Skip to content

Added layers#183

Open
Karinon wants to merge 20 commits into
mainfrom
feat/textures
Open

Added layers#183
Karinon wants to merge 20 commits into
mainfrom
feat/textures

Conversation

@Karinon

@Karinon Karinon commented Jun 18, 2026

Copy link
Copy Markdown
Collaborator

This PR introduces the following features:

  • added layer functionality

    • the old Masks-Subsection with coastlines, lat-lon grids and land/sea masks has been replaced by a Layer subsection: All of the previous functionalities is now a layer on the layer stack. You can drag and drop layers how you like
    • The currently visualized data can be put as an image (not as data) into the layer stack (I am going to refer to it as snapshot in the following text for the sake of clarity). Afterwards you can load another dataset or variable and still keep the old snapshot in the stack.
    • The snapshot is stored in the browser session storage and should also survive a reload of the page
    • Snapshots can be exported as (WGS 84)-Geotiffs and re-imported again
    • The layer stack supports imports of (WGS 84)-Geotiffs, PNGs and JPGs. For PNGs and JPGs it works best when the images are equirectangular.
  • Dataset info sorts coordinates and variables into group-dropdowns if the dataset has groups.

  • added 10m coastlines and 15 degrees lat/lon-grids

regarding Geotiff Export:

One could look at like more or less like it is a "regridding": GeoTIFF export does not regrid regular lat/lon grids; those use a direct texture export path at the source texture’s bounds and resolution = 1 lat/lon = 1 pixel. For other grid types, export is effectively a visual rasterization into an equirectangular GeoTIFF, not a scientific regridding algorithm. The fallback path renders the visible grid geometry into an offscreen buffer, using source texture size when possible and otherwise defaulting geometry-backed exports to 8192x4096, clamped by GPU limits. Mesh grids are therefore sampled through GPU projection, rasterization, interpolation, and the active colormap. Irregular point grids are the weakest case because they export rendered point splats rather than true cells or an interpolated continuous field, so gaps, overlaps, and point size can affect the result.

@Karinon Karinon requested review from d70-t and wachsylon June 18, 2026 14:03
@wachsylon

Copy link
Copy Markdown
Collaborator

Just amazing.

@Karinon

Karinon commented Jun 19, 2026

Copy link
Copy Markdown
Collaborator Author

@d70-t:

@wachsylon and I are both on vacation for the next two weeks. We are not going merge this during this time. Feel free to test it out

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants