Skip to content
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

Implement data interface for Raster types #546

Open
2 tasks done
philippjfr opened this issue Mar 8, 2016 · 2 comments
Open
2 tasks done

Implement data interface for Raster types #546

philippjfr opened this issue Mar 8, 2016 · 2 comments
Assignees
Milestone

Comments

@philippjfr
Copy link
Member

philippjfr commented Mar 8, 2016

In issue #542 we introduce a general API for N-dimensional, dense/gridded data. This provides the opportunity to adapt our existing Raster Element to use these dense data interfaces, which will allow these types to support various out-of-memory array storage formats including xarray, dask and iris directly and unify conversions between sparse and dense Element types.

Development of this feature should roughly follow these steps:

  • Implement raw array based interfaces for backward compatibility.
  • Update plotting code and operations to use new API rather than using .data directly
@philippjfr philippjfr added this to the v1.5.0 milestone Mar 8, 2016
@philippjfr philippjfr modified the milestones: v1.6.0, v1.5.0 Apr 20, 2016
@philippjfr philippjfr self-assigned this Mar 4, 2017
@philippjfr philippjfr modified the milestones: v2.0, v1.7.0 Mar 15, 2017
@philippjfr
Copy link
Member Author

philippjfr commented Apr 14, 2017

This has now to a large part been implemented but Raster itself is still a straggler and has not been converted. My suggestion for 2.0 is the following, the Raster Element will simply use the GridInterface, XArrayInterface and IrisInterface, when passed a raw array indices are added and it will use one of those interfaces to store the data. This preserves the semantics of Raster, i.e. integer indexed by default, but also provides a type that does not assume anything about the sampling of the data, which can be useful because Image and QuadMesh do and you might just want something quick and dirty.

@jlstevens
Copy link
Contributor

Sounds like a reasonable suggestion to me.

@philippjfr philippjfr modified the milestones: v2.0, v1.11 May 25, 2018
@philippjfr philippjfr modified the milestones: v1.11.0, v1.12.0 Dec 4, 2018
@philippjfr philippjfr modified the milestones: v1.12.0, v1.13.0 Mar 22, 2019
@philippjfr philippjfr modified the milestones: v1.13.0, v1.13.x Mar 3, 2020
@philippjfr philippjfr modified the milestones: v1.14.x, v2.0 May 23, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants