govips VS imaging

We chose to use govips which is a cgo wrapper around the libvips image manipulation library. The majority of new development for services in our backend is written using baseplate.go. But Go is not an ideal choice for media processing as it cannot keep up with the performance of native code. The most widely used image-processing libraries like libmagick are primarily written in C or C++. Speed was a major factor in selecting libvips in order to keep latency low on CDN cache misses for images. In our tests, libvips was 3–4 times faster than libmagick on basic image processing operations. Content-aware smart cropping was implemented by porting smartcrop.js to Go. This is the only operation implemented in pure Go.

Hi! You are willing to dive into some cgo we have a couple of feature requests :) https://github.com/davidbyttow/govips

There are a number of projects using manipulation libraries like https://github.com/disintegration/imaging which is already MIT licensed, and then there are various transcoders which I am unfamiliar with, but you will want to consider if you add enough value to make use of those unimportant to the decision to pay.

Sorry for not putting this code in my post. The input isn't from an *os.File, but rather from HTTP request captured by gin and it's on MultipartForm.File. The reason being why I had to do all these shenanigans is because I have to preprocess the image using this module: https://github.com/disintegration/imaging

Imaging