, 0, 1), (1, 0, 0, 0.5)]
    bcolors = [(0, 0, 1, 1), (0, 0, 1, 0.5)]
    alphas = [None, 0.2]
    kw = dict(ms=100, mew=20)
    for i, alpha in enumerate(alphas):
        for j, rcolor in enumerate(rcolors):
            for k, bcolor in enumerate(bcolors):
                axs[i].plot(j+1, k+1, 'o', mfc=bcolor, mec=rcolor,
                            alpha=alpha, **kw)
                axs[i].plot(j+1, k+3, 'x', mec=rcolor, alpha=alpha, **kw)
    for ax in axs:
        ax.axis([-1, 4, 0, 5])


@image_comparison(['mollweide_grid'], remove_text=True)
def test_mollweide_grid():
    # test that both horizontal and vertical gridlines appear on the Mollweide
    # projection
    fig = plt.figure()
    ax = fig.add_subplot(projection='mollweide')
    ax.grid()


def test_mollweide_forward_inverse_closure():
    # test that the round-trip Mollweide forward->inverse transformation is an
    # approximate identity
    fig = plt.figure()
    ax = fig.add_subplot(projection='mollweide')

    # set up 1-degree grid in longitude, latitude
    lon = np.linspace(-np.pi, np.pi, 360)
    lat = np.linspace(-np.pi / 2.0, np.pi / 2.0, 180)
    lon, lat = np.meshgrid(lon, lat)
    ll = np.vstack((lon.flatten(), lat.flatten())).T

    # perform forward transform
    xy = ax.transProjection.transform(ll)

    # perform inverse transform
    ll2 = ax.transProjection.inverted().transform(xy)

    # compare
    np.testing.assert_array_almost_equal(ll, ll2, 3)


def test_mollweide_inverse_forward_closure():
    # test that the round-trip Mollweide inverse->forward transformation is an
    # approximate identity
    fig = plt.figure()
    ax = fig.add_subplot(projection='mollweide')

    # set up grid in x, y
    x = np.linspace(0, 1, 500)
    x, y = np.meshgrid(x, x)
    xy = np.vstack((x.flatten(), y.flatten())).T

    # perform inverse transform
    ll = ax.transProjection.inverted().transform(xy)

    # perform forward transform
    xy2 = ax.transProjection.transform(ll)

    # compare
    np.testing.assert_array_almost_equal(xy, xy2, 3)


@image_comparison(['test_alpha'], remove_text=True)
def test_alpha():
    np.random.seed(0)
    data = np.random.random(50)

    fig, ax = plt.subplots()

    # alpha=.5 markers, solid line
    ax.plot(data, '-D', color=[1, 0, 0], mfc=[1, 0, 0, .5],
            markersize=20, lw=10)

    # everything solid by kwarg
    ax.plot(data + 2, '-D', color=[1, 0, 0, .5], mfc=[1, 0, 0, .5],
            markersize=20, lw=10,
            alpha=1)

    # everything alpha=.5 by kwarg
    ax.plot(data + 4, '-D', color=[1, 0, 0], mfc=[1, 0, 0],
            markersize=20, lw=10,
            alpha=.5)

    # everything alpha=.5 by colors
    ax.plot(data + 6, '-D', color=[1, 0, 0, .5], mfc=[1, 0, 0, .5],
            markersize=20, lw=10)

    # alpha=.5 line, solid markers
    ax.plot(data + 8, '-D', color=[1, 0, 0, .5], mfc=[1, 0, 0],
            markersize=20, lw=10)


@image_comparison(['eventplot', 'eventplot'], remove_text=True)
def test_eventplot():
    np.random.seed(0)

    data1 = np.random.random([32, 20]).tolist()
    data2 = np.random.random([6, 20]).tolist()
    data = data1 + data2
    num_datasets = len(data)

    colors1 = [[0, 1, .7]] * len(data1)
    colors2 = [[1, 0, 0],
               [0, 1, 0],
               [0, 0, 1],
               [1, .75, 0],
               [1, 0, 1],
               [0, 1, 1]]
    colors = colors1 + colors2

    lineoffsets1 = 12 + np.arange(0, len(data1)) * .33
    lineoffsets2 = [-15, -3, 1, 1.5, 6, 10]
    lineoffsets = lineoffsets1.tolist() + lineoffsets2

    linelengths1 = [.33] * len(data1)
    linelengths2 = [5, 2, 1, 1, 3, 1.5]
    linelengths = linelengths1 + linelengths2

    fig = plt.figure()
    axobj = fig.add_subplot()
    colls = axobj.eventplot(data, colors=colors, lineoffsets=lineoffsets,
                            linelengths=linelengths)

    num_collections = len(colls)
    assert num_collections == num_datasets

    # Reuse testcase from above for a labeled data test
    data = {"pos": data, "c": colors, "lo": lineoffsets, "ll": linelengths}
    fig = plt.figure()
    axobj = fig.add_subplot()
    colls = axobj.eventplot("pos", colors="c", lineoffsets="lo",
                            linelengths="ll", data=data)
    num_collections = len(colls)
    assert num_collections == num_datasets


@image_comparison(['test_eventplot_defaults.png'], remove_text=True)
def test_eventplot_defaults():
    """
    test that eventplot produces the correct output given the default params
    (see bug #3728)
    """
    np.random.seed(0)

    data1 = np.random.random([32, 20]).tolist()
    data2 = np.random.random([6, 20]).tolist()
    data = data1 + data2

    fig = plt.figure()
    axobj = fig.add_subplot()
    axobj.eventplot(data)


@pytest.mark.parametrize(('colors'), [
    ('0.5',),  # string color with multiple characters: not OK before #8193 fix
    ('tab:orange', 'tab:pink', 'tab:cyan', 'bLacK'),  # case-insensitive
    ('red', (0, 1, 0), None, (1, 0, 1, 0.5)),  # a tricky case mixing types
])
def test_eventplot_colors(colors):
    """Test the *colors* parameter of eventplot. Inspired by issue #8193."""
    data = [[0], [1], [2], [3]]  # 4 successive events of different nature

    # Build the list of the expected colors
    expected = [c if c is not None else 'C0' for c in colors]
    # Convert the list into an array of RGBA values
    # NB: ['rgbk'] is not a valid argument for to_rgba_array, while 'rgbk' is.
    if len(expected) == 1:
        expected = expected[0]
    expected = np.broadcast_to(mcolors.to_rgba_array(expected), (len(data), 4))

    fig, ax = plt.subplots()
    if len(colors) == 1:  # tuple with a single string (like '0.5' or 'rgbk')
        colors = colors[0]
    collections = ax.eventplot(data, colors=colors)

    for coll, color in zip(collections, expected):
        assert_allclose(coll.get_color(), color)


@image_comparison(['test_eventplot_problem_kwargs.png'], remove_text=True)
def test_eventplot_problem_kwargs(recwarn):
    """
    test that 'singular' versions of LineCollection props raise an
    IgnoredKeywordWarning rather than overriding the 'plural' versions (e.g.
    to prevent 'color' from overriding 'colors', see issue #4297)
    """
    np.random.seed(0)

    data1 = np.random.random([20]).tolist()
    data2 = np.random.random([10]).tolist()
    data = [data1, data2]

    fig = plt.figure()
    axobj = fig.add_subplot()

    axobj.eventplot(data,
                    colors=['r', 'b'],
                    color=['c', 'm'],
                    linewidths=[2, 1],
                    linewidth=[1, 2],
                    linestyles=['solid', 'dashed'],
                    linestyle=['dashdot', 'dotted'])

    # check that three IgnoredKeywordWarnings were raised
    assert len(recwarn) == 3
    assert all(issubclass(wi.category, MatplotlibDeprecationWarning)
               for wi in recwarn)


def test_empty_eventplot():
    fig, ax = plt.subplots(1, 1)
    ax.eventplot([[]], colors=[(0.0, 0.0, 0.0, 0.0)])
    plt.draw()


@pytest.mark.parametrize('data', [[[]], [[], [0, 1]], [[0, 1], []]])
@pytest.mark.parametrize(
    'orientation', ['_empty', 'vertical', 'horizontal', None, 'none'])
def test_eventplot_orientation(data, orientation):
    """Introduced when fixing issue #6412."""
    opts = {} if orientation == "_empty" else {'orientation': orientation}
    fig, ax = plt.subplots(1, 1)
    with (pytest.warns(MatplotlibDeprecationWarning)
          if orientation in [None, 'none'] else nullcontext()):
        ax.eventplot(data, **opts)
    plt.draw()


@image_comparison(['marker_styles.png'], remove_text=True)
def test_marker_styles():
    fig, ax = plt.subplots()
    for y, marker in enumerate(sorted(matplotlib.markers.MarkerStyle.markers,
                                      key=lambda x: str(type(x))+str(x))):
        ax.plot((y % 2)*5 + np.arange(10)*10, np.ones(10)*10*y, linestyle='',
                marker=marker, markersize=10+y/5, label=marker)


@image_comparison(['rc_markerfill.png'])
def test_markers_fillstyle_rcparams():
    fig, ax = plt.subplots()
    x = np.arange(7)
    for idx, (style, marker) in enumerate(
            [('top', 's'), ('bottom', 'o'), ('none', '^')]):
        matplotlib.rcParams['markers.fillstyle'] = style
        ax.plot(x+idx, marker=marker)


@image_comparison(['vertex_markers.png'], remove_text=True)
def test_vertex_markers():
    data = list(range(10))
    marker_as_tuple = ((-1, -1), (1, -1), (1, 1), (-1, 1))
    marker_as_list = [(-1, -1), (1, -1), (1, 1), (-1, 1)]
    fig, ax = plt.subplots()
    ax.plot(data, linestyle='', marker=marker_as_tuple, mfc='k')
    ax.plot(data[::-1], linestyle='', marker=marker_as_list, mfc='b')
    ax.set_xlim([-1, 10])
    ax.set_ylim([-1, 10])


@image_comparison(['vline_hline_zorder', 'errorbar_zorder'],
                  tol=0 if platform.machine() == 'x86_64' else 0.02)
def test_eb_line_zorder():
    x = list(range(10))

    # First illustrate basic pyplot interface, using defaults where possible.
    fig = plt.figure()
    ax = fig.gca()
    ax.plot(x, lw=10, zorder=5)
    ax.axhline(1, color='red', lw=10, zorder=1)
    ax.axhline(5, color='green', lw=10, zorder=10)
    ax.axvline(7, color='m', lw=10, zorder=7)
    ax.axvline(2, color='k', lw=10, zorder=3)

    ax.set_title("axvline and axhline zorder test")

    # Now switch to a more OO interface to exercise more features.
    fig = plt.figure()
    ax = fig.gca()
    x = list(range(10))
    y = np.zeros(10)
    yerr = list(range(10))
    ax.errorbar(x, y, yerr=yerr, zorder=5, lw=5, color='r')
    for j in range(10):
        ax.axhline(j, lw=5, color='k', zorder=j)
        ax.axhline(-j, lw=5, color='k', zorder=j)

    ax.set_title("errorbar zorder test")


@check_figures_equal()
def test_axline(fig_test, fig_ref):
    ax = fig_test.subplots()
    ax.set(xlim=(-1, 1), ylim=(-1, 1))
    ax.axline((0, 0), (1, 1))
    ax.axline((0, 0), (1, 0), color='C1')
    ax.axline((0, 0.5), (1, 0.5), color='C2')
    # slopes
    ax.axline((-0.7, -0.5), slope=0, color='C3')
    ax.axline((1, -0.5), slope=-0.5, color='C4')
    ax.axline((-0.5, 1), slope=float('inf'), color='C5')

    ax = fig_ref.subplots()
    ax.set(xlim=(-1, 1), ylim=(-1, 1))
    ax.plot([-1, 1], [-1, 1])
    ax.axhline(0, color='C1')
    ax.axhline(0.5, color='C2')
    # slopes
    ax.axhline(-0.5, color='C3')
    ax.plot([-1, 1], [0.5, -0.5], color='C4')
    ax.axvline(-0.5, color='C5')


@check_figures_equal()
def test_axline_transaxes(fig_test, fig_ref):
    ax = fig_test.subplots()
    ax.set(xlim=(-1, 1), ylim=(-1, 1))
    ax.axline((0, 0), slope=1, transform=ax.transAxes)
    ax.axline((1, 0.5), slope=1, color='C1', transform=ax.transAxes)
    ax.axline((0.5, 0.5), slope=0, color='C2', transform=ax.transAxes)
    ax.axline((0.5, 0), (0.5, 1), color='C3', transform=ax.transAxes)

    ax = fig_ref.subplots()
    ax.set(xlim=(-1, 1), ylim=(-1, 1))
    ax.plot([-1, 1], [-1, 1])
    ax.plot([0, 1], [-1, 0], color='C1')
    ax.plot([-1, 1], [0, 0], color='C2')
    ax.plot([0, 0], [-1, 1], color='C3')


@check_figures_equal()
def test_axline_transaxes_panzoom(fig_test, fig_ref):
    # test that it is robust against pan/zoom and
    # figure resize after plotting
    ax = fig_test.subplots()
    ax.set(xlim=(-1, 1), ylim=(-1, 1))
    ax.axline((0, 0), slope=1, transform=ax.transAxes)
    ax.axline((0.5, 0.5), slope=2, color='C1', transform=ax.transAxes)
    ax.axline((0.5, 0.5), slope=0, color='C2', transform=ax.transAxes)
    ax.set(xlim=(0, 5), ylim=(0, 10))
    fig_test.set_size_inches(3, 3)

    ax = fig_ref.subplots()
    ax.set(xlim=(0, 5), ylim=(0, 10))
    fig_ref.set_size_inches(3, 3)
    ax.plot([0, 5], [0, 5])
    ax.plot([0, 5], [0, 10], color='C1')
    ax.plot([0, 5], [5, 5], color='C2')


def test_axline_args():
    """Exactly one of *xy2* and *slope* must be specified."""
    fig, ax = plt.subplots()
    with pytest.raises(TypeError):
        ax.axline((0, 0))  # missing second parameter
    with pytest.raises(TypeError):
        ax.axline((0, 0), (1, 1), slope=1)  # redundant parameters
    ax.set_xscale('log')
    with pytest.raises(TypeError):
        ax.axline((0, 0), slope=1)
    ax.set_xscale('linear')
    ax.set_yscale('log')
    with pytest.raises(TypeError):
        ax.axline((0, 0), slope=1)
    ax.set_yscale('linear')
    with pytest.raises(ValueError):
        ax.axline((0, 0), (0, 0))  # two identical points are not allowed
        plt.draw()


@image_comparison(['vlines_basic', 'vlines_with_nan', 'vlines_masked'],
                  extensions=['png'])
def test_vlines():
    # normal
    x1 = [2, 3, 4, 5, 7]
    y1 = [2, -6, 3, 8, 2]
    fig1, ax1 = plt.subplots()
    ax1.vlines(x1, 0, y1, colors='g', linewidth=5)

    # GH #7406
    x2 = [2, 3, 4, 5, 6, 7]
    y2 = [2, -6, 3, 8, np.nan, 2]
    fig2, (ax2, ax3, ax4) = plt.subplots(nrows=3, figsize=(4, 8))
    ax2.vlines(x2, 0, y2, colors='g', linewidth=5)

    x3 = [2, 3, 4, 5, 6, 7]
    y3 = [np.nan, 2, -6, 3, 8, 2]
    ax3.vlines(x3, 0, y3, colors='r', linewidth=3, linestyle='--')

    x4 = [2, 3, 4, 5, 6, 7]
    y4 = [np.nan, 2, -6, 3, 8, np.nan]
    ax4.vlines(x4, 0, y4, colors='k', linewidth=2)

