0.5.0 release

.github/workflows/release.yaml

@@ -22,7 +22,7 @@ jobs:
       id: cache
       with:
         path: ${{ env.pythonLocation }}
-        key: ${{ env.pythonLocation }}-${{ hashFiles('setup.cfg') }}
+        key: ${{ env.pythonLocation }}-${{ hashFiles('pyproject.toml') }}
     - name: Install pip dependencies
       if: steps.cache.outputs.cache-hit != 'true'
       run: pip install .[tests]
@@ -48,7 +48,7 @@ jobs:
       id: cache
       with:
         path: ${{ env.pythonLocation }}
-        key: ${{ env.pythonLocation }}-${{ hashFiles('setup.cfg') }}
+        key: ${{ env.pythonLocation }}-${{ hashFiles('pyproject.toml') }}
     - name: Install pip dependencies
       if: steps.cache.outputs.cache-hit != 'true'
       run: pip install .[datasets,tests]
@@ -74,7 +74,7 @@ jobs:
       id: cache
       with:
         path: ${{ env.pythonLocation }}
-        key: ${{ env.pythonLocation }}-${{ hashFiles('setup.cfg') }}
+        key: ${{ env.pythonLocation }}-${{ hashFiles('pyproject.toml') }}
     - name: Install pip dependencies
       if: steps.cache.outputs.cache-hit != 'true'
       run: pip install .[docs,tests] planetary_computer pystac

.github/workflows/tutorials.yaml

@@ -26,7 +26,7 @@ jobs:
       id: cache
       with:
         path: ${{ env.pythonLocation }}
-        key: ${{ env.pythonLocation }}-${{ hashFiles('setup.cfg') }}
+        key: ${{ env.pythonLocation }}-${{ hashFiles('pyproject.toml') }}
     - name: Install pip dependencies
       if: steps.cache.outputs.cache-hit != 'true'
       run: pip install .[docs,tests] planetary_computer pystac

docs/tutorials/getting_started.ipynb

@@ -189,6 +189,7 @@
    "outputs": [],
    "source": [
     "chesapeake_root = os.path.join(tempfile.gettempdir(), \"chesapeake\")\n",
+    "os.makedirs(chesapeake_root, exist_ok=True)\n",
     "chesapeake = ChesapeakeDE(chesapeake_root, crs=naip.crs, res=naip.res, download=True)"
    ]
   },

tests/datasets/test_advance.py

@@ -81,6 +81,7 @@ def test_mock_missing_module(
             dataset[0]
 
     def test_plot(self, dataset: ADVANCE) -> None:
+        pytest.importorskip("scipy", minversion="1.6.2")
         x = dataset[0].copy()
         dataset.plot(x, suptitle="Test")
         plt.close()

tests/datasets/test_skippd.py

@@ -18,6 +18,8 @@
 import torchgeo.datasets.utils
 from torchgeo.datasets import SKIPPD
 
+pytest.importorskip("h5py", minversion="3")
+
 
 def download_url(url: str, root: str, *args: str, **kwargs: str) -> None:
     shutil.copy(url, root)

tests/trainers/test_detection.py

@@ -17,6 +17,9 @@
 from torchgeo.main import main
 from torchgeo.trainers import ObjectDetectionTask
 
+# MAP metric requires pycocotools to be installed
+pytest.importorskip("pycocotools")
+
 
 class PredictObjectDetectionDataModule(NASAMarineDebrisDataModule):
     def setup(self, stage: str) -> None:

tests/trainers/test_moco.py

@@ -128,20 +128,16 @@ def test_weight_str(self, mocked_weights: WeightsEnum) -> None:
 
     @pytest.mark.slow
     def test_weight_enum_download(self, weights: WeightsEnum) -> None:
-        match = "num classes .* != num classes in pretrained model"
-        with pytest.warns(UserWarning, match=match):
-            MoCoTask(
-                model=weights.meta["model"],
-                weights=weights,
-                in_channels=weights.meta["in_chans"],
-            )
+        MoCoTask(
+            model=weights.meta["model"],
+            weights=weights,
+            in_channels=weights.meta["in_chans"],
+        )
 
     @pytest.mark.slow
     def test_weight_str_download(self, weights: WeightsEnum) -> None:
-        match = "num classes .* != num classes in pretrained model"
-        with pytest.warns(UserWarning, match=match):
-            MoCoTask(
-                model=weights.meta["model"],
-                weights=str(weights),
-                in_channels=weights.meta["in_chans"],
-            )
+        MoCoTask(
+            model=weights.meta["model"],
+            weights=str(weights),
+            in_channels=weights.meta["in_chans"],
+        )

tests/trainers/test_regression.py

@@ -66,6 +66,9 @@ def create_model(*args: Any, **kwargs: Any) -> Module:
     def test_trainer(
         self, monkeypatch: MonkeyPatch, name: str, fast_dev_run: bool
     ) -> None:
+        if name == "skippd":
+            pytest.importorskip("h5py", minversion="3")
+
         config = os.path.join("tests", "conf", name + ".yaml")
 
         monkeypatch.setattr(timm, "create_model", self.create_model)

tests/trainers/test_simclr.py

@@ -126,20 +126,16 @@ def test_weight_str(self, mocked_weights: WeightsEnum) -> None:
 
     @pytest.mark.slow
     def test_weight_enum_download(self, weights: WeightsEnum) -> None:
-        match = "num classes .* != num classes in pretrained model"
-        with pytest.warns(UserWarning, match=match):
-            SimCLRTask(
-                model=weights.meta["model"],
-                weights=weights,
-                in_channels=weights.meta["in_chans"],
-            )
+        SimCLRTask(
+            model=weights.meta["model"],
+            weights=weights,
+            in_channels=weights.meta["in_chans"],
+        )
 
     @pytest.mark.slow
     def test_weight_str_download(self, weights: WeightsEnum) -> None:
-        match = "num classes .* != num classes in pretrained model"
-        with pytest.warns(UserWarning, match=match):
-            SimCLRTask(
-                model=weights.meta["model"],
-                weights=str(weights),
-                in_channels=weights.meta["in_chans"],
-            )
+        SimCLRTask(
+            model=weights.meta["model"],
+            weights=str(weights),
+            in_channels=weights.meta["in_chans"],
+        )

torchgeo/__init__.py

@@ -11,4 +11,4 @@
 """
 
 __author__ = "Adam J. Stewart"
-__version__ = "0.5.0.dev0"
+__version__ = "0.5.0"

torchgeo/datamodules/chesapeake.py

@@ -93,7 +93,7 @@ def __init__(
         """
         # This is a rough estimate of how large of a patch we will need to sample in
         # EPSG:3857 in order to guarantee a large enough patch in the local CRS.
-        self.original_patch_size = patch_size * 2
+        self.original_patch_size = patch_size * 3
         kwargs["transforms"] = _Transform(K.CenterCrop(patch_size))
 
         super().__init__(