fr/fr_env/lib/python3.8/site-packages/matplotlib/tests/test_lines.py

266 lines
8.1 KiB
Python

"""
Tests specific to the lines module.
"""
import itertools
import timeit
from cycler import cycler
import numpy as np
from numpy.testing import assert_array_equal
import pytest
import matplotlib
import matplotlib.lines as mlines
from matplotlib.markers import MarkerStyle
from matplotlib.path import Path
import matplotlib.pyplot as plt
from matplotlib.testing.decorators import image_comparison, check_figures_equal
def test_segment_hits():
"""Test a problematic case."""
cx, cy = 553, 902
x, y = np.array([553., 553.]), np.array([95., 947.])
radius = 6.94
assert_array_equal(mlines.segment_hits(cx, cy, x, y, radius), [0])
# Runtimes on a loaded system are inherently flaky. Not so much that a rerun
# won't help, hopefully.
@pytest.mark.flaky(reruns=3)
def test_invisible_Line_rendering():
"""
GitHub issue #1256 identified a bug in Line.draw method
Despite visibility attribute set to False, the draw method was not
returning early enough and some pre-rendering code was executed
though not necessary.
Consequence was an excessive draw time for invisible Line instances
holding a large number of points (Npts> 10**6)
"""
# Creates big x and y data:
N = 10**7
x = np.linspace(0, 1, N)
y = np.random.normal(size=N)
# Create a plot figure:
fig = plt.figure()
ax = plt.subplot(111)
# Create a "big" Line instance:
l = mlines.Line2D(x, y)
l.set_visible(False)
# but don't add it to the Axis instance `ax`
# [here Interactive panning and zooming is pretty responsive]
# Time the canvas drawing:
t_no_line = min(timeit.repeat(fig.canvas.draw, number=1, repeat=3))
# (gives about 25 ms)
# Add the big invisible Line:
ax.add_line(l)
# [Now interactive panning and zooming is very slow]
# Time the canvas drawing:
t_invisible_line = min(timeit.repeat(fig.canvas.draw, number=1, repeat=3))
# gives about 290 ms for N = 10**7 pts
slowdown_factor = t_invisible_line / t_no_line
slowdown_threshold = 2 # trying to avoid false positive failures
assert slowdown_factor < slowdown_threshold
def test_set_line_coll_dash():
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
np.random.seed(0)
# Testing setting linestyles for line collections.
# This should not produce an error.
ax.contour(np.random.randn(20, 30), linestyles=[(0, (3, 3))])
@image_comparison(['line_dashes'], remove_text=True)
def test_line_dashes():
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
ax.plot(range(10), linestyle=(0, (3, 3)), lw=5)
def test_line_colors():
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
ax.plot(range(10), color='none')
ax.plot(range(10), color='r')
ax.plot(range(10), color='.3')
ax.plot(range(10), color=(1, 0, 0, 1))
ax.plot(range(10), color=(1, 0, 0))
fig.canvas.draw()
def test_linestyle_variants():
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
for ls in ["-", "solid", "--", "dashed",
"-.", "dashdot", ":", "dotted"]:
ax.plot(range(10), linestyle=ls)
fig.canvas.draw()
def test_valid_linestyles():
line = mlines.Line2D([], [])
with pytest.raises(ValueError):
line.set_linestyle('aardvark')
@image_comparison(['drawstyle_variants.png'], remove_text=True)
def test_drawstyle_variants():
fig, axs = plt.subplots(6)
dss = ["default", "steps-mid", "steps-pre", "steps-post", "steps", None]
# We want to check that drawstyles are properly handled even for very long
# lines (for which the subslice optimization is on); however, we need
# to zoom in so that the difference between the drawstyles is actually
# visible.
for ax, ds in zip(axs.flat, dss):
ax.plot(range(2000), drawstyle=ds)
ax.set(xlim=(0, 2), ylim=(0, 2))
def test_valid_drawstyles():
line = mlines.Line2D([], [])
with pytest.raises(ValueError):
line.set_drawstyle('foobar')
def test_set_drawstyle():
x = np.linspace(0, 2*np.pi, 10)
y = np.sin(x)
fig, ax = plt.subplots()
line, = ax.plot(x, y)
line.set_drawstyle("steps-pre")
assert len(line.get_path().vertices) == 2*len(x)-1
line.set_drawstyle("default")
assert len(line.get_path().vertices) == len(x)
@image_comparison(['line_collection_dashes'], remove_text=True, style='mpl20')
def test_set_line_coll_dash_image():
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
np.random.seed(0)
ax.contour(np.random.randn(20, 30), linestyles=[(0, (3, 3))])
@image_comparison(['marker_fill_styles.png'], remove_text=True)
def test_marker_fill_styles():
colors = itertools.cycle([[0, 0, 1], 'g', '#ff0000', 'c', 'm', 'y',
np.array([0, 0, 0])])
altcolor = 'lightgreen'
y = np.array([1, 1])
x = np.array([0, 9])
fig, ax = plt.subplots()
for j, marker in enumerate(mlines.Line2D.filled_markers):
for i, fs in enumerate(mlines.Line2D.fillStyles):
color = next(colors)
ax.plot(j * 10 + x, y + i + .5 * (j % 2),
marker=marker,
markersize=20,
markerfacecoloralt=altcolor,
fillstyle=fs,
label=fs,
linewidth=5,
color=color,
markeredgecolor=color,
markeredgewidth=2)
ax.set_ylim([0, 7.5])
ax.set_xlim([-5, 155])
def test_markerfacecolor_fillstyle():
"""Test that markerfacecolor does not override fillstyle='none'."""
l, = plt.plot([1, 3, 2], marker=MarkerStyle('o', fillstyle='none'),
markerfacecolor='red')
assert l.get_fillstyle() == 'none'
assert l.get_markerfacecolor() == 'none'
@image_comparison(['scaled_lines'], style='default')
def test_lw_scaling():
th = np.linspace(0, 32)
fig, ax = plt.subplots()
lins_styles = ['dashed', 'dotted', 'dashdot']
cy = cycler(matplotlib.rcParams['axes.prop_cycle'])
for j, (ls, sty) in enumerate(zip(lins_styles, cy)):
for lw in np.linspace(.5, 10, 10):
ax.plot(th, j*np.ones(50) + .1 * lw, linestyle=ls, lw=lw, **sty)
def test_nan_is_sorted():
line = mlines.Line2D([], [])
assert line._is_sorted(np.array([1, 2, 3]))
assert line._is_sorted(np.array([1, np.nan, 3]))
assert not line._is_sorted([3, 5] + [np.nan] * 100 + [0, 2])
@check_figures_equal()
def test_step_markers(fig_test, fig_ref):
fig_test.subplots().step([0, 1], "-o")
fig_ref.subplots().plot([0, 0, 1], [0, 1, 1], "-o", markevery=[0, 2])
@check_figures_equal(extensions=('png',))
def test_markevery(fig_test, fig_ref):
np.random.seed(42)
t = np.linspace(0, 3, 14)
y = np.random.rand(len(t))
casesA = [None, 4, (2, 5), [1, 5, 11],
[0, -1], slice(5, 10, 2), 0.3, (0.3, 0.4),
np.arange(len(t))[y > 0.5]]
casesB = ["11111111111111", "10001000100010", "00100001000010",
"01000100000100", "10000000000001", "00000101010000",
"11011011011110", "01010011011101", "01110001110110"]
axsA = fig_ref.subplots(3, 3)
axsB = fig_test.subplots(3, 3)
for ax, case in zip(axsA.flat, casesA):
ax.plot(t, y, "-gD", markevery=case)
for ax, case in zip(axsB.flat, casesB):
me = np.array(list(case)).astype(int).astype(bool)
ax.plot(t, y, "-gD", markevery=me)
def test_marker_as_markerstyle():
fig, ax = plt.subplots()
line, = ax.plot([2, 4, 3], marker=MarkerStyle("D"))
fig.canvas.draw()
assert line.get_marker() == "D"
# continue with smoke tests:
line.set_marker("s")
fig.canvas.draw()
line.set_marker(MarkerStyle("o"))
fig.canvas.draw()
# test Path roundtrip
triangle1 = Path([[-1., -1.], [1., -1.], [0., 2.], [0., 0.]], closed=True)
line2, = ax.plot([1, 3, 2], marker=MarkerStyle(triangle1), ms=22)
line3, = ax.plot([0, 2, 1], marker=triangle1, ms=22)
assert_array_equal(line2.get_marker().vertices, triangle1.vertices)
assert_array_equal(line3.get_marker().vertices, triangle1.vertices)
@check_figures_equal()
def test_odd_dashes(fig_test, fig_ref):
fig_test.add_subplot().plot([1, 2], dashes=[1, 2, 3])
fig_ref.add_subplot().plot([1, 2], dashes=[1, 2, 3, 1, 2, 3])