Read and write layered TIFF ImageSourceData and ImageResources tags

Psdtags is a Python library to read and write the Adobe Photoshop(r) specific
ImageResources (#34377) and ImageSourceData (#37724) TIFF tags, which contain
image resource blocks, layer and mask information found in a typical layered
TIFF file created by Photoshop.

The format is specified in the
Adobe Photoshop TIFF Technical Notes (March 22, 2002) <https://download.osgeo.org/libtiff/doc/TIFFphotoshop.pdf>_
and
Adobe Photoshop File Formats Specification (November 2019) <https://www.adobe.com/devnet-apps/photoshop/fileformatashtml/>_.

Adobe Photoshop is a registered trademark of Adobe Systems Inc.

:Author: Christoph Gohlke <https://www.cgohlke.com>_
:License: BSD-3-Clause
:Version: 2025.5.10
:DOI: 10.5281/zenodo.7879187 <https://doi.org/10.5281/zenodo.7879187>_

Quickstart

Install the psdtags package and all dependencies from the
Python Package Index <https://pypi.org/project/psdtags/>_::

python -m pip install -U psdtags[all]

View the layer image and metadata stored in a layered TIFF file::

python -m psdtags file.tif

See Examples_ for using the programming interface.

Source code, examples, and support are available on
GitHub <https://github.com/cgohlke/psdtags>_.

Requirements

This revision was tested with the following requirements and dependencies
(other versions may work):

CPython <https://www.python.org>_ 3.10.11, 3.11.9, 3.12.10, 3.13.3 64-bit
NumPy <https://pypi.org/project/numpy/>_ 2.2.5
Imagecodecs <https://pypi.org/project/imagecodecs/>_ 2025.3.30
(required for compressing/decompressing image data)
Tifffile <https://pypi.org/project/tifffile/>_ 2025.5.10
(required for reading/writing tags from/to TIFF files)
Matplotlib <https://pypi.org/project/matplotlib/>_ 3.10.3
(required for plotting)

Revisions

2025.5.10

Support Python 3.14.

2025.1.1

Improve type hints.
Support Python 3.13.

2024.5.24

Fix docstring examples not correctly rendered on GitHub.

2024.2.22

Fix reading PsdBoolean (#10).
Fix order of PsdReferencePoint coordinates (breaking).
Allow reading unaligned PsdLayer blending_ranges.

2024.1.15

Fix multi-threading.

2024.1.8

Add option to compress layer channels in multiple threads.
Improve logging.
Drop support for Python 3.9 and numpy < 1.23 (NEP29).

2023.8.24

Fix channel data in layer and pattern blocks must be in big-endian order.

2023.6.15

Use PsdThumbnailFormat enum for PsdThumbnailBlock.format.

2023.4.30

Few API changes (breaking).
Improve object repr.
Drop support for Python 3.8 and numpy < 1.21 (NEP29).

2023.2.18

Allow unknown PsdKeys (#5).

2023.2.8

Change PsdPoint and PsdReferencePoint signatures (breaking).
Add helper function to create composite from layers.

2022.8.25

Update metadata.

2022.2.11

Fix struct padding.
Support TiffImageResources.

2022.2.2

Various API changes (breaking).
Handle additional layer information.
Preserve structures of unknown format as opaque bytes.
Add options to skip tag structures of unknown format.
Add abstract base class for tag structures.
Add classes for many structures.

2022.1.18

Various API changes (breaking).
Various fixes for writing TiffImageSourceData.
Support filter masks.
Add option to change channel compression on write.
Warn when skipping ResourceKey sections.

2022.1.14

Initial release.

Notes

The API is not stable yet and might change between revisions.

This library has been tested with a limited number of files only.

Additional layer information is not yet supported.

Consider psd-tools <https://github.com/psd-tools/psd-tools>_ and
pytoshop <https://github.com/mdboom/pytoshop>_ for working with
Adobe Photoshop PSD files.

Layered TIFF files can be read or written by Photoshop, Affinity Photo, and
Krita.

See also Reading and writing a Photoshop TIFF <https://www.amyspark.me/blog/posts/2021/11/14/reading-and-writing-tiff-psds.html>_.

Examples

Read the ImageSourceData tag value from a layered TIFF file and iterate over
all the channels:

isd = TiffImageSourceData.fromtiff('layered.tif')
for layer in isd.layers:
... layer.name
... for channel in layer.channels:
... ch = channel.data # a numpy array
...
'Background'
'Reflect1'
'Reflect2'
'image'
'Layer 1'
'ORight'
'I'
'IShadow'
'O'

Read the ImageResources tag value from the TIFF file, iterate over the blocks,
and get the thumbnail image:

res = TiffImageResources.fromtiff('layered.tif')
for block in res.blocks:
... blockname = block.name
...

res.thumbnail().shape
(90, 160, 3)

Write the image, ImageSourceData and ImageResources to a new layered TIFF file:

from tifffile import imread, imwrite
image = imread('layered.tif')
imwrite(
... '_layered.tif',
... image,
... byteorder=isd.byteorder, # must match ImageSourceData
... photometric='rgb', # must match ImageSourceData
... metadata=None, # do not write any tifffile specific metadata
... extratags=[isd.tifftag(maxworkers=4), res.tifftag()],
... )

Verify that the new layered TIFF file contains readable ImageSourceData:

assert isd == TiffImageSourceData.fromtiff('_layered.tif')
assert res == TiffImageResources.fromtiff('_layered.tif')

View the layer and mask information as well as the image resource blocks in
a layered TIFF file from a command line::

python -m psdtags layered.tif

Refer to the layered_tiff.py example in the source distribution for
creating a layered TIFF file from individual layer images.

Dependencies of psdtags (4).
Dependency	Optional
numpy	false
imagecodecs	true
matplotlib	true
tifffile	true

Package profile

psdtags

Releases

Release table

PyPI Downloads

Dependencies

Details

Project URLs

Dependencies

Package description