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
Converters #47
Converters #47
Conversation
We can introduce this command later!
* Introducing Scale Pixels
Conflicts: src/main/java/net/imagej/ops/normalize/NormalizeImgRT.java src/main/java/net/imagej/ops/scalepixel/ScalePixelsIterableRT.java
* Remove RealType restriction on generics
We need to be more specific about what we want to create. the assumption that the OpService actually can figure out what to create given the input parameters may not always hold.
this implementation is easier to understand and even more flexible
@hinerm's work on the convert-numbers branch (see also #54) may impact this work. From our planning doc: "Could also have SJC converters that convert whole images, as well as single element type objects." |
@dietzc Unless you want feedback sooner, I will wait on this one till the January hackathon. |
@ctrueden thats completely fine. I want to go over this stuff anyway as soon as the automatic "op-converters" are available (hackathon?!). |
needs more polishing! |
REVIEW NEEDED
@dscho I tried to implement Normalizers and Converters in a clean way, especially such that we can optimize each and every op. Can you please review, comment, etc on the design. I'm aware of the fact that the commit messages aren't perfect, but I really wanted to see progress while I'm here and we really have to discuss design issues. (I will clean up the history later of course :-))
You can start reviewing the design by following the examples provided in the ConvertIterableTest I guess.
Anyway: We also need to figure out a way to auto-generated e.g. fast variants of GenericScaleIterableRT (ArrayImg/PlanarImg etc, for all the type combinations). We will see this pattern occur more frequently: Given one or two ArrayImgs of same or different type apply a pixel-wise function on them. Maybe we can have a more generic way for creating such classes? any idea?
to make this clear: this is not intended to be merged. I will redo the commit history after we agreed if the design is OK. I'm really nitpicking on the design here because I have no idea if this is good or not. Please simply start from the provided test-case and have a look.