    # tweak the x-axis so we can see the lines better
    for ax in [ax1, ax2, ax3, ax4]:
        ax.set_xlim(0, 10)

    # check that the y-lims are all automatically the same
    assert ax1.get_ylim() == ax2.get_ylim()
    assert ax1.get_ylim() == ax3.get_ylim()
    assert ax1.get_ylim() == ax4.get_ylim()

    fig3, ax5 = plt.subplots()
    x5 = np.ma.masked_equal([2, 4, 6, 8, 10, 12], 8)
    ymin5 = np.ma.masked_equal([0, 1, -1, 0, 2, 1], 2)
    ymax5 = np.ma.masked_equal([13, 14, 15, 16, 17, 18], 18)
    ax5.vlines(x5, ymin5, ymax5, colors='k', linewidth=2)
    ax5.set_xlim(0, 15)


def test_vlines_default():
    fig, ax = plt.subplots()
    with mpl.rc_context({'lines.color': 'red'}):
        lines = ax.vlines(0.5, 0, 1)
        assert mpl.colors.same_color(lines.get_color(), 'red')


@image_comparison(['hlines_basic', 'hlines_with_nan', 'hlines_masked'],
                  extensions=['png'])
def test_hlines():
    # normal
    y1 = [2, 3, 4, 5, 7]
    x1 = [2, -6, 3, 8, 2]
    fig1, ax1 = plt.subplots()
    ax1.hlines(y1, 0, x1, colors='g', linewidth=5)

    # GH #7406
    y2 = [2, 3, 4, 5, 6, 7]
    x2 = [2, -6, 3, 8, np.nan, 2]
    fig2, (ax2, ax3, ax4) = plt.subplots(nrows=3, figsize=(4, 8))
    ax2.hlines(y2, 0, x2, colors='g', linewidth=5)

    y3 = [2, 3, 4, 5, 6, 7]
    x3 = [np.nan, 2, -6, 3, 8, 2]
    ax3.hlines(y3, 0, x3, colors='r', linewidth=3, linestyle='--')

    y4 = [2, 3, 4, 5, 6, 7]
    x4 = [np.nan, 2, -6, 3, 8, np.nan]
    ax4.hlines(y4, 0, x4, colors='k', linewidth=2)

    # tweak the y-axis so we can see the lines better
    for ax in [ax1, ax2, ax3, ax4]:
        ax.set_ylim(0, 10)

    # check that the x-lims are all automatically the same
    assert ax1.get_xlim() == ax2.get_xlim()
    assert ax1.get_xlim() == ax3.get_xlim()
    assert ax1.get_xlim() == ax4.get_xlim()

    fig3, ax5 = plt.subplots()
    y5 = np.ma.masked_equal([2, 4, 6, 8, 10, 12], 8)
    xmin5 = np.ma.masked_equal([0, 1, -1, 0, 2, 1], 2)
    xmax5 = np.ma.masked_equal([13, 14, 15, 16, 17, 18], 18)
    ax5.hlines(y5, xmin5, xmax5, colors='k', linewidth=2)
    ax5.set_ylim(0, 15)


def test_hlines_default():
    fig, ax = plt.subplots()
    with mpl.rc_context({'lines.color': 'red'}):
        lines = ax.hlines(0.5, 0, 1)
        assert mpl.colors.same_color(lines.get_color(), 'red')


@pytest.mark.parametrize('data', [[1, 2, 3, np.nan, 5],
                                  np.ma.masked_equal([1, 2, 3, 4, 5], 4)])
@check_figures_equal(extensions=["png"])
def test_lines_with_colors(fig_test, fig_ref, data):
    test_colors = ['red', 'green', 'blue', 'purple', 'orange']
    fig_test.add_subplot(2, 1, 1).vlines(data, 0, 1,
                                         colors=test_colors, linewidth=5)
    fig_test.add_subplot(2, 1, 2).hlines(data, 0, 1,
                                         colors=test_colors, linewidth=5)

    expect_xy = [1, 2, 3, 5]
    expect_color = ['red', 'green', 'blue', 'orange']
    fig_ref.add_subplot(2, 1, 1).vlines(expect_xy, 0, 1,
                                        colors=expect_color, linewidth=5)
    fig_ref.add_subplot(2, 1, 2).hlines(expect_xy, 0, 1,
                                        colors=expect_color, linewidth=5)


@image_comparison(['step_linestyle', 'step_linestyle'], remove_text=True)
def test_step_linestyle():
    x = y = np.arange(10)

    # First illustrate basic pyplot interface, using defaults where possible.
    fig, ax_lst = plt.subplots(2, 2)
    ax_lst = ax_lst.flatten()

    ln_styles = ['-', '--', '-.', ':']

    for ax, ls in zip(ax_lst, ln_styles):
        ax.step(x, y, lw=5, linestyle=ls, where='pre')
        ax.step(x, y + 1, lw=5, linestyle=ls, where='mid')
        ax.step(x, y + 2, lw=5, linestyle=ls, where='post')
        ax.set_xlim([-1, 5])
        ax.set_ylim([-1, 7])

    # Reuse testcase from above for a labeled data test
    data = {"X": x, "Y0": y, "Y1": y+1, "Y2": y+2}
    fig, ax_lst = plt.subplots(2, 2)
    ax_lst = ax_lst.flatten()
    ln_styles = ['-', '--', '-.', ':']
    for ax, ls in zip(ax_lst, ln_styles):
        ax.step("X", "Y0", lw=5, linestyle=ls, where='pre', data=data)
        ax.step("X", "Y1", lw=5, linestyle=ls, where='mid', data=data)
        ax.step("X", "Y2", lw=5, linestyle=ls, where='post', data=data)
        ax.set_xlim([-1, 5])
        ax.set_ylim([-1, 7])


@image_comparison(['mixed_collection'], remove_text=True)
def test_mixed_collection():
    # First illustrate basic pyplot interface, using defaults where possible.
    fig, ax = plt.subplots()

    c = mpatches.Circle((8, 8), radius=4, facecolor='none', edgecolor='green')

    # PDF can optimize this one
    p1 = mpl.collections.PatchCollection([c], match_original=True)
    p1.set_offsets([[0, 0], [24, 24]])
    p1.set_linewidths([1, 5])

    # PDF can't optimize this one, because the alpha of the edge changes
    p2 = mpl.collections.PatchCollection([c], match_original=True)
    p2.set_offsets([[48, 0], [-32, -16]])
    p2.set_linewidths([1, 5])
    p2.set_edgecolors([[0, 0, 0.1, 1.0], [0, 0, 0.1, 0.5]])

    ax.patch.set_color('0.5')
    ax.add_collection(p1)
    ax.add_collection(p2)

    ax.set_xlim(0, 16)
    ax.set_ylim(0, 16)


def test_subplot_key_hash():
    ax = plt.subplot(np.int32(5), np.int64(1), 1)
    ax.twinx()
    assert ax.get_subplotspec().get_geometry() == (5, 1, 0, 0)


@image_comparison(
    ["specgram_freqs.png", "specgram_freqs_linear.png",
     "specgram_noise.png", "specgram_noise_linear.png"],
    remove_text=True, tol=0.07, style="default")
def test_specgram():
    """Test axes.specgram in default (psd) mode."""

    # use former defaults to match existing baseline image
    matplotlib.rcParams['image.interpolation'] = 'nearest'

    n = 1000
    Fs = 10.

    fstims = [[Fs/4, Fs/5, Fs/11], [Fs/4.7, Fs/5.6, Fs/11.9]]
    NFFT_freqs = int(10 * Fs / np.min(fstims))
    x = np.arange(0, n, 1/Fs)
    y_freqs = np.concatenate(
        np.sin(2 * np.pi * np.multiply.outer(fstims, x)).sum(axis=1))

    NFFT_noise = int(10 * Fs / 11)
    np.random.seed(0)
    y_noise = np.concatenate([np.random.standard_normal(n), np.random.rand(n)])

    all_sides = ["default", "onesided", "twosided"]
    for y, NFFT in [(y_freqs, NFFT_freqs), (y_noise, NFFT_noise)]:
        noverlap = NFFT // 2
        pad_to = int(2 ** np.ceil(np.log2(NFFT)))
        for ax, sides in zip(plt.figure().subplots(3), all_sides):
            ax.specgram(y, NFFT=NFFT, Fs=Fs, noverlap=noverlap,
                        pad_to=pad_to, sides=sides)
        for ax, sides in zip(plt.figure().subplots(3), all_sides):
            ax.specgram(y, NFFT=NFFT, Fs=Fs, noverlap=noverlap,
                        pad_to=pad_to, sides=sides,
                        scale="linear", norm=matplotlib.colors.LogNorm())


@image_comparison(
    ["specgram_magnitude_freqs.png", "specgram_magnitude_freqs_linear.png",
     "specgram_magnitude_noise.png", "specgram_magnitude_noise_linear.png"],
    remove_text=True, tol=0.07, style="default")
def test_specgram_magnitude():
    """Test axes.specgram in magnitude mode."""

    # use former defaults to match existing baseline image
    matplotlib.rcParams['image.interpolation'] = 'nearest'

    n = 1000
    Fs = 10.

    fstims = [[Fs/4, Fs/5, Fs/11], [Fs/4.7, Fs/5.6, Fs/11.9]]
    NFFT_freqs = int(100 * Fs / np.min(fstims))
    x = np.arange(0, n, 1/Fs)
    y = np.sin(2 * np.pi * np.multiply.outer(fstims, x)).sum(axis=1)
    y[:, -1] = 1
    y_freqs = np.hstack(y)

    NFFT_noise = int(10 * Fs / 11)
    np.random.seed(0)
    y_noise = np.concatenate([np.random.standard_normal(n), np.random.rand(n)])

    all_sides = ["default", "onesided", "twosided"]
    for y, NFFT in [(y_freqs, NFFT_freqs), (y_noise, NFFT_noise)]:
        noverlap = NFFT // 2
        pad_to = int(2 ** np.ceil(np.log2(NFFT)))
        for ax, sides in zip(plt.figure().subplots(3), all_sides):
            ax.specgram(y, NFFT=NFFT, Fs=Fs, noverlap=noverlap,
                        pad_to=pad_to, sides=sides, mode="magnitude")
        for ax, sides in zip(plt.figure().subplots(3), all_sides):
            ax.specgram(y, NFFT=NFFT, Fs=Fs, noverlap=noverlap,
                        pad_to=pad_to, sides=sides, mode="magnitude",
                        scale="linear", norm=matplotlib.colors.LogNorm())


@image_comparison(
    ["specgram_angle_freqs.png", "specgram_phase_freqs.png",
     "specgram_angle_noise.png", "specgram_phase_noise.png"],
    remove_text=True, tol=0.07, style="default")
def test_specgram_angle():
    """Test axes.specgram in angle and phase modes."""

    # use former defaults to match existing baseline image
    matplotlib.rcParams['image.interpolation'] = 'nearest'

    n = 1000
    Fs = 10.

    fstims = [[Fs/4, Fs/5, Fs/11], [Fs/4.7, Fs/5.6, Fs/11.9]]
    NFFT_freqs = int(10 * Fs / np.min(fstims))
    x = np.arange(0, n, 1/Fs)
    y = np.sin(2 * np.pi * np.multiply.outer(fstims, x)).sum(axis=1)
    y[:, -1] = 1
    y_freqs = np.hstack(y)

    NFFT_noise = int(10 * Fs / 11)
    np.random.seed(0)
    y_noise = np.concatenate([np.random.standard_normal(n), np.random.rand(n)])

    all_sides = ["default", "onesided", "twosided"]
    for y, NFFT in [(y_freqs, NFFT_freqs), (y_noise, NFFT_noise)]:
        noverlap = NFFT // 2
        pad_to = int(2 ** np.ceil(np.log2(NFFT)))
        for mode in ["angle", "phase"]:
            for ax, sides in zip(plt.figure().subplots(3), all_sides):
                ax.specgram(y, NFFT=NFFT, Fs=Fs, noverlap=noverlap,
                            pad_to=pad_to, sides=sides, mode=mode)
                with pytest.raises(ValueError):
                    ax.specgram(y, NFFT=NFFT, Fs=Fs, noverlap=noverlap,
                                pad_to=pad_to, sides=sides, mode=mode,
                                scale="dB")


def test_specgram_fs_none():
    """Test axes.specgram when Fs is None, should not throw error."""
    spec, freqs, t, im = plt.specgram(np.ones(300), Fs=None, scale='linear')
    xmin, xmax, freq0, freq1 = im.get_extent()
    assert xmin == 32 and xmax == 96


@check_figures_equal(extensions=["png"])
def test_specgram_origin_rcparam(fig_test, fig_ref):
    """Test specgram ignores image.origin rcParam and uses origin 'upper'."""
    t = np.arange(500)
    signal = np.sin(t)

    plt.rcParams["image.origin"] = 'upper'

    # Reference: First graph using default origin in imshow (upper),
    fig_ref.subplots().specgram(signal)

    # Try to overwrite the setting trying to flip the specgram
    plt.rcParams["image.origin"] = 'lower'

    # Test: origin='lower' should be ignored
    fig_test.subplots().specgram(signal)


def test_specgram_origin_kwarg():
    """Ensure passing origin as a kwarg raises a TypeError."""
    t = np.arange(500)
    signal = np.sin(t)

    with pytest.raises(TypeError):
        plt.specgram(signal, origin='lower')


@image_comparison(
    ["psd_freqs.png", "csd_freqs.png", "psd_noise.png", "csd_noise.png"],
    remove_text=True, tol=0.002)
def test_psd_csd():
    n = 10000
    Fs = 100.

    fstims = [[Fs/4, Fs/5, Fs/11], [Fs/4.7, Fs/5.6, Fs/11.9]]
    NFFT_freqs = int(1000 * Fs / np.min(fstims))
    x = np.arange(0, n, 1/Fs)
    ys_freqs = np.sin(2 * np.pi * np.multiply.outer(fstims, x)).sum(axis=1)

    NFFT_noise = int(1000 * Fs / 11)
    np.random.seed(0)
    ys_noise = [np.random.standard_normal(n), np.random.rand(n)]

    all_kwargs = [{"sides": "default"},
                  {"sides": "onesided", "return_line": False},
                  {"sides": "twosided", "return_line": True}]
    for ys, NFFT in [(ys_freqs, NFFT_freqs), (ys_noise, NFFT_noise)]:
        noverlap = NFFT // 2
        pad_to = int(2 ** np.ceil(np.log2(NFFT)))
        for ax, kwargs in zip(plt.figure().subplots(3), all_kwargs):
            ret = ax.psd(np.concatenate(ys), NFFT=NFFT, Fs=Fs,
                         noverlap=noverlap, pad_to=pad_to, **kwargs)
            assert len(ret) == 2 + kwargs.get("return_line", False)
            ax.set(xlabel="", ylabel="")
        for ax, kwargs in zip(plt.figure().subplots(3), all_kwargs):
            ret = ax.csd(*ys, NFFT=NFFT, Fs=Fs,
                         noverlap=noverlap, pad_to=pad_to, **kwargs)
            assert len(ret) == 2 + kwargs.get("return_line", False)
            ax.set(xlabel="", ylabel="")


@image_comparison(
    ["magnitude_spectrum_freqs_linear.png",
     "magnitude_spectrum_freqs_dB.png",
     "angle_spectrum_freqs.png",
     "phase_spectrum_freqs.png",
     "magnitude_spectrum_noise_linear.png",
     "magnitude_spectrum_noise_dB.png",
     "angle_spectrum_noise.png",
     "phase_spectrum_noise.png"],
    remove_text=True)
def test_spectrum():
    n = 10000
    Fs = 100.

    fstims1 = [Fs/4, Fs/5, Fs/11]
    NFFT = int(1000 * Fs / min(fstims1))
    pad_to = int(2 ** np.ceil(np.log2(NFFT)))

    x = np.arange(0, n, 1/Fs)
    y_freqs = ((np.sin(2 * np.pi * np.outer(x, fstims1)) * 10**np.arange(3))
               .sum(axis=1))
    np.random.seed(0)
    y_noise = np.hstack([np.random.standard_normal(n), np.random.rand(n)]) - .5

    all_sides = ["default", "onesided", "twosided"]
    kwargs = {"Fs": Fs, "pad_to": pad_to}
    for y in [y_freqs, y_noise]:
        for ax, sides in zip(plt.figure().subplots(3), all_sides):
            spec, freqs, line = ax.magnitude_spectrum(y, sides=sides, **kwargs)
            ax.set(xlabel="", ylabel="")
        for ax, sides in zip(plt.figure().subplots(3), all_sides):
            spec, freqs, line = ax.magnitude_spectrum(y, sides=sides, **kwargs,
                                                      scale="dB")
            ax.set(xlabel="", ylabel="")
        for ax, sides in zip(plt.figure().subplots(3), all_sides):
            spec, freqs, line = ax.angle_spectrum(y, sides=sides, **kwargs)
            ax.set(xlabel="", ylabel="")
        for ax, sides in zip(plt.figure().subplots(3), all_sides):
            spec, freqs, line = ax.phase_spectrum(y, sides=sides, **kwargs)
            ax.set(xlabel="", ylabel="")


@image_comparison(['twin_spines.png'], remove_text=True)
def test_twin_spines():

    def make_patch_spines_invisible(ax):
        ax.set_frame_on(True)
        ax.patch.set_visible(False)
        ax.spines[:].set_visible(False)

    fig = plt.figure(figsize=(4, 3))
    fig.subplots_adjust(right=0.75)

    host = fig.add_subplot()
    par1 = host.twinx()
    par2 = host.twinx()

    # Offset the right spine of par2.  The ticks and label have already been
    # placed on the right by twinx above.
    par2.spines.right.set_position(("axes", 1.2))
    # Having been created by twinx, par2 has its frame off, so the line of
    # its detached spine is invisible.  First, activate the frame but make
    # the patch and spines invisible.
    make_patch_spines_invisible(par2)
    # Second, show the right spine.
    par2.spines.right.set_visible(True)

    p1, = host.plot([0, 1, 2], [0, 1, 2], "b-")
    p2, = par1.plot([0, 1, 2], [0, 3, 2], "r-")
    p3, = par2.plot([0, 1, 2], [50, 30, 15], "g-")

    host.set_xlim(0, 2)
    host.set_ylim(0, 2)
    par1.set_ylim(0, 4)
    par2.set_ylim(1, 65)

    host.yaxis.label.set_color(p1.get_color())
    par1.yaxis.label.set_color(p2.get_color())
    par2.yaxis.label.set_color(p3.get_color())

    tkw = dict(size=4, width=1.5)
    host.tick_params(axis='y', colors=p1.get_color(), **tkw)
    par1.tick_params(axis='y', colors=p2.get_color(), **tkw)
    par2.tick_params(axis='y', colors=p3.get_color(), **tkw)
    host.tick_params(axis='x', **tkw)


@image_comparison(['twin_spines_on_top.png', 'twin_spines_on_top.png'],
                  remove_text=True)
def test_twin_spines_on_top():
    matplotlib.rcParams['axes.linewidth'] = 48.0
    matplotlib.rcParams['lines.linewidth'] = 48.0

    fig = plt.figure()
    ax1 = fig.add_subplot(1, 1, 1)

    data = np.array([[1000, 1100, 1200, 1250],
                     [310, 301, 360, 400]])

    ax2 = ax1.twinx()

    ax1.plot(data[0], data[1]/1E3, color='#BEAED4')
    ax1.fill_between(data[0], data[1]/1E3, color='#BEAED4', alpha=.8)

    ax2.plot(data[0], data[1]/1E3, color='#7FC97F')
    ax2.fill_between(data[0], data[1]/1E3, color='#7FC97F', alpha=.5)

    # Reuse testcase from above for a labeled data test
    data = {"i": data[0], "j": data[1]/1E3}
    fig = plt.figure()
    ax1 = fig.add_subplot(1, 1, 1)
    ax2 = ax1.twinx()
    ax1.plot("i", "j", color='#BEAED4', data=data)
    ax1.fill_between("i", "j", color='#BEAED4', alpha=.8, data=data)
    ax2.plot("i", "j", color='#7FC97F', data=data)
    ax2.fill_between("i", "j", color='#7FC97F', alpha=.5, data=data)


@pytest.mark.parametrize("grid_which, major_visible, minor_visible", [
    ("both", True, True),
    ("major", True, False),
    ("minor", False, True),
])
def test_rcparam_grid_minor(grid_which, major_visible, minor_visible):
    mpl.rcParams.update({"axes.grid": True, "axes.grid.which": grid_which})
    fig, ax = plt.subplots()
    fig.canvas.draw()
    assert all(tick.gridline.get_visible() == major_visible
               for tick in ax.xaxis.majorTicks)
    assert all(tick.gridline.get_visible() == minor_visible
               for tick in ax.xaxis.minorTicks)


def test_grid():
    fig, ax = plt.subplots()
    ax.grid()
    fig.canvas.draw()
    assert ax.xaxis.majorTicks[0].gridline.get_visible()
    ax.grid(visible=False)
    fig.canvas.draw()
    assert not ax.xaxis.majorTicks[0].gridline.get_visible()
    ax.grid(visible=True)
    fig.canvas.draw()
    assert ax.xaxis.majorTicks[0].gridline.get_visible()
    ax.grid()
    fig.canvas.draw()
    assert not ax.xaxis.majorTicks[0].gridline.get_visible()


def test_vline_limit():
    fig = plt.figure()
    ax = fig.gca()
    ax.axvline(0.5)
    ax.plot([-0.1, 0, 0.2, 0.1])
    assert_allclose(ax.get_ylim(), (-.1, .2))


@pytest.mark.parametrize('fv, fh, args', [[plt.axvline, plt.axhline, (1,)],
                                          [plt.axvspan, plt.axhspan, (1, 1)]])
def test_axline_minmax(fv, fh, args):
    bad_lim = matplotlib.dates.num2date(1)
    # Check vertical functions
    with pytest.raises(ValueError, match='ymin must be a single scalar value'):
        fv(*args, ymin=bad_lim, ymax=1)
    with pytest.raises(ValueError, match='ymax must be a single scalar value'):
        fv(*args, ymin=1, ymax=bad_lim)
    # Check horizontal functions
    with pytest.raises(ValueError, match='xmin must be a single scalar value'):
        fh(*args, xmin=bad_lim, xmax=1)
    with pytest.raises(ValueError, match='xmax must be a single scalar value'):
        fh(*args, xmin=1, xmax=bad_lim)


def test_empty_shared_subplots():
    # empty plots with shared axes inherit limits from populated plots
    fig, axs = plt.subplots(nrows=1, ncols=2, sharex=True, sharey=True)
    axs[0].plot([1, 2, 3], [2, 4, 6])
    x0, x1 = axs[1].get_xlim()
    y0, y1 = axs[1].get_ylim()
    assert x0 <= 1
    assert x1 >= 3
    assert y0 <= 2
    assert y1 >= 6


def test_shared_with_aspect_1():
    # allow sharing one axis
    for adjustable in ['box', 'datalim']:
        fig, axs = plt.subplots(nrows=2, sharex=True)
        axs[0].set_aspect(2, adjustable=adjustable, share=True)
        assert axs[1].get_aspect() == 2
        assert axs[1].get_adjustable() == adjustable

        fig, axs = plt.subplots(nrows=2, sharex=True)
        axs[0].set_aspect(2, adjustable=adjustable)
        assert axs[1].get_aspect() == 'auto'


def test_shared_with_aspect_2():
    # Share 2 axes only with 'box':
    fig, axs = plt.subplots(nrows=2, sharex=True, sharey=True)
    axs[0].set_aspect(2, share=True)
    axs[0].plot([1, 2], [3, 4])
    axs[1].plot([3, 4], [1, 2])
    plt.draw()  # Trigger apply_aspect().
    assert axs[0].get_xlim() == axs[1].get_xlim()
    assert axs[0].get_ylim() == axs[1].get_ylim()


def test_shared_with_aspect_3():
    # Different aspect ratios:
    for adjustable in ['box', 'datalim']:
        fig, axs = plt.subplots(nrows=2, sharey=True)
        axs[0].set_aspect(2, adjustable=adjustable)
        axs[1].set_aspect(0.5, adjustable=adjustable)
        axs[0].plot([1, 2], [3, 4])
        axs[1].plot([3, 4], [1, 2])
        plt.draw()  # Trigger apply_aspect().
        assert axs[0].get_xlim() != axs[1].get_xlim()
        assert axs[0].get_ylim() == axs[1].get_ylim()
        fig_aspect = fig.bbox_inches.height / fig.bbox_inches.width
        for ax in axs:
            p = ax.get_position()
            box_aspect = p.height / p.width
            lim_aspect = ax.viewLim.height / ax.viewLim.width
            expected = fig_aspect * box_aspect / lim_aspect
            assert round(expected, 4) == round(ax.get_aspect(), 4)


@pytest.mark.parametrize('twin', ('x', 'y'))
def test_twin_with_aspect(twin):
    fig, ax = plt.subplots()
    # test twinx or twiny
    ax_twin = getattr(ax, 'twin{}'.format(twin))()
    ax.set_aspect(5)
    ax_twin.set_aspect(2)
    assert_array_equal(ax.bbox.extents,
                       ax_twin.bbox.extents)


def test_relim_visible_only():
    x1 = (0., 10.)
    y1 = (0., 10.)
    x2 = (-10., 20.)
    y2 = (-10., 30.)

    fig = matplotlib.figure.Figure()
    ax = fig.add_subplot()
    ax.plot(x1, y1)
    assert ax.get_xlim() == x1
    assert ax.get_ylim() == y1
    line, = ax.plot(x2, y2)
    assert ax.get_xlim() == x2
    assert ax.get_ylim() == y2
    line.set_visible(False)
    assert ax.get_xlim() == x2
    assert ax.get_ylim() == y2

    ax.relim(visible_only=True)
    ax.autoscale_view()

    assert ax.get_xlim() == x1
    assert ax.get_ylim() == y1


def test_text_labelsize():
    """
    tests for issue #1172
    """
    fig = plt.figure()
    ax = fig.gca()
    ax.tick_params(labelsize='large')
    ax.tick_params(direction='out')


@image_comparison(['pie_default.png'])
def test_pie_default():
    # The slices will be ordered and plotted counter-clockwise.
    labels = 'Frogs', 'Hogs', 'Dogs', 'Logs'
    sizes = [15, 30, 45, 10]
    colors = ['yellowgreen', 'gold', 'lightskyblue', 'lightcoral']
    explode = (0, 0.1, 0, 0)  # only "explode" the 2nd slice (i.e. 'Hogs')
    fig1, ax1 = plt.subplots(figsize=(8, 6))
    ax1.pie(sizes, explode=explode, labels=labels, colors=colors,
            autopct='%1.1f%%', shadow=True, startangle=90)


@image_comparison(['pie_linewidth_0', 'pie_linewidth_0', 'pie_linewidth_0'],
                  extensions=['png'])
def test_pie_linewidth_0():
    # The slices will be ordered and plotted counter-clockwise.
    labels = 'Frogs', 'Hogs', 'Dogs', 'Logs'
    sizes = [15, 30, 45, 10]
    colors = ['yellowgreen', 'gold', 'lightskyblue', 'lightcoral']
    explode = (0, 0.1, 0, 0)  # only "explode" the 2nd slice (i.e. 'Hogs')

    plt.pie(sizes, explode=explode, labels=labels, colors=colors,
            autopct='%1.1f%%', shadow=True, startangle=90,
            wedgeprops={'linewidth': 0})
    # Set aspect ratio to be equal so that pie is drawn as a circle.
    plt.axis('equal')

    # Reuse testcase from above for a labeled data test
    data = {"l": labels, "s": sizes, "c": colors, "ex": explode}
    fig = plt.figure()
    ax = fig.gca()
    ax.pie("s", explode="ex", labels="l", colors="c",
           autopct='%1.1f%%', shadow=True, startangle=90,
           wedgeprops={'linewidth': 0}, data=data)
    ax.axis('equal')

    # And again to test the pyplot functions which should also be able to be
    # called with a data kwarg
    plt.figure()
    plt.pie("s", explode="ex", labels="l", colors="c",
            autopct='%1.1f%%', shadow=True, startangle=90,
            wedgeprops={'linewidth': 0}, data=data)
    plt.axis('equal')


@image_comparison(['pie_center_radius.png'])
def test_pie_center_radius():
    # The slices will be ordered and plotted counter-clockwise.
    labels = 'Frogs', 'Hogs', 'Dogs', 'Logs'
    sizes = [15, 30, 45, 10]
    colors = ['yellowgreen', 'gold', 'lightskyblue', 'lightcoral']
    explode = (0, 0.1, 0, 0)  # only "explode" the 2nd slice (i.e. 'Hogs')

    plt.pie(sizes, explode=explode, labels=labels, colors=colors,
            autopct='%1.1f%%', shadow=True, startangle=90,
            wedgeprops={'linewidth': 0}, center=(1, 2), radius=1.5)

    plt.annotate("Center point", xy=(1, 2), xytext=(1, 1.5),
                 arrowprops=dict(arrowstyle="->",
                                 connectionstyle="arc3"))
    # Set aspect ratio to be equal so that pie is drawn as a circle.
    plt.axis('equal')


@image_comparison(['pie_linewidth_2.png'])
def test_pie_linewidth_2():
    # The slices will be ordered and plotted counter-clockwise.
    labels = 'Frogs', 'Hogs', 'Dogs', 'Logs'
    sizes = [15, 30, 45, 10]
    colors = ['yellowgreen', 'gold', 'lightskyblue', 'lightcoral']
    explode = (0, 0.1, 0, 0)  # only "explode" the 2nd slice (i.e. 'Hogs')

    plt.pie(sizes, explode=explode, labels=labels, colors=colors,
            autopct='%1.1f%%', shadow=True, startangle=90,
            wedgeprops={'linewidth': 2})
    # Set aspect ratio to be equal so that pie is drawn as a circle.
    plt.axis('equal')


@image_comparison(['pie_ccw_true.png'])
def test_pie_ccw_true():
    # The slices will be ordered and plotted counter-clockwise.
    labels = 'Frogs', 'Hogs', 'Dogs', 'Logs'
    sizes = [15, 30, 45, 10]
    colors = ['yellowgreen', 'gold', 'lightskyblue', 'lightcoral']
    explode = (0, 0.1, 0, 0)  # only "explode" the 2nd slice (i.e. 'Hogs')

    plt.pie(sizes, explode=explode, labels=labels, colors=colors,
            autopct='%1.1f%%', shadow=True, startangle=90,
            counterclock=True)
    # Set aspect ratio to be equal so that pie is drawn as a circle.
    plt.axis('equal')


