You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In this minor release, we update default file/directory/... naming to use "loris" instead of "loris2". Please check your configuration when you deploy this new release.
This release drops support for Python 2.7, which went out-of-support at the end of 2019. Loris is now Python 3 only.
Features:
Loris now supports downloading images as tif. To enable this, add "tif" to your transforms.target_formats setting. Thanks @cbutcosk for the patch! (#427)
Loris now allows decimal points in the size parameter, e.g. pct:12.5. Thanks @lutzhelm for the patch! (#445)
Loris uses the image identifier to set a Content-Disposition header in responses. This will cause browsers to download images with a more meaningful filename. (#476)
Bugfixes:
Loris no longer hangs if the Kakadu or OpenJPEG transformation process gets stuck. (#442)
There are some changes to the way caches are laid out. Caches will need to be rebuilt, but this should happen automatically as users request images from Loris.
Previously, trying to install Loris with pip 10 would fail with an ImportError; this allows installing Loris with pip 10. If you’re not using pip 10, there is no reason to upgrade.
Previously, if you tried to request a PNG with transparency as a non-transparent format (e.g. JPEG), Loris would return a 500 error. Now it converts the image to RGB before returning. Thanks to @dpancic for reporting the issue.
Not completely IIIF Image API 2.1 compliant (no support for max and related features), this release includes several bugfixes and new efficiencies in routing and resolving. Thanks to @bcail and 19 others! for the heavy lifting!.
The is a pre-release that adds several features to the resolvers, and a handful of bugfixes. It does not implement all IIIF 2.1 features yet--don't be fooled by the version number.