I would like to know what tools are currently available to optimize images for web in Haiku, especially to reduce their file size. I have already found a couple command-line utilities in HaikuDepot, which I still haven’t had the time to try, but would like to know which ones you use, if you are using any at all, of course.
On macOS, I have settled with JPEGmini and ImageOptim (which in turn allows combine the use of several standalone utilities like pngcrush). On iOS, from which I currently build and publish my Pelican based website, I ended up creating my own command-line utility in Python:
It allows to automate some common tasks like reducing width/height, reducing the number of colors in PNG images, converting PNG to JPEG, and saving optimized versions (as far as Pillow allows) of those files. It works on a single image or in a folder of images.
What do you people use on Haiku? Is there anything like ImageOptim?
In any case, you will probably end using that, cimg, higher level options like opencv or some weird libs (for me).
Btw, last time i had to do mass operations with images i used pillow to load, process and compact ~60k images into a binary blob for neural networks… Seems like you always end using python + some image lib in C.
Right. As I said in another thread around here, Python is a cool language, provided that the ecosystem is available for this kind of tasks, Python allows us to have a higher level abstraction and still benefit from the speed and technical expertise coming from compiled packages like Pillow.
optipng and pngcrush are in HaikuDepot.
There’s also a OptPNGTranslator that uses optipng. Install it and every native app can save optimized PNGs. Neato!
(@lelldorin: Can you make sure to have the [quote] and [/quote] in your replies are on their own line? Otherwise the quoting doesn’t quite work and I have to edit your post… )
Here are the set of tools it interfaces with (macOS version:
Basically, it’s a window that accepts a list of files into it and then there is a minimal interface that has very few configurations, allowing to use these tools without any complication if they are available in the system. I believe some of them were provided by the installer itself, but not entirely sure.
for this kind of tasks, Python allows us to have a higher level abstraction and still benefit from the speed and technical expertise coming from compiled packages like Pillow.
