All formats

pyvips

LanguagePython
LicenseMIT (libvips: LGPL-2.1)
Repositorygithub.com/libvips/pyvips
Documentationpyvips docs, heifsave options
Depends onlibvips built with libheif; libheif with libde265 (decode) and x265 (encode)
Read / writeyes / yes, 8/10/12-bit, alpha, multiple images, thumbnails

pyvips is the Python binding of libvips, which processes images in a streaming way: fast, low memory, and good for converting many phone photos or generating thumbnails on a server. HEIC goes through libheif.

HEIC support depends on how libvips and libheif were built. The prebuilt pyvips-binary wheels and the Windows libvips builds include libheif without an HEVC decoder (only AV1, for AVIF), so they can’t open HEIC. Use a system libvips: Homebrew’s vips and Linux packages with the libheif HEVC plugins work. Check with vips heifload IMG_0001.heic x.jpg.

Installation#

Shell
# Debian / Ubuntu
sudo apt install libvips-dev libheif-plugin-libde265 libheif-plugin-x265
pip install pyvips
# macOS
brew install vips && pip install pyvips

Basic usage: convert HEIC to JPEG#

Python
import pyvips

image = pyvips.Image.new_from_file("IMG_0001.heic")
print(image.width, image.height, image.bands)
image.write_to_file("IMG_0001.jpg", Q=90)

libheif applies the rotation stored in the file, so the result is already upright.

How to make a thumbnail from a HEIC photo#

thumbnail can use the small preview image stored in the file when it is big enough, which is much faster than decoding the full photo:

Python
thumb = pyvips.Image.thumbnail("IMG_0001.heic", 400)  # fit in 400 x 400
thumb.write_to_file("thumb.jpg", Q=85)

How to save an image as HEIC#

Python
image = pyvips.Image.new_from_file("photo.png")
image.heifsave("photo.heic", Q=50, compression="hevc")  # Q 50 is about JPEG Q 90

Other options: lossless=True, bitdepth=10, and subsample_mode="off" for 4:4:4 chroma. compression="av1" (or a .avif file name) writes AVIF.

How to read Exif data#

Python
image = pyvips.Image.new_from_file("IMG_0001.heic")
if image.get_typeof("exif-data") != 0:
    exif = image.get("exif-data")              # raw Exif bytes
    print(image.get("exif-ifd0-Model"))        # parsed fields are also available

How to read all images in a file#

Python
first = pyvips.Image.new_from_file("burst.heic")
n = first.get("n-pages")
for page in range(n):
    img = pyvips.Image.new_from_file("burst.heic", page=page)
    print(page, img.width, img.height)

Tips#