8000 DOC: move 2.0 whats_new to prev_whats_new · matplotlib/matplotlib@fef856e · GitHub
[go: up one dir, main page]
Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly
Appearance settings

Commit fef856e

Browse files
tacaswelltacaswell
committed
DOC: move 2.0 whats_new to prev_whats_new
1 parent 3c25356 commit fef856e

File tree

2 files changed

+320
-316
lines changed

2 files changed

+320
-316
lines changed

‎doc/users/prev_whats_new/whats_new_2.0.0.rst

Lines changed: 317 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,317 @@
1+
.. _whats-new-2-0-0:
2+
3+
New in matplotlib 2.0
4+
=====================
5+
6+
.. note::
7+
8+
matplotlib 2.0 supports Python 2.7, and 3.4+
9+
10+
11+
12+
Default style changes
13+
---------------------
14+
15+
The major changes in v2.0 are related to overhauling the default styles.
16+
17+
.. toctree::
18+
:maxdepth: 2
19+
20+
../dflt_style_changes
21+
22+
23+
Improved color conversion API and RGBA support
24+
----------------------------------------------
25+
26+
The :mod:`~matplotlib.colors` gained a new color conversion API with
27+
full support for the alpha channel. The main public functions are
28+
:func:`~matplotlib.colors.is_color_like`, :func:`matplotlib.colors.to_rgba`,
29+
:func:`matplotlib.colors.to_rgba_array` and :func:`~matplotlib.colors.to_hex`.
30+
RGBA quadruplets are encoded in hex format as `#rrggbbaa`.
31+
32+
A side benefit is that the Qt options editor now allows setting the alpha
33+
channel of the artists as well.
34+
35+
36+
New Configuration (rcParams)
37+
----------------------------
38+
39+
New rcparams added
40+
41+
+---------------------------------+--------------------------------------------------+
42+
| Parameter | Description |
43+
+=================================+==================================================+
44+
|`date.autoformatter.year` | format string for 'year' scale dates |
45+
+---------------------------------+--------------------------------------------------+
46+
|`date.autoformatter.month` | format string for 'month' scale dates |
47+
+---------------------------------+--------------------------------------------------+
48+
|`date.autoformatter.day` | format string for 'day' scale dates |
49+
+---------------------------------+--------------------------------------------------+
50+
|`date.autoformatter.hour` | format string for 'hour' scale times |
51+
+---------------------------------+--------------------------------------------------+
52+
|`date.autoformatter.minute` | format string for 'minute' scale times |
53+
+---------------------------------+--------------------------------------------------+
54+
|`date.autoformatter.second` | format string for 'second' scale times |
55+
+---------------------------------+--------------------------------------------------+
56+
|`date.autoformatter.microsecond` | format string for 'microsecond' scale times |
57+
+---------------------------------+--------------------------------------------------+
58+
|`scatter.marker` | default marker for scatter plot |
59+
+---------------------------------+--------------------------------------------------+
60+
|`svg.hashsalt` | see note |
61+
+---------------------------------+--------------------------------------------------+
62+
|`xtick.top`, `xtick.minor.top`, | Control where major and minor ticks are drawn. |
63+
|`xtick.major.top` | The global values are `and` ed with the |
64+
|`xtick.bottom`, | corresponding major/minor values. |
65+
|`xtick.minor.bottom`, | |
66+
|`xtick.major.bottom` | |
67+
|`ytick.left`, `ytick.minor.left`,| |
68+
|`ytick.major.left` | |
69+
|`ytick.right`, | |
70+
|`ytick.minor.right`, | |
71+
|`ytick.major.right` | |
72+
+---------------------------------+--------------------------------------------------+
73+
|`hist.bins` | The default number of bins to use in |
74+
| | `~matplotlib.axes.Axes.hist`. This can be an |
75+
| | `int`, a list of floats, or ``'auto'`` if numpy |
76+
| | >= 1.11 is installed. |
77+
+---------------------------------+--------------------------------------------------+
78+
|`lines.scale_dashes` | Whether the line dash patterns should scale with |
79+
| | linewidth. |
80+
+---------------------------------+--------------------------------------------------+
81+
|`axes.formatter.offset_threshold`| Minimum number of digits saved in tick labels |
82+
| | that triggers using an offset. |
83+
+---------------------------------+--------------------------------------------------+
84+
85+
86+
87+
Added ``svg.hashsalt`` key to rcParams
88+
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
89+
90+
If ``svg.hashsalt`` is ``None`` (which it is by default), the svg
91+
backend uses ``uuid4`` to generate the hash salt. If it is not
92+