@image_comparison(['pie_frame_grid.png'])
def test_pie_frame_grid():
    # The slices will be ordered and plotted counter-clockwise.
    labels = 'Frogs', 'Hogs', 'Dogs', 'Logs'
    sizes = [15, 30, 45, 10]
    colors = ['yellowgreen', 'gold', 'lightskyblue', 'lightcoral']
    # only "explode" the 2nd slice (i.e. 'Hogs')
    explode = (0, 0.1, 0, 0)

    plt.pie(sizes, explode=explode, labels=labels, colors=colors,
            autopct='%1.1f%%', shadow=True, startangle=90,
            wedgeprops={'linewidth': 0},
            frame=True, center=(2, 2))

    plt.pie(sizes[::-1], explode=explode, labels=labels, colors=colors,
            autopct='%1.1f%%', shadow=True, startangle=90,
            wedgeprops={'linewidth': 0},
            frame=True, center=(5, 2))

    plt.pie(sizes, explode=explode[::-1], labels=labels, colors=colors,
            autopct='%1.1f%%', shadow=True, startangle=90,
            wedgeprops={'linewidth': 0},
            frame=True, center=(3, 5))
    # Set aspect ratio to be equal so that pie is drawn as a circle.
    plt.axis('equal')


@image_comparison(['pie_rotatelabels_true.png'])
def test_pie_rotatelabels_true():
    # The slices will be ordered and plotted counter-clockwise.
    labels = 'Hogwarts', 'Frogs', 'Dogs', 'Logs'
    sizes = [15, 30, 45, 10]
    colors = ['yellowgreen', 'gold', 'lightskyblue', 'lightcoral']
    explode = (0, 0.1, 0, 0)  # only "explode" the 2nd slice (i.e. 'Hogs')

    plt.pie(sizes, explode=explode, labels=labels, colors=colors,
            autopct='%1.1f%%', shadow=True, startangle=90,
            rotatelabels=True)
    # Set aspect ratio to be equal so that pie is drawn as a circle.
    plt.axis('equal')


@image_comparison(['pie_no_label.png'])
def test_pie_nolabel_but_legend():
    labels = 'Frogs', 'Hogs', 'Dogs', 'Logs'
    sizes = [15, 30, 45, 10]
    colors = ['yellowgreen', 'gold', 'lightskyblue', 'lightcoral']
    explode = (0, 0.1, 0, 0)  # only "explode" the 2nd slice (i.e. 'Hogs')
    plt.pie(sizes, explode=explode, labels=labels, colors=colors,
            autopct='%1.1f%%', shadow=True, startangle=90, labeldistance=None,
            rotatelabels=True)
    plt.axis('equal')
    plt.ylim(-1.2, 1.2)
    plt.legend()


def test_pie_textprops():
    data = [23, 34, 45]
    labels = ["Long name 1", "Long name 2", "Long name 3"]

    textprops = dict(horizontalalignment="center",
                     verticalalignment="top",
                     rotation=90,
                     rotation_mode="anchor",
                     size=12, color="red")

    _, texts, autopct = plt.gca().pie(data, labels=labels, autopct='%.2f',
                                      textprops=textprops)
    for labels in [texts, autopct]:
        for tx in labels:
            assert tx.get_ha() == textprops["horizontalalignment"]
            assert tx.get_va() == textprops["verticalalignment"]
            assert tx.get_rotation() == textprops["rotation"]
            assert tx.get_rotation_mode() == textprops["rotation_mode"]
            assert tx.get_size() == textprops["size"]
            assert tx.get_color() == textprops["color"]


def test_pie_get_negative_values():
    # Test the ValueError raised when feeding negative values into axes.pie
    fig, ax = plt.subplots()
    with pytest.raises(ValueError):
        ax.pie([5, 5, -3], explode=[0, .1, .2])


def test_normalize_kwarg_warn_pie():
    fig, ax = plt.subplots()
    with pytest.warns(MatplotlibDeprecationWarning):
        ax.pie(x=[0], normalize=None)


def test_normalize_kwarg_pie():
    fig, ax = plt.subplots()
    x = [0.3, 0.3, 0.1]
    t1 = ax.pie(x=x, normalize=True)
    assert abs(t1[0][-1].theta2 - 360.) < 1e-3
    t2 = ax.pie(x=x, normalize=False)
    assert abs(t2[0][-1].theta2 - 360.) > 1e-3


@image_comparison(['set_get_ticklabels.png'])
def test_set_get_ticklabels():
    # test issue 2246
    fig, ax = plt.subplots(2)
    ha = ['normal', 'set_x/yticklabels']

    ax[0].plot(np.arange(10))
    ax[0].set_title(ha[0])

    ax[1].plot(np.arange(10))
    ax[1].set_title(ha[1])

    # set ticklabel to 1 plot in normal way
    ax[0].set_xticks(range(10))
    ax[0].set_yticks(range(10))
    ax[0].set_xticklabels(['a', 'b', 'c', 'd'] + 6 * [''])
    ax[0].set_yticklabels(['11', '12', '13', '14'] + 6 * [''])

    # set ticklabel to the other plot, expect the 2 plots have same label
    # setting pass get_ticklabels return value as ticklabels argument
    ax[1].set_xticks(ax[0].get_xticks())
    ax[1].set_yticks(ax[0].get_yticks())
    ax[1].set_xticklabels(ax[0].get_xticklabels())
    ax[1].set_yticklabels(ax[0].get_yticklabels())


def test_subsampled_ticklabels():
    # test issue 11937
    fig, ax = plt.subplots()
    ax.plot(np.arange(10))
    ax.xaxis.set_ticks(np.arange(10) + 0.1)
    ax.locator_params(nbins=5)
    ax.xaxis.set_ticklabels([c for c in "bcdefghijk"])
    plt.draw()
    labels = [t.get_text() for t in ax.xaxis.get_ticklabels()]
    assert labels == ['b', 'd', 'f', 'h', 'j']


def test_mismatched_ticklabels():
    fig, ax = plt.subplots()
    ax.plot(np.arange(10))
    ax.xaxis.set_ticks([1.5, 2.5])
    with pytest.raises(ValueError):
        ax.xaxis.set_ticklabels(['a', 'b', 'c'])


def test_empty_ticks_fixed_loc():
    # Smoke test that [] can be used to unset all tick labels
    fig, ax = plt.subplots()
    ax.bar([1, 2], [1, 2])
    ax.set_xticks([1, 2])
    ax.set_xticklabels([])


@image_comparison(['retain_tick_visibility.png'])
def test_retain_tick_visibility():
    fig, ax = plt.subplots()
    plt.plot([0, 1, 2], [0, -1, 4])
    plt.setp(ax.get_yticklabels(), visible=False)
    ax.tick_params(axis="y", which="both", length=0)


def test_tick_label_update():
    # test issue 9397

    fig, ax = plt.subplots()

    # Set up a dummy formatter
    def formatter_func(x, pos):
        return "unit value" if x == 1 else ""
    ax.xaxis.set_major_formatter(plt.FuncFormatter(formatter_func))

    # Force some of the x-axis ticks to be outside of the drawn range
    ax.set_xticks([-1, 0, 1, 2, 3])
    ax.set_xlim(-0.5, 2.5)

    ax.figure.canvas.draw()
    tick_texts = [tick.get_text() for tick in ax.xaxis.get_ticklabels()]
    assert tick_texts == ["", "", "unit value", "", ""]


@image_comparison(['o_marker_path_snap.png'], savefig_kwarg={'dpi': 72})
def test_o_marker_path_snap():
    fig, ax = plt.subplots()
    ax.margins(.1)
    for ms in range(1, 15):
        ax.plot([1, 2, ], np.ones(2) + ms, 'o', ms=ms)

    for ms in np.linspace(1, 10, 25):
        ax.plot([3, 4, ], np.ones(2) + ms, 'o', ms=ms)


def test_margins():
    # test all ways margins can be called
    data = [1, 10]
    xmin = 0.0
    xmax = len(data) - 1.0
    ymin = min(data)
    ymax = max(data)

    fig1, ax1 = plt.subplots(1, 1)
    ax1.plot(data)
    ax1.margins(1)
    assert ax1.margins() == (1, 1)
    assert ax1.get_xlim() == (xmin - (xmax - xmin) * 1,
                              xmax + (xmax - xmin) * 1)
    assert ax1.get_ylim() == (ymin - (ymax - ymin) * 1,
                              ymax + (ymax - ymin) * 1)

    fig2, ax2 = plt.subplots(1, 1)
    ax2.plot(data)
    ax2.margins(0.5, 2)
    assert ax2.margins() == (0.5, 2)
    assert ax2.get_xlim() == (xmin - (xmax - xmin) * 0.5,
                              xmax + (xmax - xmin) * 0.5)
    assert ax2.get_ylim() == (ymin - (ymax - ymin) * 2,
                              ymax + (ymax - ymin) * 2)

    fig3, ax3 = plt.subplots(1, 1)
    ax3.plot(data)
    ax3.margins(x=-0.2, y=0.5)
    assert ax3.margins() == (-0.2, 0.5)
    assert ax3.get_xlim() == (xmin - (xmax - xmin) * -0.2,
                              xmax + (xmax - xmin) * -0.2)
    assert ax3.get_ylim() == (ymin - (ymax - ymin) * 0.5,
                              ymax + (ymax - ymin) * 0.5)


def test_set_margin_updates_limits():
    mpl.style.use("default")
    fig, ax = plt.subplots()
    ax.plot([1, 2], [1, 2])
    ax.set(xscale="log", xmargin=0)
    assert ax.get_xlim() == (1, 2)


def test_length_one_hist():
    fig, ax = plt.subplots()
    ax.hist(1)
    ax.hist([1])


def test_pathological_hexbin():
    # issue #2863
    mylist = [10] * 100
    fig, ax = plt.subplots(1, 1)
    ax.hexbin(mylist, mylist)
    fig.savefig(io.BytesIO())  # Check that no warning is emitted.


def test_color_None():
    # issue 3855
    fig, ax = plt.subplots()
    ax.plot([1, 2], [1, 2], color=None)


def test_color_alias():
    # issues 4157 and 4162
    fig, ax = plt.subplots()
    line = ax.plot([0, 1], c='lime')[0]
    assert 'lime' == line.get_color()


def test_numerical_hist_label():
    fig, ax = plt.subplots()
    ax.hist([range(15)] * 5, label=range(5))
    ax.legend()


def test_unicode_hist_label():
    fig, ax = plt.subplots()
    a = (b'\xe5\xbe\x88\xe6\xbc\x82\xe4\xba\xae, ' +
         b'r\xc3\xb6m\xc3\xa4n ch\xc3\xa4r\xc3\xa1ct\xc3\xa8rs')
    b = b'\xd7\xa9\xd7\x9c\xd7\x95\xd7\x9d'
    labels = [a.decode('utf-8'),
              'hi aardvark',
              b.decode('utf-8'),
              ]

    ax.hist([range(15)] * 3, label=labels)
    ax.legend()


def test_move_offsetlabel():
    data = np.random.random(10) * 1e-22

    fig, ax = plt.subplots()
    ax.plot(data)
    fig.canvas.draw()
    before = ax.yaxis.offsetText.get_position()
    assert ax.yaxis.offsetText.get_horizontalalignment() == 'left'
    ax.yaxis.tick_right()
    fig.canvas.draw()
    after = ax.yaxis.offsetText.get_position()
    assert after[0] > before[0] and after[1] == before[1]
    assert ax.yaxis.offsetText.get_horizontalalignment() == 'right'

    fig, ax = plt.subplots()
    ax.plot(data)
    fig.canvas.draw()
    before = ax.xaxis.offsetText.get_position()
    assert ax.xaxis.offsetText.get_verticalalignment() == 'top'
    ax.xaxis.tick_top()
    fig.canvas.draw()
    after = ax.xaxis.offsetText.get_position()
    assert after[0] == before[0] and after[1] > before[1]
    assert ax.xaxis.offsetText.get_verticalalignment() == 'bottom'


@image_comparison(['rc_spines.png'], savefig_kwarg={'dpi': 40})
def test_rc_spines():
    rc_dict = {
        'axes.spines.left': False,
        'axes.spines.right': False,
        'axes.spines.top': False,
        'axes.spines.bottom': False}
    with matplotlib.rc_context(rc_dict):
        plt.subplots()  # create a figure and axes with the spine properties


@image_comparison(['rc_grid.png'], savefig_kwarg={'dpi': 40})
def test_rc_grid():
    fig = plt.figure()
    rc_dict0 = {
        'axes.grid': True,
        'axes.grid.axis': 'both'
    }
    rc_dict1 = {
        'axes.grid': True,
        'axes.grid.axis': 'x'
    }
    rc_dict2 = {
        'axes.grid': True,
        'axes.grid.axis': 'y'
    }
    dict_list = [rc_dict0, rc_dict1, rc_dict2]

    for i, rc_dict in enumerate(dict_list, 1):
        with matplotlib.rc_context(rc_dict):
            fig.add_subplot(3, 1, i)


def test_rc_tick():
    d = {'xtick.bottom': False, 'xtick.top': True,
         'ytick.left': True, 'ytick.right': False}
    with plt.rc_context(rc=d):
        fig = plt.figure()
        ax1 = fig.add_subplot(1, 1, 1)
        xax = ax1.xaxis
        yax = ax1.yaxis
        # tick1On bottom/left
        assert not xax._major_tick_kw['tick1On']
        assert xax._major_tick_kw['tick2On']
        assert not xax._minor_tick_kw['tick1On']
        assert xax._minor_tick_kw['tick2On']

        assert yax._major_tick_kw['tick1On']
        assert not yax._major_tick_kw['tick2On']
        assert yax._minor_tick_kw['tick1On']
        assert not yax._minor_tick_kw['tick2On']


def test_rc_major_minor_tick():
    d = {'xtick.top': True, 'ytick.right': True,  # Enable all ticks
         'xtick.bottom': True, 'ytick.left': True,
         # Selectively disable
         'xtick.minor.bottom': False, 'xtick.major.bottom': False,
         'ytick.major.left': False, 'ytick.minor.left': False}
    with plt.rc_context(rc=d):
        fig = plt.figure()
        ax1 = fig.add_subplot(1, 1, 1)
        xax = ax1.xaxis
        yax = ax1.yaxis
        # tick1On bottom/left
        assert not xax._major_tick_kw['tick1On']
        assert xax._major_tick_kw['tick2On']
        assert not xax._minor_tick_kw['tick1On']
        assert xax._minor_tick_kw['tick2On']

        assert not yax._major_tick_kw['tick1On']
        assert yax._major_tick_kw['tick2On']
        assert not yax._minor_tick_kw['tick1On']
        assert yax._minor_tick_kw['tick2On']


def test_square_plot():
    x = np.arange(4)
    y = np.array([1., 3., 5., 7.])
    fig, ax = plt.subplots()
    ax.plot(x, y, 'mo')
    ax.axis('square')
    xlim, ylim = ax.get_xlim(), ax.get_ylim()
    assert np.diff(xlim) == np.diff(ylim)
    assert ax.get_aspect() == 1
    assert_array_almost_equal(
        ax.get_position(original=True).extents, (0.125, 0.1, 0.9, 0.9))
    assert_array_almost_equal(
        ax.get_position(original=False).extents, (0.2125, 0.1, 0.8125, 0.9))


def test_bad_plot_args():
    with pytest.raises(ValueError):
        plt.plot(None)
    with pytest.raises(ValueError):
        plt.plot(None, None)
    with pytest.raises(ValueError):
        plt.plot(np.zeros((2, 2)), np.zeros((2, 3)))
    with pytest.raises(ValueError):
        plt.plot((np.arange(5).reshape((1, -1)), np.arange(5).reshape(-1, 1)))


