matplotlib · QuLogic · Dec 13, 2016 · Dec 7, 2016 · Dec 8, 2016 · Dec 8, 2016
diff --git a/doc/api/animation_api.rst b/doc/api/animation_api.rst
@@ -1,12 +1,333 @@
-*********
-animation
-*********
+======================
+ ``animation`` module
+======================
 
+.. automodule:: matplotlib.animation
 
-:mod:`matplotlib.animation`
-===========================
+.. contents:: Table of Contents
+   :depth: 1
+   :local:
+   :backlinks: entry
 
-.. automodule:: matplotlib.animation
-   :members:
-   :undoc-members:
-   :show-inheritance:
+
+Animation
+=========
+
+The easiest way to make a live animation in matplotlib is to use one of the
+`Animation` classes.
+
+.. autosummary::
+   :toctree: _as_gen
+   :nosignatures:
+
+   FuncAnimation
+   ArtistAnimation
+
+In both cases it is critical to keep a reference to the instance
+object.  The animation is advanced by a timer (typically from the host
+GUI framework) which the `Animation` object holds the only reference
+to.  If you do not hold a reference to the `Animation` object, it (and
+hence the timers), will be garbage collected which will stop the
+animation.
+
+To save an animation to disk use
+
+.. autosummary::
+   :toctree: _as_gen
+   :nosignatures:
+
+   Animation.save
+   Animation.to_html5_video
+
+See :ref:`ani_writer_classes` below for details about what movie formats are supported.
+
+
+``FuncAnimation``
+-----------------
+
+The inner workings of `FuncAnimation` is more-or-less::
+
+  for d in frames:
+     artists = func(d, *fargs)
+     fig.canvas.draw_idle()
+     plt.pause(interval)
+
+
+with details to handle 'blitting' (to dramatically improve the live
+performance), to be non-blocking, handle repeats, multiple animated
+axes, and easily save the animation to a movie file.
+
+'Blitting' is a `old technique
+<https://en.wikipedia.org/wiki/Bit_blit>`__ in computer graphics.  The
+general gist is to take an existing bit map (in our case a mostly
+rasterized figure) and then 'blit' one more artist on top.  Thus, by
+managing a saved 'clean' bitmap, we can only re-draw the few artists
+that are changing at each frame and possibly save significant amounts of
+time.  When using blitting (by passing ``blit=True``) the core loop of
+`FuncAnimation` gets a bit more complicated ::
+
+   ax = fig.gca()
+
+   def update_blit(artists):
+       fig.canvas.restore_region(bg_cache)
+       for a in artists:
+           a.axes.draw_artist(a)
+
+       ax.figure.canvas.blit(ax.bbox)
+
+   artists = init_func()
+
+   for a in artists:
+      a.set_animated(True)
+
+   fig.canvas.draw()
+   bg_cache = fig.canvas.copy_from_bbox(ax.bbox)
+
+   for f in frames:
+       artists = func(f, *fargs)
+       update_blit(artists)
+       plt.pause(interval)
+
+This is of course leaving out many details (such as updating the
+background when the figure is resized or fully re-drawn).  However,
+this hopefully minimalist example gives a sense of how ``init_func``
+and ``func`` are used inside of `FuncAnimation` and the theory of how
+'blitting' works.
+
+The expected signature on ``func`` and ``init_func`` is very simple to
+keep `FuncAnimation` out of your book keeping and plotting logic, but
+this means that the callable objects you pass in must know what
+artists they should be working on.  There are several approaches to
+handling this, of varying complexity and encapsulation.  The simplest
+approach, which works quite well in the case of a script, is to define the
+artist at a global scope and let Python sort things out.  For example ::
+
+   import numpy as np
+   import matplotlib.pyplot as plt
+   from matplotlib.animation import FuncAnimation
+
+   fig, ax = plt.subplots()
+   xdata, ydata = [], []
+   ln, = plt.plot([], [], 'ro', animated=True)
+
+   def init():
+       ax.set_xlim(0, 2*np.pi)
+       ax.set_ylim(-1, 1)
+       return ln,
+
+   def update(frame):
+       xdata.append(frame)
+       ydata.append(np.sin(frame))
+       ln.set_data(xdata, ydata)
+       return ln,
+
+   ani = FuncAnimation(fig, update, frames=np.linspace(0, 2*np.pi, 128),
+                       init_func=init, blit=True)
+   plt.show()
+
+
+The second method is to us `functools.partial` to 'bind' artists to
+function.  A third method is to use closures to build up the required
+artists and functions.  A fourth method is to create a class.
+
+
+
+
+Examples
+~~~~~~~~
+
+.. toctree::
+   :maxdepth: 1
+
+   ../examples/animation/animate_decay
+   ../examples/animation/bayes_update
+   ../examples/animation/double_pendulum_animated
+   ../examples/animation/dynamic_image
+   ../examples/animation/histogram
+   ../examples/animation/rain
+   ../examples/animation/random_data
+   ../examples/animation/simple_3danim
+   ../examples/animation/simple_anim
+   ../examples/animation/strip_chart_demo
+   ../examples/animation/unchained
+
+``ArtistAnimation``
+-------------------
+
+
+Examples
+~~~~~~~~
+
+.. toctree::
