Open astrofrog opened 12 years ago
@ChrisBeaumont - do you think this would be useful? If so, I'm happy to implement this.
hmm, I'm inclined to think that the dendrogram itself should be expressed in pixel units, and the WCS gets used when building the catalog of extracted properties (i.e. code that doesn't yet exist in Python). But this might be my bias towards the way things currently work in IDL.
This would be a convenience on the generated dendrogram, not something used at compute time - i.e. once you have a Leaf, you can optionally ask for world coordinates (but coords
by default just returns pixel coordinates). This is just an idea though, I don't feel too strongly about it, but I've had to help several users convert the pixel coordinates to world coordinates, and it seems that it's something that would be useful to have as actual properties of the dendrogram components.
My feelings are still mixed, since I think this is less generic than what a dendrogram fundamentally is. I think that the as-yet-unwritten property cataloging module cover this and many other conversions to meaningful units.
We should start a discussion of what the basic structure analysis API should look like, and then decide whether we still need a one-off, leaf.world_coords attribute
You're right, we should think first about the analysis API. Related to this issue, we could consider having sub-classes in future for specific cases, e.g. FITSDendrogram which are WCS aware, while the generic class is not.
I've created a basic page for this here https://github.com/dendrograms/dendro-core/wiki/Structure-analysis-API-proposal
On Thu, Jun 13, 2013 at 9:38 AM, Thomas Robitaille <notifications@github.com
wrote:
You're right, we should think first about the analysis API. Related to this issue, we could consider having sub-classes in future for specific cases, e.g. FITSDendrogram which are WCS aware, while the generic class is not.
— Reply to this email directly or view it on GitHubhttps://github.com/dendrograms/dendro-core/issues/4#issuecomment-19392103 .
Chris Beaumont Graduate Student Institute for Astronomy University of Hawaii at Manoa 2680 Woodlawn Drive Honolulu, HI 96822 www.ifa.hawaii.edu/~beaumont
It would be nice to allow a WCS object to be passed for the data, and be able to return the world coordinates of elements inside structures - e.g. the world coordinates of all voxels inside a leaf from a 3-d cube. This could use a different attribute to what would normally be used to get the pixel coordinates (e.g.
leaf.world_coords
instead ofleaf.coords
, or something like this).