Center the coordinates to pixels for rasterio backend #1468
Conversation
Rasterio uses edge-based coordinates, which contradict the treatment of coordinates in xarray as being centered over the pixels. This centers them with an offset of half the resolution.
|
Looks good to me, and very important (for those of us who care about half a pixel! :-)... |
|
Thanks for the PR. I'll let @fmaussion comment on the substance here but given the confusion and the change in behavior, let's add a note in |
|
Thanks for the feedback @jhamman - just added the documentation. |
doc/whats-new.rst
Outdated
| @@ -21,6 +21,10 @@ v0.9.7 (unreleased) | |||
| Enhancements | |||
| ~~~~~~~~~~~~ | |||
|
|
|||
| - :py:func:`~xarray.open_rasterio` method now shifts the rasterio | |||
| coordinates so that they are centered in each pixel. | |||
| By `Greg Brener <https://github.com/gbrener>`_. | |||
There was a problem hiding this comment.
This change should rather appear in "Bug fixes" I think
| file's geoinformation. | ||
| file's geoinformation, shifted to the center of each pixel (see | ||
| `"PixelIsArea" Raster Space | ||
| <http://web.archive.org/web/20160326194152/http://remotesensing.org/geotiff/spec/geotiff2.5.html#2.5.2>`_ |
There was a problem hiding this comment.
Thanks for looking into this. Isn't there a better ref than on the web archive?
There was a problem hiding this comment.
Happy to help @fmaussion . The web archive link is used by the GeoTIFF website (first link here: http://trac.osgeo.org/geotiff). A quick Google search didn't yield anything better.
xarray/backends/rasterio_.py
Outdated
| coords['y'] = np.linspace(start=y0, num=ny, stop=(y0 + (ny - 1) * dy)) | ||
| coords['x'] = np.linspace(start=x0, num=nx, stop=(x0 + (nx - 1) * dx)) | ||
| coords['y'] = np.linspace(start=y0 + dy/2, | ||
| num=ny, |
There was a problem hiding this comment.
This was just to satisfy flake8's line-length requirement. The three lines were for consistency, but I changed it (in my recent commit) to be two lines.
xarray/backends/rasterio_.py
Outdated
| num=ny, | ||
| stop=(y0 + (ny - 1) * dy) + dy/2) | ||
| coords['x'] = np.linspace(start=x0 + dx/2, | ||
| num=nx, |
There was a problem hiding this comment.
This was just to satisfy flake8's line-length requirement. The three lines were for consistency, but I changed it (in my recent commit) to be two lines.
|
Definitely a bad bug which could justify a minor release soon :-( |
|
Thanks for the feedback @fmaussion . I addressed your suggestions and made some small cleanups. Please let me know if there's anything else. |
|
I'm not sure we really have a cell centering convention for xarray, except for plotting routines, where we chose the cell-centered convention from CF-conventions. But this does seem like a marginal improvement, so +1 from me. In the future we may want to switch to describing bounds fully using |
|
+1 from me as well. |
|
Thanks! |
Rasterio uses edge-based coordinates, which is a different convention from how xarray treats coordinates (based on description here: http://xarray.pydata.org/en/stable/plotting.html#coordinates). This PR centers them, offsetting by half of the resolution.
git diff upstream/master | flake8 --diff- [ ] Closes #xxxxwhats-new.rstfor all changes andapi.rstfor new APICCing @fmaussion since he may be interested.