+   :maxdepth: 1
+
+   ../examples/animation/basic_example
+   ../examples/animation/basic_example_writer
+   ../examples/animation/dynamic_image2
+
+
+
+
+Writer Classes
+==============
+
+
+
+The provided writers fall into two broad categories: pipe-based and
+file-based.  The pipe-based writers stream the captured frames over a
+pipe to an external process.  The pipe-based variants tend to be more
+performant, but may not work on all systems.
+
+.. autosummary::
+   :toctree: _as_gen
+   :nosignatures:
+
+
+   FFMpegWriter
+   ImageMagickFileWriter
+   AVConvWriter
+
+Alternatively the file-based writers save temporary files for each
+frame which are stitched into a single file at the end.  Although
+slower, these writers can be easier to debug.
+
+.. autosummary::
+   :toctree: _as_gen
+   :nosignatures:
+
+   FFMpegFileWriter
+   ImageMagickWriter
+   AVConvFileWriter
+
+
+Fundamentally, a `MovieWriter` provides a way to grab sequential frames
+from the same underlying `~matplotlib.figure.Figure` object.  The base
+class `MovieWriter` implements 3 methods and a context manager.  The
+only difference between the pipe-based and file-based writers is in the
+arguments to their respective ``setup`` methods.
+
+
+.. autosummary::
+   :toctree: _as_gen
+   :nosignatures:
+
+   MovieWriter.setup
+   FileMovieWriter.setup
+   MovieWriter.grab_frame
+   MovieWriter.finish
+   MovieWriter.saving
+
+
+The ``setup()`` method is used to prepare the writer (possibly opening
+a pipe), successive calls to ``grab_frame()`` capture a single frame
+at a time and ``finish()`` finalizes the movie and writes the output
+file to disk.  For example ::
+
+   moviewriter = MovieWriter(...)
+   moviewriter.setup(fig=fig, 'my_movie.ext', dpi=100)
+   for j in range(n):
+       update_figure(n)
+       moviewriter.grab_frame()
+   moviewriter.finish()
+
+
+If using the writer classes directly (not through `Animation.save`), it is strongly encouraged
+to use the `~MovieWriter.saving` context manager ::
+
+  with moviewriter.saving(fig, 'myfile.mp4', dpi=100):
+      for j in range(n):
+          update_figure(n)
+          moviewriter.grab_frame()
+
+
+to ensures that setup and cleanup are performed as necessary.
+
+
+:ref:`animation-moviewriter`
+
+
+.. _ani_writer_classes:
+
+Helper Classes
+==============
+
+
+Animation Base Classes
+----------------------
+
+
+.. autosummary::
+   :toctree: _as_gen
+   :nosignatures:
+
+   Animation
+   TimedAnimation
+
+
+Custom Animation classes
+------------------------
+
+:ref:`animation-subplots`
+
+Writer Registry
+---------------
+
+A module-level registry is provided to map between the name of the
+writer and the class to allow a string to be passed to
+`Animation.save` instead of a writer instance.
+
+.. autosummary::
+   :toctree: _as_gen
+   :nosignatures:
+
+   MovieWriterRegistry
+
+Writer Base Classes
+-------------------
+
+To reduce code duplication base classes
+
+.. autosummary::
+   :toctree: _as_gen
+   :nosignatures:
+
+   AbstractMovieWriter
+   MovieWriter
+   FileMovieWriter
+
+and mixins are provided
+
+.. autosummary::
+   :toctree: _as_gen
+   :nosignatures:
+
+   AVConvBase
+   FFMpegBase
+   ImageMagickBase
+
+See the source code for how to easily implement new `MovieWriter`
+classes.
+
+
+Inheritance Diagrams
+====================
+
+.. inheritance-diagram:: matplotlib.animation.FuncAnimation matplotlib.animation.ArtistAnimation
+   :private-bases:
+
+.. inheritance-diagram:: matplotlib.animation.AVConvFileWriter matplotlib.animation.AVConvWriter matplotlib.animation.FFMpegFileWriter matplotlib.animation.FFMpegWriter matplotlib.animation.ImageMagickFileWriter matplotlib.animation.ImageMagickWriter
+   :private-bases:
+
+
+
+Deprecated
+==========
+
+
+.. autosummary::
+   :toctree: _as_gen
+   :nosignatures:
+
+   MencoderBase
+   MencoderFileWriter
+   MencoderWriter
diff --git a/doc/api/api_changes.rst b/doc/api/api_changes.rst
@@ -141,8 +141,8 @@ the kwarg is None which internally sets it to the 'auto' string,
 triggering a new algorithm for adjusting the maximum according
 to the axis length relative to the ticklabel font size.
 
-`matplotlib.ticker.LogFormatter` gains minor_thresholds kwarg
--------------------------------------------------------------
+`matplotlib.ticker.LogFormatter`: two new kwargs
+------------------------------------------------
 
 Previously, minor ticks on log-scaled axes were not labeled by
 default.  An algorithm has been added to the
@@ -151,6 +151,9 @@ ticks between integer powers of the base.  The algorithm uses
 two parameters supplied in a kwarg tuple named 'minor_thresholds'.
 See the docstring for further explanation.
 
+To improve support for axes using `~matplotlib.ticker.SymmetricLogLocator`,
+a 'linthresh' kwarg was added.
+
 
 New defaults for 3D quiver function in mpl_toolkits.mplot3d.axes3d.py
 ---------------------------------------------------------------------