``None``, it must be a string that is used as the hash salt instead of
93+
``uuid4``. This allows for deterministic SVG output.
94+
95+
96+
Removed the ``svg.image_noscale`` rcParam
97+
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
98+
99+
As a result of the extensive changes to image handling, the
100+
``svg.image_noscale`` rcParam has been removed. The same
101+
functionality may be achieved by setting ``interpolation='none'`` on
102+
individual images or globally using the ``image.interpolation``
103+
rcParam.
104+
105+
106+
Qualitative colormaps
107+
---------------------
108+
109+
ColorBrewer's "qualitative" colormaps ("Accent", "Dark2", "Paired",
110+
"Pastel1", "Pastel2", "Set1", "Set2", "Set3") were intended for discrete
111+
categorical data, with no implication of value, and therefore have been
112+
converted to ``ListedColormap`` instead of ``LinearSegmentedColormap``, so
113+
the colors will no longer be interpolated and they can be used for
114+
choropleths, labeled image features, etc.
115+
116+
117+
118+
Axis offset label now responds to `labelcolor`
119+
----------------------------------------------
120+
121+
Axis offset labels are now colored the same as axis tick markers when `labelcolor` is altered.
122+
123+
Improved offset text choice
124+
---------------------------
125+
The default offset-text choice was changed to only use significant digits that
126+
are common to all ticks (e.g. 1231..1239 -> 1230, instead of 1231), except when
127+
they straddle a relatively large multiple of a power of ten, in which case that
128+
multiple is chosen (e.g. 1999..2001->2000).
129+
130+
131+
Style parameter blacklist
132+
-------------------------
133+
134+
In order to prevent unexpected consequences from using a style, style
135+
files are no longer able to set parameters that affect things
136+
unrelated to style. These parameters include::
137+
138+
'interactive', 'backend', 'backend.qt4', 'webagg.port',
139+
'webagg.port_retries', 'webagg.open_in_browser', 'backend_fallback',
140+
'toolbar', 'timezone', 'datapath', 'figure.max_open_warning',
141+
'savefig.directory', 'tk.window_focus', 'docstring.hardcopy'
142+
143+
144+
Change in default font
145+
----------------------
146+
147+
The default font used by matplotlib in text has been changed to DejaVu Sans and
148+
DejaVu Serif for the sans-serif and serif families, respectively. The DejaVu
149+
font family is based on the previous matplotlib default --Bitstream Vera-- but
150+
includes a much wider range of characters.
151+
152+
The default mathtext font has been changed from Computer Modern to the DejaVu
153+
family to maintain consistency with regular text. Two new options for the
154+
``mathtext.fontset`` configuration parameter have been added: ``dejavusans``
155+
(default) and ``dejavuserif``. Both of these options use DejaVu glyphs whenever
156+
possible and fall back to STIX symbols when a glyph is not found in DejaVu. To
157+
return to the previous behavior, set the rcParam ``mathtext.fontset`` to ``cm``.
158+
159+
160+
Faster text rendering
161+
---------------------
162+
163+
Rendering text in the Agg backend is now less fuzzy and about 20%
164+
faster to draw.
165+
166+
167+
Improvements for the Qt figure options editor
168+
---------------------------------------------
169+
170+
Various usability improvements were implemented for the Qt figure options
171+
editor, among which:
172+
173+
- Line style entries are now sorted without duplicates.
174+
- The colormap and normalization limits can now be set for images.
175+
- Line edits for floating values now display only as many digits as necessary
176+
to avoid precision loss. An important bug was also fixed regarding input
177+
validation using Qt5 and a locale where the decimal separator is ",".
178+
- The axes selector now uses shorter, more user-friendly names for axes, and
179+
does not crash if there are no axes.
180+
- Line and image entries using the default labels ("_lineX", "_imageX") are now
181+
sorted numerically even when there are more than 10 entries.
182+
183+
184+
Improved image support
185+
----------------------
186+
187+
Prior to version 2.0, matplotlib resampled images by first applying
188+
the color map and then resizing the result. Since the resampling was
189+
performed on the colored image, this introduced colors in the output
190+
image that didn't actually exist in the color map. Now, images are
191+
resampled first (and entirely in floating-point, if the input image is
192+
floating-point), and then the color map is applied.
193+
194+
In order to make this important change, the image handling code was
195+
almost entirely rewritten. As a side effect, image resampling uses
196+
less memory and fewer datatype conversions than before.
197+
198+
The experimental private feature where one could "skew" an image by
199+
setting the private member ``_image_skew_coordinate`` has been
200+
removed. Instead, images will obey the transform of the axes on which
201+
they are drawn.
202+
203+
Non-linear scales on image plots
204+
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
205+
206+
:func:`imshow` now draws data at the requested points in data space after the
207+
application of non-linear scales.
208+
209+
The image on the left demonstrates the new, correct behavior.
210+
The old behavior can be recreated using :func:`pcolormesh` as
211+
demonstrated on the right.
212+
213+
214+
.. plot::
215+
216+
import numpy as np
217+
import matplotlib.pyplot as plt
218+
219+
data = np.arange(30).reshape(5, 6)
220+
x = np.linspace(0, 6, 7)
221+
y = 10**np.linspace(0, 5, 6)
222+
X, Y = np.meshgrid(x, y)
223+
224+
fig, (ax1, ax2) = plt.subplots(ncols=2, figsize=(8, 4))
225+
226+
ax1.imshow(data, aspect="auto", extent=(0, 6, 1e0, 1e5), interpolation='nearest')
227+
ax1.set_yscale('log')
228+
ax1.set_title('Using ax.imshow')
229+
230+
ax2.pcolormesh(x, y, np.flipud(data))
231+
ax2.set_yscale('log')
232+
ax2.set_title('Using ax.pcolormesh')
233+
ax2.autoscale('tight')
234+
235+
plt.show()
236+
237+
238+
This can be understood by analogy to plotting a histogram with linearly spaced bins
239+
with a logarithmic x-axis. Equal sized bins will be displayed as wider for small
240+
*x* and narrower for large *x*.
241+
242+
243+
244+
Support for HiDPI (Retina) displays in the NbAgg and WebAgg backends
245+
--------------------------------------------------------------------
246+
247+
The NbAgg and WebAgg backends will now use the full resolution of your
248+
high-pixel-density display.
249+
250+
Change in the default animation codec
251+
-------------------------------------
252+
253+
The default animation codec has been changed from ``mpeg4`` to ``h264``,
254+
which is more efficient. It can be set via the ``animation.codec`` rcParam.
255+
256+
Deprecated support for mencoder in animation
257+
--------------------------------------------
258+
259+
The use of mencoder for writing video files with mpl is problematic;
260+
switching to ffmpeg is strongly advised. All support for mencoder
261+
will be removed in version 2.2.
262+
263+
Boxplot Zorder Keyword Argument
264+
-------------------------------
265+
266+
The ``zorder`` parameter now exists for :func:`boxplot`. This allows the zorder
267+
of a boxplot to be set in the plotting function call.
268+
269+
::
270+
271+
boxplot(np.arange(10), zorder=10)
272+
273+
Filled ``+`` and ``x`` markers
274+
------------------------------
275+
276+
New fillable *plus* and *x* markers have been added. See
277+
the :mod:`~matplotlib.markers` module and
278+
:ref:`marker reference <sphx_glr_gallery_lines_bars_and_markers_marker_reference.py>`
279+
examples.
280+
281+
`rcount` and `ccount` for `plot_surface()`
282+
------------------------------------------
283+
284+
As of v2.0, mplot3d's :func:`~mpl_toolkits.mplot3d.axes3d.plot_surface` now
285+
accepts `rcount` and `ccount` arguments for controlling the sampling of the
286+
input data for plotting. These arguments specify the maximum number of
287+
evenly spaced samples to take from the input data. These arguments are
288+
also the new default sampling method for the function, and is
289+
considered a style change.
290+
291+
The old `rstride` and `cstride` arguments, which specified the size of the
292+
evenly spaced samples, become the default when 'classic' mode is invoked,
293+
and are still available for use. There are no plans for deprecating these
294+
arguments.
295+
296+
Streamplot Zorder Keyword Argument Changes
297+
------------------------------------------
298+
299+
The ``zorder`` parameter for :func:`streamplot` now has default
300+
value of ``None`` instead of ``2``. If ``None`` is given as ``zorder``,
301+
:func:`streamplot` has a default ``zorder`` of
302+
``matplotlib.lines.Line2D.zorder``.
303+
304+
.. _gc_get_hatch_color_wn:
305+
306+
Extension to `matplotlib.backend_bases.GraphicsContextBase`
307+
-----------------------------------------------------------
308+
309+
To support standardizing hatch behavior across the backends we ship
310+
the `matplotlib.backend_bases.GraphicsContextBase.get_hatch_color`
311+
method as added to `matplotlib.backend_bases.GraphicsContextBase`.
312+
This is only used during the render process in the backends we ship so
313+
will not break any third-party backends.
314+
315+
If you maintain a third-party backend which extends
316+
`~matplotlib.backend_bases.GraphicsContextBase` this method is now
317+
available to you and should be used to color hatch patterns.

0 commit comments

Comments
 (0)
0