Features

• Merge multiple images into one APNG file. (It use Pillow to convert images into PNG format)

• Read APNG file and extract each frames into PNG file.

• It doesn’t do any optimization but only concat the images. This might be changed in the future.

Dependencies

• Pillow - Optional. You can still use pyAPNG without PIL but it can only read PNG files.

Development dependencies

• pngcheck

• See requirements.txt for other dev-dependencies.

Installation

From pypi:

pip install apng

Usage

Convert a series of images into APNG animation:

from apng import APNG

APNG.from_files(["1.jpg", "2.jpg", "3.jpg"], delay=100).save("result.png")

Use different delays:

from apng import APNG

files = [
  ("1.jpg", 100),
  ("2.jpg", 200),
  ("3.jpg", 300)
]

im = APNG()
for file, delay in files:
  im.append_file(file, delay=delay)
im.save("result.png")

Extract frames from an APNG file:

from apng import APNG

im = APNG.open("animation.png")
for i, (png, control) in enumerate(im.frames):
  png.save("{i}.png".format(i=i))

Add a text chunk to the PNG file:

from apng import PNG, make_text_chunk

im = PNG.open("image.png")
im.chunks.append(make_text_chunk(key="Comment", value="Some text"))
im.save("image.png")

Performance

If you want to convert some large JPGs into animation, the library has to convert your JPGs into PNGs then merge them into a single animation APNG file. The problems are:

1. It is extremely slow.

2. The file size of the APNG is extremely large. Probably 5x of the original or more.

In this case, I suggest trying an animation format called “ugoira”, which is implemented by Pixiv.net. There is also an image viewer named “HoneyView” which can view it locally.

Document

http://pyapng.readthedocs.io/en/latest/

Todos

• Add optimizer?

Changelog

• 0.3.4 (Mar 11, 2020)

  • Fix: exclude test files from the package.

• 0.3.3 (Feb 11, 2019)

  • Fix: failed to extract frames containing multiple fdAT chunks.

• 0.3.2 (Jul 20, 2018)

  • Add: make_text_chunk function.

  • Add: Chunk data class.

  • Change: now parse_chunks yields Chunk instead of a tuple. This should be safe since Chunk is a namedtuple.

• 0.3.1 (May 13, 2018)

  • Add: universal wheel.

• 0.3.0 (May 13, 2018)

  • Support Python 2.

  • Add: PNG method open_any, from_bytes.

  • Add: APNG method append_file, from_bytes.

  • Add: module function parse_chunks.

  • Drop: module function `is_png` and `chunks`.

  • Change: `PNG.open` now only reads PNG images. To read non-PNG images, use `PNG.open_any`.

  • Change: `APNG.append` now only accepts `PNG` instance. To append PNG files, use `APNG.append_file`.

• 0.2.1 (Apr 19, 2018)

  • Add: support num_plays. (#4)

• 0.2.0 (Dec 8, 2017)

  • Add test.

  • Add documents.

  • Add: support path-like object.

  • Fix: some chunks must appear before IDAT. (#1)

  • Fix: change chunks order in APNG. Some chunks are moved to the end of the file.

  • Fix: remove tRNS hack.

  • Fix: is_png shouldn’t move file pointer. (#2)

• 0.1.0 (May 30, 2016)

  • First release.