jcupitt
Joined 12 years ago
Joined 12 years ago
https://www.rubydoc.info/gems/ruby-vips/Vips.block_untrusted
You can also set the env var `VIPS_BLOCK_UNTRUSTED`, which might be easier.
You can block or allow specific load operations, so you can limit format support to just the types you need:
https://www.rubydoc.info/gems/ruby-vips/Vips#block-class_met...
That might be even better.
There was a post on libvips.org about this a while ago:
https://www.libvips.org/2022/05/28/What's-new-in-8.13.html
And a note about it in the checklist for devs:
https://www.libvips.org/API/current/developer-checklist.html...
I've found jemalloc works best for long running libvips processes, fwiw.
- I'd consider adding support for libvips smartcrop, ICC colour management, HDR, image alpha
- can you join pipelines? you could have one making a background, one making a foreground, then composite them, for example
- your benchmarks don't measure peak memory use, it can be a useful thing to test, especially if people are processing large images (eg. the output of every modern smartphone sigh)
- have you benchmarked against nip4, the official libvips GUI? it has a CLI mode you can run with GNU parallel
- I agree, a CLI mode for your tool sounds useful
https://sharp.pixelplumbing.com/install/#prebuilt-binaries
it can sometimes need to compile the C++ shim that sits between node and libvips, but that's rare.
https://sharp.pixelplumbing.com/install/#prebuilt-binaries
It can usually also download a precompiled binary for the C++ shim that sits between node and libvips, but if your node / arch / etc. is not supported, it'll compile that (that's what the build.js file you linked does).
image = pyvips.Image.new_from_file(filename, access="sequential")
scale = min(200 / image.width, 200 / image.height)
thumbnail = image.colourspace("oklab").resize(scale).extract_bands(0, n=3)
rgbf = thumbnail.write_to_memory()
That'll stream the source image and make a RGBRGBRGB memory buffer of single precision floats. You could perhaps use kernel="linear" and avoid any ringing from lanczos3.
I think I would downsample in a linear light space, like scRGB. Averaging there means averaging photons, which will surely be better than OKLab. Maybe switch to OKLab for clustering. Though of course I've not tested it.
It includes support for UltraHDR (HDR and SDR in one JPEG file), camera RAW images, and the Oklab colourspace. This should all be coming to sharp in the next six month or so.
Commercial dev is about results and money, and that makes AI almost impossible to resist.
OSS has always been about process. The pleasure is in tinkering with code, collaborating with friends on a fun new idea, and helping people. I don't care about speed.