-
Notifications
You must be signed in to change notification settings - Fork 77
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
New CTL programs #144
Comments
Passing a scalar or vector and getting it transformed on stdout would be great yeah! |
@aforsythe I'm getting a github 404 on your aces-unit-test repo page |
@aforsythe @KelSolaar do you imagine this functionality should be part of a separate program independent of ctlrender or as alternate input/output path added to ctlrender, or something else ? |
@michaeldsmith Sorry the repo was private. Public now. re: stdout ... I'm open. I can see arguments both ways. |
Oh sorry, I was not clear, yes on existing |
I tend to agree ... thinking about it, the command line interface we'd want would probably be nearly identical to ctlrender, so why not just build on top of that. |
I've come across a few scenarios where having programs capable of tasks not currently supported by the built-in tools in CTL would be highly advantageous.
The first scenario involves the desire to input a scalar, vector, or array of numbers directly into CTL and observe the numerical output. At present, we utilize CTL through Matlab by creating a temporary image, processing it with ctlrender, and then importing the results back into Matlab. I believe @KelSolaar employs a similar method with colour-science. Ideally, a program that accepts numerical data directly might be more efficient, although I acknowledge there might be intricate details that need to be addressed.
The second scenario is the need to execute a CTL program without feeding an image into it. This functionality would be particularly beneficial for conducting unit tests, such as the ones I have outlined here: https://github.com/ampas/aces-unittest-test. Currently, my approach involves using ctlrender, even though the image data itself isn't being utilized, which seems like an inelegant solution.
I'd be interested in hearing your opinions. Thanks.
The text was updated successfully, but these errors were encountered: