Add own data to CDL dataset. #429
Conversation
|
Hmmm, I think we should add an explicit test for the cmap to get around this test coverage thing. Update: still not quite sure why test coverage would drop, but #430 should fix this. |
Coverage is dropping because the new fake data does not have a cmap. We should add this to the data. Also, it seems like the actual |
| @@ -58,6 +58,8 @@ class CDL(RasterDataset): | |||
| (2008, "e6aa3967e379b98fd30c26abe9696053"), | |||
| ] | |||
|
|
|||
| cmap: Dict[int, Tuple[int, int, int, int]] = {} | |||
There was a problem hiding this comment.
This is already defined at the RasterDataset level, I don't think we need it here too
| @@ -31,9 +31,21 @@ def dataset( | |||
| monkeypatch.setattr( # type: ignore[attr-defined] | |||
| torchgeo.datasets.cdl, "download_url", download_url | |||
| ) | |||
| cmap = { | |||
There was a problem hiding this comment.
This cmap should be added directly to the file so we can test that TorchGeo can read the cmap from a file. See https://github.com/microsoft/torchgeo/tree/main/tests/data#raster-data for the code to add this to the file.
|
|
||
|
|
||
| directories = ["2020_30m_cdls", "2021_30m_cdls"] | ||
| raster_extensions = [[".tif", ".tif.ovr"], [".tif", ".tif.ovr"]] |
There was a problem hiding this comment.
Can probably simplify this to a single list, no need to duplicate it
* add own data * data.py with cmap * remove .img test data format
* add own data * data.py with cmap * remove .img test data format
As requested in #415 , this PR adds a data.py and new fake data.