@pytest.mark.parametrize(
    "xy, cls", [
        ((), mpl.image.AxesImage),  # (0, N)
        (((3, 7), (2, 6)), mpl.image.AxesImage),  # (xmin, xmax)
        ((range(5), range(4)), mpl.image.AxesImage),  # regular grid
        (([1, 2, 4, 8, 16], [0, 1, 2, 3]),  # irregular grid
         mpl.image.PcolorImage),
        ((np.random.random((4, 5)), np.random.random((4, 5))),  # 2D coords
         mpl.collections.QuadMesh),
    ]
)
@pytest.mark.parametrize(
    "data", [np.arange(12).reshape((3, 4)), np.random.rand(3, 4, 3)]
)
def test_pcolorfast(xy, data, cls):
    fig, ax = plt.subplots()
    assert type(ax.pcolorfast(*xy, data)) == cls


def test_shared_scale():
    fig, axs = plt.subplots(2, 2, sharex=True, sharey=True)

    axs[0, 0].set_xscale("log")
    axs[0, 0].set_yscale("log")

    for ax in axs.flat:
        assert ax.get_yscale() == 'log'
        assert ax.get_xscale() == 'log'

    axs[1, 1].set_xscale("linear")
    axs[1, 1].set_yscale("linear")

    for ax in axs.flat:
        assert ax.get_yscale() == 'linear'
        assert ax.get_xscale() == 'linear'


def test_shared_bool():
    with pytest.raises(TypeError):
        plt.subplot(sharex=True)
    with pytest.raises(TypeError):
        plt.subplot(sharey=True)


def test_violin_point_mass():
    """Violin plot should handle point mass pdf gracefully."""
    plt.violinplot(np.array([0, 0]))


def generate_errorbar_inputs():
    base_xy = cycler('x', [np.arange(5)]) + cycler('y', [np.ones(5)])
    err_cycler = cycler('err', [1,
                                [1, 1, 1, 1, 1],
                                [[1, 1, 1, 1, 1],
                                 [1, 1, 1, 1, 1]],
                                np.ones(5),
                                np.ones((2, 5)),
                                None
                                ])
    xerr_cy = cycler('xerr', err_cycler)
    yerr_cy = cycler('yerr', err_cycler)

    empty = ((cycler('x', [[]]) + cycler('y', [[]])) *
             cycler('xerr', [[], None]) * cycler('yerr', [[], None]))
    xerr_only = base_xy * xerr_cy
    yerr_only = base_xy * yerr_cy
    both_err = base_xy * yerr_cy * xerr_cy

    return [*xerr_only, *yerr_only, *both_err, *empty]


@pytest.mark.parametrize('kwargs', generate_errorbar_inputs())
def test_errorbar_inputs_shotgun(kwargs):
    ax = plt.gca()
    eb = ax.errorbar(**kwargs)
    eb.remove()


@image_comparison(["dash_offset"], remove_text=True)
def test_dash_offset():
    fig, ax = plt.subplots()
    x = np.linspace(0, 10)
    y = np.ones_like(x)
    for j in range(0, 100, 2):
        ax.plot(x, j*y, ls=(j, (10, 10)), lw=5, color='k')


def test_title_pad():
    # check that title padding puts the title in the right
    # place...
    fig, ax = plt.subplots()
    ax.set_title('aardvark', pad=30.)
    m = ax.titleOffsetTrans.get_matrix()
    assert m[1, -1] == (30. / 72. * fig.dpi)
    ax.set_title('aardvark', pad=0.)
    m = ax.titleOffsetTrans.get_matrix()
    assert m[1, -1] == 0.
    # check that it is reverted...
    ax.set_title('aardvark', pad=None)
    m = ax.titleOffsetTrans.get_matrix()
    assert m[1, -1] == (matplotlib.rcParams['axes.titlepad'] / 72. * fig.dpi)


def test_title_location_roundtrip():
    fig, ax = plt.subplots()
    # set default title location
    plt.rcParams['axes.titlelocation'] = 'center'
    ax.set_title('aardvark')
    ax.set_title('left', loc='left')
    ax.set_title('right', loc='right')

    assert 'left' == ax.get_title(loc='left')
    assert 'right' == ax.get_title(loc='right')
    assert 'aardvark' == ax.get_title(loc='center')

    with pytest.raises(ValueError):
        ax.get_title(loc='foo')
    with pytest.raises(ValueError):
        ax.set_title('fail', loc='foo')


@image_comparison(["loglog.png"], remove_text=True, tol=0.02)
def test_loglog():
    fig, ax = plt.subplots()
    x = np.arange(1, 11)
    ax.loglog(x, x**3, lw=5)
    ax.tick_params(length=25, width=2)
    ax.tick_params(length=15, width=2, which='minor')


@pytest.mark.parametrize("new_api", [False, True])
@image_comparison(["test_loglog_nonpos.png"], remove_text=True, style='mpl20')
def test_loglog_nonpos(new_api):
    fig, axs = plt.subplots(3, 3)
    x = np.arange(1, 11)
    y = x**3
    y[7] = -3.
    x[4] = -10
    for (i, j), ax in np.ndenumerate(axs):
        mcx = ['mask', 'clip', ''][j]
        mcy = ['mask', 'clip', ''][i]
        if new_api:
            if mcx == mcy:
                if mcx:
                    ax.loglog(x, y**3, lw=2, nonpositive=mcx)
                else:
                    ax.loglog(x, y**3, lw=2)
            else:
                ax.loglog(x, y**3, lw=2)
                if mcx:
                    ax.set_xscale("log", nonpositive=mcx)
                if mcy:
                    ax.set_yscale("log", nonpositive=mcy)
        else:
            kws = {}
            if mcx:
                kws['nonposx'] = mcx
            if mcy:
                kws['nonposy'] = mcy
            with (pytest.warns(MatplotlibDeprecationWarning) if kws
                  else nullcontext()):
                ax.loglog(x, y**3, lw=2, **kws)


@pytest.mark.style('default')
def test_axes_margins():
    fig, ax = plt.subplots()
    ax.plot([0, 1, 2, 3])
    assert ax.get_ybound()[0] != 0

    fig, ax = plt.subplots()
    ax.bar([0, 1, 2, 3], [1, 1, 1, 1])
    assert ax.get_ybound()[0] == 0

    fig, ax = plt.subplots()
    ax.barh([0, 1, 2, 3], [1, 1, 1, 1])
    assert ax.get_xbound()[0] == 0

    fig, ax = plt.subplots()
    ax.pcolor(np.zeros((10, 10)))
    assert ax.get_xbound() == (0, 10)
    assert ax.get_ybound() == (0, 10)

    fig, ax = plt.subplots()
    ax.pcolorfast(np.zeros((10, 10)))
    assert ax.get_xbound() == (0, 10)
    assert ax.get_ybound() == (0, 10)

    fig, ax = plt.subplots()
    ax.hist(np.arange(10))
    assert ax.get_ybound()[0] == 0

    fig, ax = plt.subplots()
    ax.imshow(np.zeros((10, 10)))
    assert ax.get_xbound() == (-0.5, 9.5)
    assert ax.get_ybound() == (-0.5, 9.5)


@pytest.fixture(params=['x', 'y'])
def shared_axis_remover(request):
    def _helper_x(ax):
        ax2 = ax.twinx()
        ax2.remove()
        ax.set_xlim(0, 15)
        r = ax.xaxis.get_major_locator()()
        assert r[-1] > 14

    def _helper_y(ax):
        ax2 = ax.twiny()
        ax2.remove()
        ax.set_ylim(0, 15)
        r = ax.yaxis.get_major_locator()()
        assert r[-1] > 14

    return {"x": _helper_x, "y": _helper_y}[request.param]


@pytest.fixture(params=['gca', 'subplots', 'subplots_shared', 'add_axes'])
def shared_axes_generator(request):
    # test all of the ways to get fig/ax sets
    if request.param == 'gca':
        fig = plt.figure()
        ax = fig.gca()
    elif request.param == 'subplots':
        fig, ax = plt.subplots()
    elif request.param == 'subplots_shared':
        fig, ax_lst = plt.subplots(2, 2, sharex='all', sharey='all')
        ax = ax_lst[0][0]
    elif request.param == 'add_axes':
        fig = plt.figure()
        ax = fig.add_axes([.1, .1, .8, .8])
    return fig, ax


def test_remove_shared_axes(shared_axes_generator, shared_axis_remover):
    # test all of the ways to get fig/ax sets
    fig, ax = shared_axes_generator
    shared_axis_remover(ax)


def test_remove_shared_axes_relim():
    fig, ax_lst = plt.subplots(2, 2, sharex='all', sharey='all')
    ax = ax_lst[0][0]
    orig_xlim = ax_lst[0][1].get_xlim()
    ax.remove()
    ax.set_xlim(0, 5)
    assert_array_equal(ax_lst[0][1].get_xlim(), orig_xlim)


def test_shared_axes_autoscale():
    l = np.arange(-80, 90, 40)
    t = np.random.random_sample((l.size, l.size))

    fig, (ax1, ax2) = plt.subplots(2, 1, sharex=True, sharey=True)

    ax1.set_xlim(-1000, 1000)
    ax1.set_ylim(-1000, 1000)
    ax1.contour(l, l, t)

    ax2.contour(l, l, t)
    assert not ax1.get_autoscalex_on() and not ax2.get_autoscalex_on()
    assert not ax1.get_autoscaley_on() and not ax2.get_autoscaley_on()
    assert ax1.get_xlim() == ax2.get_xlim() == (-1000, 1000)
    assert ax1.get_ylim() == ax2.get_ylim() == (-1000, 1000)


def test_adjust_numtick_aspect():
    fig, ax = plt.subplots()
    ax.yaxis.get_major_locator().set_params(nbins='auto')
    ax.set_xlim(0, 1000)
    ax.set_aspect('equal')
    fig.canvas.draw()
    assert len(ax.yaxis.get_major_locator()()) == 2
    ax.set_ylim(0, 1000)
    fig.canvas.draw()
    assert len(ax.yaxis.get_major_locator()()) > 2


@image_comparison(["auto_numticks.png"], style='default')
def test_auto_numticks():
    # Make tiny, empty subplots, verify that there are only 3 ticks.
    plt.subplots(4, 4)


@image_comparison(["auto_numticks_log.png"], style='default')
def test_auto_numticks_log():
    # Verify that there are not too many ticks with a large log range.
    fig, ax = plt.subplots()
    matplotlib.rcParams['axes.autolimit_mode'] = 'round_numbers'
    ax.loglog([1e-20, 1e5], [1e-16, 10])


def test_broken_barh_empty():
    fig, ax = plt.subplots()
    ax.broken_barh([], (.1, .5))


def test_broken_barh_timedelta():
    """Check that timedelta works as x, dx pair for this method."""
    fig, ax = plt.subplots()
    d0 = datetime.datetime(2018, 11, 9, 0, 0, 0)
    pp = ax.broken_barh([(d0, datetime.timedelta(hours=1))], [1, 2])
    assert pp.get_paths()[0].vertices[0, 0] == mdates.date2num(d0)
    assert pp.get_paths()[0].vertices[2, 0] == mdates.date2num(d0) + 1 / 24


def test_pandas_pcolormesh(pd):
    time = pd.date_range('2000-01-01', periods=10)
    depth = np.arange(20)
    data = np.random.rand(19, 9)

    fig, ax = plt.subplots()
    ax.pcolormesh(time, depth, data)


def test_pandas_indexing_dates(pd):
    dates = np.arange('2005-02', '2005-03', dtype='datetime64[D]')
    values = np.sin(np.array(range(len(dates))))
    df = pd.DataFrame({'dates': dates, 'values': values})

    ax = plt.gca()

    without_zero_index = df[np.array(df.index) % 2 == 1].copy()
    ax.plot('dates', 'values', data=without_zero_index)


def test_pandas_errorbar_indexing(pd):
    df = pd.DataFrame(np.random.uniform(size=(5, 4)),
                      columns=['x', 'y', 'xe', 'ye'],
                      index=[1, 2, 3, 4, 5])
    fig, ax = plt.subplots()
    ax.errorbar('x', 'y', xerr='xe', yerr='ye', data=df)


def test_pandas_index_shape(pd):
    df = pd.DataFrame({"XX": [4, 5, 6], "YY": [7, 1, 2]})
    fig, ax = plt.subplots()
    ax.plot(df.index, df['YY'])


def test_pandas_indexing_hist(pd):
    ser_1 = pd.Series(data=[1, 2, 2, 3, 3, 4, 4, 4, 4, 5])
    ser_2 = ser_1.iloc[1:]
    fig, ax = plt.subplots()
    ax.hist(ser_2)


def test_pandas_bar_align_center(pd):
    # Tests fix for issue 8767
    df = pd.DataFrame({'a': range(2), 'b': range(2)})

    fig, ax = plt.subplots(1)

    ax.bar(df.loc[df['a'] == 1, 'b'],
           df.loc[df['a'] == 1, 'b'],
           align='center')

    fig.canvas.draw()


def test_axis_set_tick_params_labelsize_labelcolor():
    # Tests fix for issue 4346
    axis_1 = plt.subplot()
    axis_1.yaxis.set_tick_params(labelsize=30, labelcolor='red',
                                 direction='out')

    # Expected values after setting the ticks
    assert axis_1.yaxis.majorTicks[0]._size == 4.0
    assert axis_1.yaxis.majorTicks[0].tick1line.get_color() == 'k'
    assert axis_1.yaxis.majorTicks[0].label1.get_size() == 30.0
    assert axis_1.yaxis.majorTicks[0].label1.get_color() == 'red'


def test_axes_tick_params_gridlines():
    # Now treating grid params like other Tick params
    ax = plt.subplot()
    ax.tick_params(grid_color='b', grid_linewidth=5, grid_alpha=0.5,
                   grid_linestyle='dashdot')
    for axis in ax.xaxis, ax.yaxis:
        assert axis.majorTicks[0].gridline.get_color() == 'b'
        assert axis.majorTicks[0].gridline.get_linewidth() == 5
        assert axis.majorTicks[0].gridline.get_alpha() == 0.5
        assert axis.majorTicks[0].gridline.get_linestyle() == '-.'


def test_axes_tick_params_ylabelside():
    # Tests fix for issue 10267
    ax = plt.subplot()
    ax.tick_params(labelleft=False, labelright=True,
                   which='major')
    ax.tick_params(labelleft=False, labelright=True,
                   which='minor')
    # expects left false, right true
    assert ax.yaxis.majorTicks[0].label1.get_visible() is False
    assert ax.yaxis.majorTicks[0].label2.get_visible() is True
    assert ax.yaxis.minorTicks[0].label1.get_visible() is False
    assert ax.yaxis.minorTicks[0].label2.get_visible() is True


def test_axes_tick_params_xlabelside():
    # Tests fix for issue 10267
    ax = plt.subplot()
    ax.tick_params(labeltop=True, labelbottom=False,
                   which='major')
    ax.tick_params(labeltop=True, labelbottom=False,
                   which='minor')
    # expects top True, bottom False
    # label1.get_visible() mapped to labelbottom
    # label2.get_visible() mapped to labeltop
    assert ax.xaxis.majorTicks[0].label1.get_visible() is False
    assert ax.xaxis.majorTicks[0].label2.get_visible() is True
    assert ax.xaxis.minorTicks[0].label1.get_visible() is False
    assert ax.xaxis.minorTicks[0].label2.get_visible() is True


def test_none_kwargs():
    ax = plt.figure().subplots()
    ln, = ax.plot(range(32), linestyle=None)
    assert ln.get_linestyle() == '-'


def test_bar_uint8():
    xs = [0, 1, 2, 3]
    b = plt.bar(np.array(xs, dtype=np.uint8), [2, 3, 4, 5], align="edge")
    for (patch, x) in zip(b.patches, xs):
        assert patch.xy[0] == x


@image_comparison(['date_timezone_x.png'], tol=1.0)
def test_date_timezone_x():
    # Tests issue 5575
    time_index = [datetime.datetime(2016, 2, 22, hour=x,
                                    tzinfo=dateutil.tz.gettz('Canada/Eastern'))
                  for x in range(3)]

    # Same Timezone
    plt.figure(figsize=(20, 12))
    plt.subplot(2, 1, 1)
    plt.plot_date(time_index, [3] * 3, tz='Canada/Eastern')

    # Different Timezone
    plt.subplot(2, 1, 2)
    plt.plot_date(time_index, [3] * 3, tz='UTC')


@image_comparison(['date_timezone_y.png'])
def test_date_timezone_y():
    # Tests issue 5575
    time_index = [datetime.datetime(2016, 2, 22, hour=x,
                                    tzinfo=dateutil.tz.gettz('Canada/Eastern'))
                  for x in range(3)]

    # Same Timezone
    plt.figure(figsize=(20, 12))
    plt.subplot(2, 1, 1)
    plt.plot_date([3] * 3,
                  time_index, tz='Canada/Eastern', xdate=False, ydate=True)

    # Different Timezone
    plt.subplot(2, 1, 2)
    plt.plot_date([3] * 3, time_index, tz='UTC', xdate=False, ydate=True)


@image_comparison(['date_timezone_x_and_y.png'], tol=1.0)
def test_date_timezone_x_and_y():
    # Tests issue 5575
    UTC = datetime.timezone.utc
    time_index = [datetime.datetime(2016, 2, 22, hour=x, tzinfo=UTC)
                  for x in range(3)]

    # Same Timezone
    plt.figure(figsize=(20, 12))
    plt.subplot(2, 1, 1)
    plt.plot_date(time_index, time_index, tz='UTC', ydate=True)

    # Different Timezone
    plt.subplot(2, 1, 2)
    plt.plot_date(time_index, time_index, tz='US/Eastern', ydate=True)


@image_comparison(['axisbelow.png'], remove_text=True)
def test_axisbelow():
    # Test 'line' setting added in 6287.
    # Show only grids, not frame or ticks, to make this test
    # independent of future change to drawing order of those elements.
    axs = plt.figure().subplots(ncols=3, sharex=True, sharey=True)
    settings = (False, 'line', True)

    for ax, setting in zip(axs, settings):
        ax.plot((0, 10), (0, 10), lw=10, color='m')
        circ = mpatches.Circle((3, 3), color='r')
        ax.add_patch(circ)
        ax.grid(color='c', linestyle='-', linewidth=3)
        ax.tick_params(top=False, bottom=False,
                       left=False, right=False)
        ax.spines[:].set_visible(False)
        ax.set_axisbelow(setting)


def test_titletwiny():
    plt.style.use('mpl20')
    fig, ax = plt.subplots(dpi=72)
    ax2 = ax.twiny()
    xlabel2 = ax2.set_xlabel('Xlabel2')
    title = ax.set_title('Title')
    fig.canvas.draw()
    renderer = fig.canvas.get_renderer()
    # ------- Test that title is put above Xlabel2 (Xlabel2 at top) ----------
    bbox_y0_title = title.get_window_extent(renderer).y0  # bottom of title
    bbox_y1_xlabel2 = xlabel2.get_window_extent(renderer).y1  # top of xlabel2
    y_diff = bbox_y0_title - bbox_y1_xlabel2
    assert np.isclose(y_diff, 3)


def test_titlesetpos():
    # Test that title stays put if we set it manually
    fig, ax = plt.subplots()
    fig.subplots_adjust(top=0.8)
    ax2 = ax.twiny()
    ax.set_xlabel('Xlabel')
    ax2.set_xlabel('Xlabel2')
    ax.set_title('Title')
    pos = (0.5, 1.11)
    ax.title.set_position(pos)
    renderer = fig.canvas.get_renderer()
    ax._update_title_position(renderer)
    assert ax.title.get_position() == pos


def test_title_xticks_top():
    # Test that title moves if xticks on top of axes.
    mpl.rcParams['axes.titley'] = None
    fig, ax = plt.subplots()
    ax.xaxis.set_ticks_position('top')
    ax.set_title('xlabel top')
    fig.canvas.draw()
    assert ax.title.get_position()[1] > 1.04


def test_title_xticks_top_both():
    # Test that title moves if xticks on top of axes.
    mpl.rcParams['axes.titley'] = None
    fig, ax = plt.subplots()
    ax.tick_params(axis="x",
                   bottom=True, top=True, labelbottom=True, labeltop=True)
    ax.set_title('xlabel top')
    fig.canvas.draw()
    assert ax.title.get_position()[1] > 1.04


def test_title_no_move_off_page():
    # If an axes is off the figure (ie. if it is cropped during a save)
    # make sure that the automatic title repositioning does not get done.
    mpl.rcParams['axes.titley'] = None
    fig = plt.figure()
    ax = fig.add_axes([0.1, -0.5, 0.8, 0.2])
    ax.tick_params(axis="x",
                   bottom=True, top=True, labelbottom=True, labeltop=True)
    tt = ax.set_title('Boo')
    fig.canvas.draw()
    assert tt.get_position()[1] == 1.0


def test_offset_label_color():
    # Tests issue 6440
    fig, ax = plt.subplots()
    ax.plot([1.01e9, 1.02e9, 1.03e9])
    ax.yaxis.set_tick_params(labelcolor='red')
    assert ax.yaxis.get_offset_text().get_color() == 'red'


def test_offset_text_visible():
    fig, ax = plt.subplots()
    ax.plot([1.01e9, 1.02e9, 1.03e9])
    ax.yaxis.set_tick_params(label1On=False, label2On=True)
    assert ax.yaxis.get_offset_text().get_visible()
    ax.yaxis.set_tick_params(label2On=False)
    assert not ax.yaxis.get_offset_text().get_visible()


def test_large_offset():
    fig, ax = plt.subplots()
    ax.plot((1 + np.array([0, 1.e-12])) * 1.e27)
    fig.canvas.draw()


def test_barb_units():
    fig, ax = plt.subplots()
    dates = [datetime.datetime(2017, 7, 15, 18, i) for i in range(0, 60, 10)]
    y = np.linspace(0, 5, len(dates))
    u = v = np.linspace(0, 50, len(dates))
    ax.barbs(dates, y, u, v)


def test_quiver_units():
    fig, ax = plt.subplots()
    dates = [datetime.datetime(2017, 7, 15, 18, i) for i in range(0, 60, 10)]
    y = np.linspace(0, 5, len(dates))
    u = v = np.linspace(0, 50, len(dates))
    ax.quiver(dates, y, u, v)


def test_bar_color_cycle():
    to_rgb = mcolors.to_rgb
    fig, ax = plt.subplots()
    for j in range(5):
        ln, = ax.plot(range(3))
        brs = ax.bar(range(3), range(3))
        for br in brs:
            assert to_rgb(ln.get_color()) == to_rgb(br.get_facecolor())


def test_tick_param_label_rotation():
    fix, (ax, ax2) = plt.subplots(1, 2)
    ax.plot([0, 1], [0, 1])
    ax2.plot([0, 1], [0, 1])
    ax.xaxis.set_tick_params(which='both', rotation=75)
    ax.yaxis.set_tick_params(which='both', rotation=90)
    for text in ax.get_xticklabels(which='both'):
        assert text.get_rotation() == 75
    for text in ax.get_yticklabels(which='both'):
        assert text.get_rotation() == 90

    ax2.tick_params(axis='x', labelrotation=53)
    ax2.tick_params(axis='y', rotation=35)
    for text in ax2.get_xticklabels(which='major'):
        assert text.get_rotation() == 53
    for text in ax2.get_yticklabels(which='major'):
        assert text.get_rotation() == 35


@pytest.mark.style('default')
def test_fillbetween_cycle():
    fig, ax = plt.subplots()

    for j in range(3):
        cc = ax.fill_between(range(3), range(3))
        target = mcolors.to_rgba('C{}'.format(j))
        assert tuple(cc.get_facecolors().squeeze()) == tuple(target)

    for j in range(3, 6):
        cc = ax.fill_betweenx(range(3), range(3))
        target = mcolors.to_rgba('C{}'.format(j))
        assert tuple(cc.get_facecolors().squeeze()) == tuple(target)

    target = mcolors.to_rgba('k')

    for al in ['facecolor', 'facecolors', 'color']:
        cc = ax.fill_between(range(3), range(3), **{al: 'k'})
        assert tuple(cc.get_facecolors().squeeze()) == tuple(target)

    edge_target = mcolors.to_rgba('k')
    for j, el in enumerate(['edgecolor', 'edgecolors'], start=6):
        cc = ax.fill_between(range(3), range(3), **{el: 'k'})
        face_target = mcolors.to_rgba('C{}'.format(j))
        assert tuple(cc.get_facecolors().squeeze()) == tuple(face_target)
        assert tuple(cc.get_edgecolors().squeeze()) == tuple(edge_target)


def test_log_margins():
    plt.rcParams['axes.autolimit_mode'] = 'data'
    fig, ax = plt.subplots()
    margin = 0.05
    ax.set_xmargin(margin)
    ax.semilogx([10, 100], [10, 100])
    xlim0, xlim1 = ax.get_xlim()
    transform = ax.xaxis.get_transform()
    xlim0t, xlim1t = transform.transform([xlim0, xlim1])
    x0t, x1t = transform.transform([10, 100])
    delta = (x1t - x0t) * margin
    assert_allclose([xlim0t + delta, xlim1t - delta], [x0t, x1t])


def test_color_length_mismatch():
    N = 5
    x, y = np.arange(N), np.arange(N)
    colors = np.arange(N+1)
    fig, ax = plt.subplots()
    with pytest.raises(ValueError):
        ax.scatter(x, y, c=colors)
    c_rgb = (0.5, 0.5, 0.5)
    ax.scatter(x, y, c=c_rgb)
    ax.scatter(x, y, c=[c_rgb] * N)


def test_eventplot_legend():
    plt.eventplot([1.0], label='Label')
    plt.legend()


def test_bar_broadcast_args():
    fig, ax = plt.subplots()
    # Check that a bar chart with a single height for all bars works.
    ax.bar(range(4), 1)
    # Check that a horizontal chart with one width works.
    ax.barh(0, 1, left=range(4), height=1)
    # Check that edgecolor gets broadcast.
    rect1, rect2 = ax.bar([0, 1], [0, 1], edgecolor=(.1, .2, .3, .4))
    assert rect1.get_edgecolor() == rect2.get_edgecolor() == (.1, .2, .3, .4)


def test_invalid_axis_limits():
    plt.plot([0, 1], [0, 1])
    with pytest.raises(ValueError):
        plt.xlim(np.nan)
    with pytest.raises(ValueError):
        plt.xlim(np.inf)
    with pytest.raises(ValueError):
        plt.ylim(np.nan)
    with pytest.raises(ValueError):
        plt.ylim(np.inf)


# Test all 4 combinations of logs/symlogs for minorticks_on()
@pytest.mark.parametrize('xscale', ['symlog', 'log'])
@pytest.mark.parametrize('yscale', ['symlog', 'log'])
def test_minorticks_on(xscale, yscale):
    ax = plt.subplot()
    ax.plot([1, 2, 3, 4])
    ax.set_xscale(xscale)
    ax.set_yscale(yscale)
    ax.minorticks_on()


def test_twinx_knows_limits():
    fig, ax = plt.subplots()

    ax.axvspan(1, 2)
    xtwin = ax.twinx()
    xtwin.plot([0, 0.5], [1, 2])
    # control axis
    fig2, ax2 = plt.subplots()

    ax2.axvspan(1, 2)
    ax2.plot([0, 0.5], [1, 2])

    assert_array_equal(xtwin.viewLim.intervalx, ax2.viewLim.intervalx)


def test_zero_linewidth():
    # Check that setting a zero linewidth doesn't error
    plt.plot([0, 1], [0, 1], ls='--', lw=0)


def test_empty_errorbar_legend():
    fig, ax = plt.subplots()
    ax.errorbar([], [], xerr=[], label='empty y')
    ax.errorbar([], [], yerr=[], label='empty x')
    ax.legend()


@check_figures_equal(extensions=["png"])
def test_plot_decimal(fig_test, fig_ref):
    x0 = np.arange(-10, 10, 0.3)
    y0 = [5.2 * x ** 3 - 2.1 * x ** 2 + 7.34 * x + 4.5 for x in x0]
    x = [Decimal(i) for i in x0]
    y = [Decimal(i) for i in y0]
    # Test image - line plot with Decimal input
    fig_test.subplots().plot(x, y)
    # Reference image
    fig_ref.subplots().plot(x0, y0)


# pdf and svg tests fail using travis' old versions of gs and inkscape.
@check_figures_equal(extensions=["png"])
def test_markerfacecolor_none_alpha(fig_test, fig_ref):
    fig_test.subplots().plot(0, "o", mfc="none", alpha=.5)
    fig_ref.subplots().plot(0, "o", mfc="w", alpha=.5)


def test_tick_padding_tightbbox():
    """Test that tick padding gets turned off if axis is off"""
    plt.rcParams["xtick.direction"] = "out"
    plt.rcParams["ytick.direction"] = "out"
    fig, ax = plt.subplots()
    bb = ax.get_tightbbox(fig.canvas.get_renderer())
    ax.axis('off')
    bb2 = ax.get_tightbbox(fig.canvas.get_renderer())
    assert bb.x0 < bb2.x0
    assert bb.y0 < bb2.y0


def test_inset():
    """
    Ensure that inset_ax argument is indeed optional
    """
    dx, dy = 0.05, 0.05
    # generate 2 2d grids for the x & y bounds
    y, x = np.mgrid[slice(1, 5 + dy, dy),
                    slice(1, 5 + dx, dx)]
    z = np.sin(x) ** 10 + np.cos(10 + y * x) * np.cos(x)

    fig, ax = plt.subplots()
    ax.pcolormesh(x, y, z[:-1, :-1])
    ax.set_aspect(1.)
    ax.apply_aspect()
    # we need to apply_aspect to make the drawing below work.

    xlim = [1.5, 2.15]
    ylim = [2, 2.5]

    rect = [xlim[0], ylim[0], xlim[1] - xlim[0], ylim[1] - ylim[0]]

    rec, connectors = ax.indicate_inset(bounds=rect)
    assert connectors is None
    fig.canvas.draw()
    xx = np.array([[1.5, 2.],
                   [2.15, 2.5]])
    assert np.all(rec.get_bbox().get_points() == xx)


def test_zoom_inset():
    dx, dy = 0.05, 0.05
    # generate 2 2d grids for the x & y bounds
    y, x = np.mgrid[slice(1, 5 + dy, dy),
                    slice(1, 5 + dx, dx)]
    z = np.sin(x)**10 + np.cos(10 + y*x) * np.cos(x)

    fig, ax = plt.subplots()
    ax.pcolormesh(x, y, z[:-1, :-1])
    ax.set_aspect(1.)
    ax.apply_aspect()
    # we need to apply_aspect to make the drawing below work.

    # Make the inset_axes...  Position axes coordinates...
    axin1 = ax.inset_axes([0.7, 0.7, 0.35, 0.35])
    # redraw the data in the inset axes...
    axin1.pcolormesh(x, y, z[:-1, :-1])
    axin1.set_xlim([1.5, 2.15])
    axin1.set_ylim([2, 2.5])
    axin1.set_aspect(ax.get_aspect())

    rec, connectors = ax.indicate_inset_zoom(axin1)
    assert len(connectors) == 4
    fig.canvas.draw()
    xx = np.array([[1.5,  2.],
                   [2.15, 2.5]])
    assert(np.all(rec.get_bbox().get_points() == xx))
    xx = np.array([[0.6325, 0.692308],
                   [0.8425, 0.907692]])
    np.testing.assert_allclose(
        axin1.get_position().get_points(), xx, rtol=1e-4)


@pytest.mark.parametrize('x_inverted', [False, True])
@pytest.mark.parametrize('y_inverted', [False, True])
def test_indicate_inset_inverted(x_inverted, y_inverted):
    """
    Test that the inset lines are correctly located with inverted data axes.
    """
    fig, (ax1, ax2) = plt.subplots(1, 2)

    x = np.arange(10)
    ax1.plot(x, x, 'o')
    if x_inverted:
        ax1.invert_xaxis()
    if y_inverted:
        ax1.invert_yaxis()

    rect, bounds = ax1.indicate_inset([2, 2, 5, 4], ax2)
    lower_left, upper_left, lower_right, upper_right = bounds

    sign_x = -1 if x_inverted else 1
    sign_y = -1 if y_inverted else 1
    assert sign_x * (lower_right.xy2[0] - lower_left.xy2[0]) > 0
    assert sign_x * (upper_right.xy2[0] - upper_left.xy2[0]) > 0
    assert sign_y * (upper_left.xy2[1] - lower_left.xy2[1]) > 0
    assert sign_y * (upper_right.xy2[1] - lower_right.xy2[1]) > 0


def test_set_position():
    fig, ax = plt.subplots()
    ax.set_aspect(3.)
    ax.set_position([0.1, 0.1, 0.4, 0.4], which='both')
    assert np.allclose(ax.get_position().width, 0.1)
    ax.set_aspect(2.)
    ax.set_position([0.1, 0.1, 0.4, 0.4], which='original')
    assert np.allclose(ax.get_position().width, 0.15)
    ax.set_aspect(3.)
    ax.set_position([0.1, 0.1, 0.4, 0.4], which='active')
    assert np.allclose(ax.get_position().width, 0.1)


def test_spines_properbbox_after_zoom():
    fig, ax = plt.subplots()
    bb = ax.spines.bottom.get_window_extent(fig.canvas.get_renderer())
    # this is what zoom calls:
    ax._set_view_from_bbox((320, 320, 500, 500), 'in',
                           None, False, False)
    bb2 = ax.spines.bottom.get_window_extent(fig.canvas.get_renderer())
    np.testing.assert_allclose(bb.get_points(), bb2.get_points(), rtol=1e-6)


def test_cartopy_backcompat():

    class Dummy(matplotlib.axes.Axes):
        ...

    class DummySubplot(matplotlib.axes.SubplotBase, Dummy):
        _axes_class = Dummy

    matplotlib.axes._subplots._subplot_classes[Dummy] = DummySubplot

    FactoryDummySubplot = matplotlib.axes.subplot_class_factory(Dummy)

    assert DummySubplot is FactoryDummySubplot


def test_gettightbbox_ignore_nan():
    fig, ax = plt.subplots()
    remove_ticks_and_titles(fig)
    ax.text(np.NaN, 1, 'Boo')
    renderer = fig.canvas.get_renderer()
    np.testing.assert_allclose(ax.get_tightbbox(renderer).width, 496)


def test_scatter_series_non_zero_index(pd):
    # create non-zero index
    ids = range(10, 18)
    x = pd.Series(np.random.uniform(size=8), index=ids)
    y = pd.Series(np.random.uniform(size=8), index=ids)
    c = pd.Series([1, 1, 1, 1, 1, 0, 0, 0], index=ids)
    plt.scatter(x, y, c)


def test_scatter_empty_data():
    # making sure this does not raise an exception
    plt.scatter([], [])
    plt.scatter([], [], s=[], c=[])


@image_comparison(['annotate_across_transforms.png'],
                  style='mpl20', remove_text=True)
def test_annotate_across_transforms():
    x = np.linspace(0, 10, 200)
    y = np.exp(-x) * np.sin(x)

    fig, ax = plt.subplots(figsize=(3.39, 3))
    ax.plot(x, y)
    axins = ax.inset_axes([0.4, 0.5, 0.3, 0.3])
    axins.set_aspect(0.2)
    axins.xaxis.set_visible(False)
    axins.yaxis.set_visible(False)
    ax.annotate("", xy=(x[150], y[150]), xycoords=ax.transData,
                xytext=(1, 0), textcoords=axins.transAxes,
                arrowprops=dict(arrowstyle="->"))


@image_comparison(['secondary_xy.png'], style='mpl20')
def test_secondary_xy():
    fig, axs = plt.subplots(1, 2, figsize=(10, 5), constrained_layout=True)

    def invert(x):
        with np.errstate(divide='ignore'):
            return 1 / x

    for nn, ax in enumerate(axs):
        ax.plot(np.arange(2, 11), np.arange(2, 11))
        if nn == 0:
            secax = ax.secondary_xaxis
        else:
            secax = ax.secondary_yaxis

        secax(0.2, functions=(invert, invert))
        secax(0.4, functions=(lambda x: 2 * x, lambda x: x / 2))
        secax(0.6, functions=(lambda x: x**2, lambda x: x**(1/2)))
        secax(0.8)


def test_secondary_fail():
    fig, ax = plt.subplots()
    ax.plot(np.arange(2, 11), np.arange(2, 11))
    with pytest.raises(ValueError):
        ax.secondary_xaxis(0.2, functions=(lambda x: 1 / x))
    with pytest.raises(ValueError):
        ax.secondary_xaxis('right')
    with pytest.raises(ValueError):
        ax.secondary_yaxis('bottom')


def test_secondary_resize():
    fig, ax = plt.subplots(figsize=(10, 5))
    ax.plot(np.arange(2, 11), np.arange(2, 11))

    def invert(x):
        with np.errstate(divide='ignore'):
            return 1 / x

    ax.secondary_xaxis('top', functions=(invert, invert))
    fig.canvas.draw()
    fig.set_size_inches((7, 4))
    assert_allclose(ax.get_position().extents, [0.125, 0.1, 0.9, 0.9])


def test_secondary_minorloc():
    fig, ax = plt.subplots(figsize=(10, 5))
    ax.plot(np.arange(2, 11), np.arange(2, 11))

    def invert(x):
        with np.errstate(divide='ignore'):
            return 1 / x

    secax = ax.secondary_xaxis('top', functions=(invert, invert))
    assert isinstance(secax._axis.get_minor_locator(),
                      mticker.NullLocator)
    secax.minorticks_on()
    assert isinstance(secax._axis.get_minor_locator(),
                      mticker.AutoMinorLocator)
    ax.set_xscale('log')
    plt.draw()
    assert isinstance(secax._axis.get_minor_locator(),
                      mticker.LogLocator)
    ax.set_xscale('linear')
    plt.draw()
    assert isinstance(secax._axis.get_minor_locator(),
                      mticker.NullLocator)


def test_secondary_formatter():
    fig, ax = plt.subplots()
    ax.set_xscale("log")
    secax = ax.secondary_xaxis("top")
    secax.xaxis.set_major_formatter(mticker.ScalarFormatter())
    fig.canvas.draw()
    assert isinstance(
        secax.xaxis.get_major_formatter(), mticker.ScalarFormatter)


def color_boxes(fig, axs):
    """
    Helper for the tests below that test the extents of various axes elements
    """
    fig.canvas.draw()

    renderer = fig.canvas.get_renderer()
    bbaxis = []
    for nn, axx in enumerate([axs.xaxis, axs.yaxis]):
        bb = axx.get_tightbbox(renderer)
        if bb:
            axisr = plt.Rectangle(
                (bb.x0, bb.y0), width=bb.width, height=bb.height,
                linewidth=0.7, edgecolor='y', facecolor="none", transform=None,
                zorder=3)
            fig.add_artist(axisr)
        bbaxis += [bb]

    bbspines = []
    for nn, a in enumerate(['bottom', 'top', 'left', 'right']):
        bb = axs.spines[a].get_window_extent(renderer)
        spiner = plt.Rectangle(
            (bb.x0, bb.y0), width=bb.width, height=bb.height,
            linewidth=0.7, edgecolor="green", facecolor="none", transform=None,
            zorder=3)
        fig.add_artist(spiner)
        bbspines += [bb]

    bb = axs.get_window_extent()
    rect2 = plt.Rectangle(
        (bb.x0, bb.y0), width=bb.width, height=bb.height,
        linewidth=1.5, edgecolor="magenta", facecolor="none", transform=None,
        zorder=2)
    fig.add_artist(rect2)
    bbax = bb

    bb2 = axs.get_tightbbox(renderer)
    rect2 = plt.Rectangle(
        (bb2.x0, bb2.y0), width=bb2.width, height=bb2.height,
        linewidth=3, edgecolor="red", facecolor="none", transform=None,
        zorder=1)
    fig.add_artist(rect2)
    bbtb = bb2
    return bbaxis, bbspines, bbax, bbtb


def test_normal_axes():
    with rc_context({'_internal.classic_mode': False}):
        fig, ax = plt.subplots(dpi=200, figsize=(6, 6))
        fig.canvas.draw()
        plt.close(fig)
        bbaxis, bbspines, bbax, bbtb = color_boxes(fig, ax)

    # test the axis bboxes
    target = [
        [123.375, 75.88888888888886, 983.25, 33.0],
        [85.51388888888889, 99.99999999999997, 53.375, 993.0]
    ]
    for nn, b in enumerate(bbaxis):
        targetbb = mtransforms.Bbox.from_bounds(*target[nn])
        assert_array_almost_equal(b.bounds, targetbb.bounds, decimal=2)

    target = [
        [150.0, 119.999, 930.0, 11.111],
        [150.0, 1080.0, 930.0, 0.0],
        [150.0, 119.9999, 11.111, 960.0],
        [1068.8888, 119.9999, 11.111, 960.0]
    ]
    for nn, b in enumerate(bbspines):
        targetbb = mtransforms.Bbox.from_bounds(*target[nn])
        assert_array_almost_equal(b.bounds, targetbb.bounds, decimal=2)

    target = [150.0, 119.99999999999997, 930.0, 960.0]
    targetbb = mtransforms.Bbox.from_bounds(*target)
    assert_array_almost_equal(bbax.bounds, targetbb.bounds, decimal=2)

    target = [85.5138, 75.88888, 1021.11, 1017.11]
    targetbb = mtransforms.Bbox.from_bounds(*target)
    assert_array_almost_equal(bbtb.bounds, targetbb.bounds, decimal=2)

    # test that get_position roundtrips to get_window_extent
    axbb = ax.get_position().transformed(fig.transFigure).bounds
    assert_array_almost_equal(axbb, ax.get_window_extent().bounds, decimal=2)


def test_nodecorator():
    with rc_context({'_internal.classic_mode': False}):
        fig, ax = plt.subplots(dpi=200, figsize=(6, 6))
        fig.canvas.draw()
        ax.set(xticklabels=[], yticklabels=[])
        bbaxis, bbspines, bbax, bbtb = color_boxes(fig, ax)

    # test the axis bboxes
    for nn, b in enumerate(bbaxis):
        assert b is None

    target = [
        [150.0, 119.999, 930.0, 11.111],
        [150.0, 1080.0, 930.0, 0.0],
        [150.0, 119.9999, 11.111, 960.0],
        [1068.8888, 119.9999, 11.111, 960.0]
    ]
    for nn, b in enumerate(bbspines):
        targetbb = mtransforms.Bbox.from_bounds(*target[nn])
        assert_allclose(b.bounds, targetbb.bounds, atol=1e-2)

    target = [150.0, 119.99999999999997, 930.0, 960.0]
    targetbb = mtransforms.Bbox.from_bounds(*target)
    assert_allclose(bbax.bounds, targetbb.bounds, atol=1e-2)

    target = [150., 120., 930., 960.]
    targetbb = mtransforms.Bbox.from_bounds(*target)
    assert_allclose(bbtb.bounds, targetbb.bounds, atol=1e-2)


def test_displaced_spine():
    with rc_context({'_internal.classic_mode': False}):
        fig, ax = plt.subplots(dpi=200, figsize=(6, 6))
        ax.set(xticklabels=[], yticklabels=[])
        ax.spines.bottom.set_position(('axes', -0.1))
        fig.canvas.draw()
        bbaxis, bbspines, bbax, bbtb = color_boxes(fig, ax)

    targets = [
        [150., 24., 930., 11.111111],
        [150.0, 1080.0, 930.0, 0.0],
        [150.0, 119.9999, 11.111, 960.0],
        [1068.8888, 119.9999, 11.111, 960.0]
    ]
    for target, bbspine in zip(targets, bbspines):
        targetbb = mtransforms.Bbox.from_bounds(*target)
        assert_allclose(bbspine.bounds, targetbb.bounds, atol=1e-2)

    target = [150.0, 119.99999999999997, 930.0, 960.0]
    targetbb = mtransforms.Bbox.from_bounds(*target)
    assert_allclose(bbax.bounds, targetbb.bounds, atol=1e-2)

    target = [150., 24., 930., 1056.]
    targetbb = mtransforms.Bbox.from_bounds(*target)
    assert_allclose(bbtb.bounds, targetbb.bounds, atol=1e-2)


def test_tickdirs():
    """
    Switch the tickdirs and make sure the bboxes switch with them
    """
    targets = [[[150.0, 120.0, 930.0, 11.1111],
                [150.0, 120.0, 11.111, 960.0]],
               [[150.0, 108.8889, 930.0, 11.111111111111114],
                [138.889, 120, 11.111, 960.0]],
               [[150.0, 114.44444444444441, 930.0, 11.111111111111114],
                [144.44444444444446, 119.999, 11.111, 960.0]]]
    for dnum, dirs in enumerate(['in', 'out', 'inout']):
        with rc_context({'_internal.classic_mode': False}):
            fig, ax = plt.subplots(dpi=200, figsize=(6, 6))
            ax.tick_params(direction=dirs)
            fig.canvas.draw()
            bbaxis, bbspines, bbax, bbtb = color_boxes(fig, ax)
            for nn, num in enumerate([0, 2]):
                targetbb = mtransforms.Bbox.from_bounds(*targets[dnum][nn])
                assert_allclose(
                    bbspines[num].bounds, targetbb.bounds, atol=1e-2)


def test_minor_accountedfor():
    with rc_context({'_internal.classic_mode': False}):
        fig, ax = plt.subplots(dpi=200, figsize=(6, 6))
        fig.canvas.draw()
        ax.tick_params(which='both', direction='out')

        bbaxis, bbspines, bbax, bbtb = color_boxes(fig, ax)
        bbaxis, bbspines, bbax, bbtb = color_boxes(fig, ax)
        targets = [[150.0, 108.88888888888886, 930.0, 11.111111111111114],
                   [138.8889, 119.9999, 11.1111, 960.0]]
        for n in range(2):
            targetbb = mtransforms.Bbox.from_bounds(*targets[n])
            assert_allclose(
                bbspines[n * 2].bounds, targetbb.bounds, atol=1e-2)

        fig, ax = plt.subplots(dpi=200, figsize=(6, 6))
        fig.canvas.draw()
        ax.tick_params(which='both', direction='out')
        ax.minorticks_on()
        ax.tick_params(axis='both', which='minor', length=30)
        fig.canvas.draw()
        bbaxis, bbspines, bbax, bbtb = color_boxes(fig, ax)
        targets = [[150.0, 36.66666666666663, 930.0, 83.33333333333334],
                   [66.6667, 120.0, 83.3333, 960.0]]

        for n in range(2):
            targetbb = mtransforms.Bbox.from_bounds(*targets[n])
            assert_allclose(
                bbspines[n * 2].bounds, targetbb.bounds, atol=1e-2)


@check_figures_equal(extensions=["png"])
def test_axis_bool_arguments(fig_test, fig_ref):
    # Test if False and "off" give the same
    fig_test.add_subplot(211).axis(False)
    fig_ref.add_subplot(211).axis("off")
    # Test if True after False gives the same as "on"
    ax = fig_test.add_subplot(212)
    ax.axis(False)
    ax.axis(True)
    fig_ref.add_subplot(212).axis("on")


def test_axis_extent_arg():
    fig, ax = plt.subplots()
    xmin = 5
    xmax = 10
    ymin = 15
    ymax = 20
    extent = ax.axis([xmin, xmax, ymin, ymax])

    # test that the docstring is correct
    assert tuple(extent) == (xmin, xmax, ymin, ymax)

    # test that limits were set per the docstring
    assert (xmin, xmax) == ax.get_xlim()
    assert (ymin, ymax) == ax.get_ylim()


def test_datetime_masked():
    # make sure that all-masked data falls back to the viewlim
    # set in convert.axisinfo....
    x = np.array([datetime.datetime(2017, 1, n) for n in range(1, 6)])
    y = np.array([1, 2, 3, 4, 5])
    m = np.ma.masked_greater(y, 0)

    fig, ax = plt.subplots()
    ax.plot(x, m)
    dt = mdates.date2num(np.datetime64('0000-12-31'))
    assert ax.get_xlim() == (730120.0 + dt, 733773.0 + dt)


def test_hist_auto_bins():
    _, bins, _ = plt.hist([[1, 2, 3], [3, 4, 5, 6]], bins='auto')
    assert bins[0] <= 1
    assert bins[-1] >= 6


def test_hist_nan_data():
    fig, (ax1, ax2) = plt.subplots(2)

    data = [1, 2, 3]
    nan_data = data + [np.nan]

    bins, edges, _ = ax1.hist(data)
    with np.errstate(invalid='ignore'):
        nanbins, nanedges, _ = ax2.hist(nan_data)

    np.testing.assert_allclose(bins, nanbins)
    np.testing.assert_allclose(edges, nanedges)


def test_hist_range_and_density():
    _, bins, _ = plt.hist(np.random.rand(10), "auto",
                          range=(0, 1), density=True)
    assert bins[0] == 0
    assert bins[-1] == 1


def test_bar_errbar_zorder():
    # Check that the zorder of errorbars is always greater than the bar they
    # are plotted on
    fig, ax = plt.subplots()
    x = [1, 2, 3]
    barcont = ax.bar(x=x, height=x, yerr=x, capsize=5, zorder=3)

    data_line, caplines, barlinecols = barcont.errorbar.lines
    for bar in barcont.patches:
        for capline in caplines:
            assert capline.zorder > bar.zorder
        for barlinecol in barlinecols:
            assert barlinecol.zorder > bar.zorder


def test_set_ticks_inverted():
    fig, ax = plt.subplots()
    ax.invert_xaxis()
    ax.set_xticks([.3, .7])
    assert ax.get_xlim() == (1, 0)


def test_aspect_nonlinear_adjustable_box():
    fig = plt.figure(figsize=(10, 10))  # Square.

    ax = fig.add_subplot()
    ax.plot([.4, .6], [.4, .6])  # Set minpos to keep logit happy.
    ax.set(xscale="log", xlim=(1, 10),
           yscale="logit", ylim=(1/11, 1/1001),
           aspect=1, adjustable="box")
    ax.margins(0)
    pos = fig.transFigure.transform_bbox(ax.get_position())
    assert pos.height / pos.width == pytest.approx(2)


def test_aspect_nonlinear_adjustable_datalim():
    fig = plt.figure(figsize=(10, 10))  # Square.

    ax = fig.add_axes([.1, .1, .8, .8])  # Square.
    ax.plot([.4, .6], [.4, .6])  # Set minpos to keep logit happy.
    ax.set(xscale="log", xlim=(1, 100),
           yscale="logit", ylim=(1 / 101, 1 / 11),
           aspect=1, adjustable="datalim")
    ax.margins(0)
    ax.apply_aspect()

    assert ax.get_xlim() == pytest.approx([1*10**(1/2), 100/10**(1/2)])
    assert ax.get_ylim() == (1 / 101, 1 / 11)


def test_box_aspect():
    # Test if axes with box_aspect=1 has same dimensions
    # as axes with aspect equal and adjustable="box"

    fig1, ax1 = plt.subplots()
    axtwin = ax1.twinx()
    axtwin.plot([12, 344])

    ax1.set_box_aspect(1)

    fig2, ax2 = plt.subplots()
    ax2.margins(0)
    ax2.plot([0, 2], [6, 8])
    ax2.set_aspect("equal", adjustable="box")

    fig1.canvas.draw()
    fig2.canvas.draw()

    bb1 = ax1.get_position()
    bbt = axtwin.get_position()
    bb2 = ax2.get_position()

    assert_array_equal(bb1.extents, bb2.extents)
    assert_array_equal(bbt.extents, bb2.extents)


def test_box_aspect_custom_position():
    # Test if axes with custom position and box_aspect
    # behaves the same independent of the order of setting those.

    fig1, ax1 = plt.subplots()
    ax1.set_position([0.1, 0.1, 0.9, 0.2])
    fig1.canvas.draw()
    ax1.set_box_aspect(1.)

    fig2, ax2 = plt.subplots()
    ax2.set_box_aspect(1.)
    fig2.canvas.draw()
    ax2.set_position([0.1, 0.1, 0.9, 0.2])

    fig1.canvas.draw()
    fig2.canvas.draw()

    bb1 = ax1.get_position()
    bb2 = ax2.get_position()

    assert_array_equal(bb1.extents, bb2.extents)


def test_bbox_aspect_axes_init():
    # Test that box_aspect can be given to axes init and produces
    # all equal square axes.
    fig, axs = plt.subplots(2, 3, subplot_kw=dict(box_aspect=1),
                            constrained_layout=True)
    fig.canvas.draw()
    renderer = fig.canvas.get_renderer()
    sizes = []
    for ax in axs.flat:
        bb = ax.get_window_extent(renderer)
        sizes.extend([bb.width, bb.height])

    assert_allclose(sizes, sizes[0])


def test_redraw_in_frame():
    fig, ax = plt.subplots(1, 1)
    ax.plot([1, 2, 3])
    fig.canvas.draw()
    ax.redraw_in_frame()


def test_invisible_axes():
    # invisible axes should not respond to events...
    fig, ax = plt.subplots()
    assert fig.canvas.inaxes((200, 200)) is not None
    ax.set_visible(False)
    assert fig.canvas.inaxes((200, 200)) is None


def test_xtickcolor_is_not_markercolor():
    plt.rcParams['lines.markeredgecolor'] = 'white'
    ax = plt.axes()
    ticks = ax.xaxis.get_major_ticks()
    for tick in ticks:
        assert tick.tick1line.get_markeredgecolor() != 'white'


def test_ytickcolor_is_not_markercolor():
    plt.rcParams['lines.markeredgecolor'] = 'white'
    ax = plt.axes()
    ticks = ax.yaxis.get_major_ticks()
    for tick in ticks:
        assert tick.tick1line.get_markeredgecolor() != 'white'


@pytest.mark.parametrize('axis', ('x', 'y'))
@pytest.mark.parametrize('auto', (True, False, None))
def test_unautoscale(axis, auto):
    fig, ax = plt.subplots()
    x = np.arange(100)
    y = np.linspace(-.1, .1, 100)
    ax.scatter(y, x)

    get_autoscale_on = getattr(ax, f'get_autoscale{axis}_on')
    set_lim = getattr(ax, f'set_{axis}lim')
    get_lim = getattr(ax, f'get_{axis}lim')

    post_auto = get_autoscale_on() if auto is None else auto

    set_lim((-0.5, 0.5), auto=auto)
    assert post_auto == get_autoscale_on()
    fig.canvas.draw()
    assert_array_equal(get_lim(), (-0.5, 0.5))


@check_figures_equal(extensions=["png"])
def test_polar_interpolation_steps_variable_r(fig_test, fig_ref):
    l, = fig_test.add_subplot(projection="polar").plot([0, np.pi/2], [1, 2])
    l.get_path()._interpolation_steps = 100
    fig_ref.add_subplot(projection="polar").plot(
        np.linspace(0, np.pi/2, 101), np.linspace(1, 2, 101))


@pytest.mark.style('default')
def test_autoscale_tiny_sticky():
    fig, ax = plt.subplots()
    ax.bar(0, 1e-9)
    fig.canvas.draw()
    assert ax.get_ylim() == (0, 1.05e-9)


def test_xtickcolor_is_not_xticklabelcolor():
    plt.rcParams['xtick.color'] = 'yellow'
    plt.rcParams['xtick.labelcolor'] = 'blue'
    ax = plt.axes()
    ticks = ax.xaxis.get_major_ticks()
    for tick in ticks:
        assert tick.tick1line.get_color() == 'yellow'
        assert tick.label1.get_color() == 'blue'


def test_ytickcolor_is_not_yticklabelcolor():
    plt.rcParams['ytick.color'] = 'yellow'
    plt.rcParams['ytick.labelcolor'] = 'blue'
    ax = plt.axes()
    ticks = ax.yaxis.get_major_ticks()
    for tick in ticks:
        assert tick.tick1line.get_color() == 'yellow'
        assert tick.label1.get_color() == 'blue'


@pytest.mark.parametrize('size', [size for size in mfont_manager.font_scalings
                                  if size is not None] + [8, 10, 12])
@pytest.mark.style('default')
def test_relative_ticklabel_sizes(size):
    mpl.rcParams['xtick.labelsize'] = size
    mpl.rcParams['ytick.labelsize'] = size
    fig, ax = plt.subplots()
    fig.canvas.draw()

    for name, axis in zip(['x', 'y'], [ax.xaxis, ax.yaxis]):
        for tick in axis.get_major_ticks():
            assert tick.label1.get_size() == axis._get_tick_label_size(name)


def test_multiplot_autoscale():
    fig = plt.figure()
    ax1, ax2 = fig.subplots(2, 1, sharex='all')
    ax1.scatter([1, 2, 3, 4], [2, 3, 2, 3])
    ax2.axhspan(-5, 5)
    xlim = ax1.get_xlim()
    assert np.allclose(xlim, [0.5, 4.5])


def test_sharing_does_not_link_positions():
    fig = plt.figure()
    ax0 = fig.add_subplot(221)
    ax1 = fig.add_axes([.6, .6, .3, .3], sharex=ax0)
    init_pos = ax1.get_position()
    fig.subplots_adjust(left=0)
    assert (ax1.get_position().get_points() == init_pos.get_points()).all()


@check_figures_equal(extensions=["pdf"])
def test_2dcolor_plot(fig_test, fig_ref):
    color = np.array([0.1, 0.2, 0.3])
    # plot with 1D-color:
    axs = fig_test.subplots(5)
    axs[0].plot([1, 2], [1, 2], c=color.reshape(-1))
    axs[1].scatter([1, 2], [1, 2], c=color.reshape(-1))
    axs[2].step([1, 2], [1, 2], c=color.reshape(-1))
    axs[3].hist(np.arange(10), color=color.reshape(-1))
    axs[4].bar(np.arange(10), np.arange(10), color=color.reshape(-1))
    # plot with 2D-color:
    axs = fig_ref.subplots(5)
    axs[0].plot([1, 2], [1, 2], c=color.reshape((1, -1)))
    axs[1].scatter([1, 2], [1, 2], c=color.reshape((1, -1)))
    axs[2].step([1, 2], [1, 2], c=color.reshape((1, -1)))
    axs[3].hist(np.arange(10), color=color.reshape((1, -1)))
    axs[4].bar(np.arange(10), np.arange(10), color=color.reshape((1, -1)))


def test_shared_axes_retick():
    fig, axs = plt.subplots(2, 2, sharex='all', sharey='all')

    for ax in axs.flat:
        ax.plot([0, 2], 'o-')

    axs[0, 0].set_xticks([-0.5, 0, 1, 1.5])  # should affect all axes xlims
    for ax in axs.flat:
        assert ax.get_xlim() == axs[0, 0].get_xlim()

    axs[0, 0].set_yticks([-0.5, 0, 2, 2.5])  # should affect all axes ylims
    for ax in axs.flat:
        assert ax.get_ylim() == axs[0, 0].get_ylim()


@pytest.mark.parametrize('ha', ['left', 'center', 'right'])
def test_ylabel_ha_with_position(ha):
    fig = Figure()
    ax = fig.subplots()
    ax.set_ylabel("test", y=1, ha=ha)
    ax.yaxis.set_label_position("right")
    assert ax.yaxis.get_label().get_ha() == ha


def test_bar_label_location_vertical():
    ax = plt.gca()
    xs, heights = [1, 2], [3, -4]
    rects = ax.bar(xs, heights)
    labels = ax.bar_label(rects)
    assert labels[0].xy == (xs[0], heights[0])
    assert labels[0].get_ha() == 'center'
    assert labels[0].get_va() == 'bottom'
    assert labels[1].xy == (xs[1], heights[1])
    assert labels[1].get_ha() == 'center'
    assert labels[1].get_va() == 'top'


def test_bar_label_location_horizontal():
    ax = plt.gca()
    ys, widths = [1, 2], [3, -4]
    rects = ax.barh(ys, widths)
    labels = ax.bar_label(rects)
    assert labels[0].xy == (widths[0], ys[0])
    assert labels[0].get_ha() == 'left'
    assert labels[0].get_va() == 'center'
    assert labels[1].xy == (widths[1], ys[1])
    assert labels[1].get_ha() == 'right'
    assert labels[1].get_va() == 'center'


def test_bar_label_location_center():
    ax = plt.gca()
    ys, widths = [1, 2], [3, -4]
    rects = ax.barh(ys, widths)
    labels = ax.bar_label(rects, label_type='center')
    assert labels[0].xy == (widths[0] / 2, ys[0])
    assert labels[0].get_ha() == 'center'
    assert labels[0].get_va() == 'center'
    assert labels[1].xy == (widths[1] / 2, ys[1])
    assert labels[1].get_ha() == 'center'
    assert labels[1].get_va() == 'center'


def test_bar_label_location_errorbars():
    ax = plt.gca()
    xs, heights = [1, 2], [3, -4]
    rects = ax.bar(xs, heights, yerr=1)
    labels = ax.bar_label(rects)
    assert labels[0].xy == (xs[0], heights[0] + 1)
    assert labels[0].get_ha() == 'center'
    assert labels[0].get_va() == 'bottom'
    assert labels[1].xy == (xs[1], heights[1] - 1)
    assert labels[1].get_ha() == 'center'
    assert labels[1].get_va() == 'top'


def test_bar_label_fmt():
    ax = plt.gca()
    rects = ax.bar([1, 2], [3, -4])
    labels = ax.bar_label(rects, fmt='%.2f')
    assert labels[0].get_text() == '3.00'
    assert labels[1].get_text() == '-4.00'


def test_bar_label_labels():
    ax = plt.gca()
    rects = ax.bar([1, 2], [3, -4])
    labels = ax.bar_label(rects, labels=['A', 'B'])
    assert labels[0].get_text() == 'A'
    assert labels[1].get_text() == 'B'


def test_patch_bounds():  # PR 19078
    fig, ax = plt.subplots()
    ax.add_patch(mpatches.Wedge((0, -1), 1.05, 60, 120, 0.1))
    bot = 1.9*np.sin(15*np.pi/180)**2
    np.testing.assert_array_almost_equal_nulp(
        np.array((-0.525, -(bot+0.05), 1.05, bot+0.1)), ax.dataLim.bounds, 16)