Benin Cashews: radiant mlhub -> source cooperative

tests/data/technoserve-cashew-benin/data.py

@@ -0,0 +1,57 @@
+#!/usr/bin/env python3
+
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License.
+
+import os
+
+import numpy as np
+import rasterio
+from rasterio.crs import CRS
+from rasterio.transform import Affine
+
+DTYPE = np.uint16
+SIZE = 2
+
+np.random.seed(0)
+
+dates = ('00_20191105',)
+all_bands = (
+    'B01',
+    'B02',
+    'B03',
+    'B04',
+    'B05',
+    'B06',
+    'B07',
+    'B08',
+    'B8A',
+    'B09',
+    'B11',
+    'B12',
+    'CLD',
+)
+profile = {
+    'driver': 'GTiff',
+    'dtype': DTYPE,
+    'width': SIZE,
+    'height': SIZE,
+    'count': 1,
+    'crs': CRS.from_epsg(32631),
+    'transform': Affine(
+        10.002549584378608,
+        0.0,
+        440853.29890114715,
+        0.0,
+        -9.99842989423825,
+        1012804.082877621,
+    ),
+}
+
+for date in dates:
+    os.makedirs(os.path.join('imagery', '00', date), exist_ok=True)
+    for band in all_bands:
+        Z = np.random.randint(np.iinfo(DTYPE).max, size=(SIZE, SIZE), dtype=DTYPE)
+        path = os.path.join('imagery', '00', date, f'{date}_{band}_10m.tif')
+        with rasterio.open(path, 'w', **profile) as src:
+            src.write(Z, 1)

tests/data/technoserve-cashew-benin/labels/00.geojson

@@ -0,0 +1,8 @@
+{
+"type": "FeatureCollection",
+"name": "cashew_benin",
+"crs": { "type": "name", "properties": { "name": "urn:ogc:def:crs:EPSG::32631" } },
+"features": [
+{ "type": "Feature", "properties": { "OBJECTID": 1, "class": 1, "Shape_Leng": 367629.52331100003, "Shape_Area": 16997542.377500001, "class_name": "Well-managed planatation" }, "geometry": { "type": "Polygon", "coordinates": [ [ [ 447131.214800000190735, 1001286.6359 ], [ 447166.272199999541044, 1001285.31299999915 ], [ 447196.037899999879301, 1001285.31299999915 ], [ 447244.324400000274181, 1001283.3286 ], [ 447256.230700000189245, 1001282.00569999963 ], [ 447253.58490000013262, 1001248.9327 ], [ 447254.907800000160933, 1001228.4275 ], [ 447252.923700000159442, 1001212.05179999955 ], [ 447087.555899999978, 1001212.333799999207 ], [ 447082.266800000332296, 1001241.656599999988 ], [ 447076.97510000038892, 1001256.208799999207 ], [ 447074.329300000332296, 1001286.6359 ], [ 447131.214800000190735, 1001286.6359 ] ] ] } }
+]
+}

tests/datasets/test_benin_cashews.py

@@ -1,9 +1,7 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import glob
 import os
-import shutil
 from pathlib import Path
 
 import matplotlib.pyplot as plt
@@ -18,44 +16,22 @@
     DatasetNotFoundError,
     RGBBandsMissingError,
 )
-
-
-class Collection:
-    def download(self, output_dir: str, **kwargs: str) -> None:
-        glob_path = os.path.join('tests', 'data', 'ts_cashew_benin', '*.tar.gz')
-        for tarball in glob.iglob(glob_path):
-            shutil.copy(tarball, output_dir)
-
-
-def fetch(dataset_id: str, **kwargs: str) -> Collection:
-    return Collection()
+from torchgeo.datasets.utils import Executable
 
 
 class TestBeninSmallHolderCashews:
     @pytest.fixture
     def dataset(
-        self, monkeypatch: MonkeyPatch, tmp_path: Path
+        self, azcopy: Executable, monkeypatch: MonkeyPatch, tmp_path: Path
     ) -> BeninSmallHolderCashews:
-        radiant_mlhub = pytest.importorskip('radiant_mlhub', minversion='0.3')
-        monkeypatch.setattr(radiant_mlhub.Collection, 'fetch', fetch)
-        source_md5 = '255efff0f03bc6322470949a09bc76db'
-        labels_md5 = 'ed2195d93ca6822d48eb02bc3e81c127'
-        monkeypatch.setitem(BeninSmallHolderCashews.image_meta, 'md5', source_md5)
-        monkeypatch.setitem(BeninSmallHolderCashews.target_meta, 'md5', labels_md5)
-        monkeypatch.setattr(BeninSmallHolderCashews, 'dates', ('2019_11_05',))
+        url = os.path.join('tests', 'data', 'technoserve-cashew-benin')
+        monkeypatch.setattr(BeninSmallHolderCashews, 'url', url)
+        monkeypatch.setattr(BeninSmallHolderCashews, 'dates', ('20191105',))
+        monkeypatch.setattr(BeninSmallHolderCashews, 'tile_height', 2)
+        monkeypatch.setattr(BeninSmallHolderCashews, 'tile_width', 2)
         root = str(tmp_path)
         transforms = nn.Identity()
-        bands = BeninSmallHolderCashews.all_bands
-
-        return BeninSmallHolderCashews(
-            root,
-            transforms=transforms,
-            bands=bands,
-            download=True,
-            api_key='',
-            checksum=True,
-            verbose=True,
-        )
+        return BeninSmallHolderCashews(root, transforms=transforms, download=True)
 
     def test_getitem(self, dataset: BeninSmallHolderCashews) -> None:
         x = dataset[0]
@@ -66,25 +42,22 @@ def test_getitem(self, dataset: BeninSmallHolderCashews) -> None:
         assert isinstance(x['y'], torch.Tensor)
 
     def test_len(self, dataset: BeninSmallHolderCashews) -> None:
-        assert len(dataset) == 72
+        assert len(dataset) == 1
 
     def test_add(self, dataset: BeninSmallHolderCashews) -> None:
         ds = dataset + dataset
         assert isinstance(ds, ConcatDataset)
-        assert len(ds) == 144
+        assert len(ds) == 2
 
     def test_already_downloaded(self, dataset: BeninSmallHolderCashews) -> None:
-        BeninSmallHolderCashews(root=dataset.root, download=True, api_key='')
+        BeninSmallHolderCashews(root=dataset.root, download=True)
 
     def test_not_downloaded(self, tmp_path: Path) -> None:
         with pytest.raises(DatasetNotFoundError, match='Dataset not found'):
             BeninSmallHolderCashews(str(tmp_path))
 
     def test_invalid_bands(self) -> None:
         with pytest.raises(AssertionError):
-            BeninSmallHolderCashews(bands=['B01', 'B02'])  # type: ignore[arg-type]
-
-        with pytest.raises(ValueError, match='is an invalid band name.'):
             BeninSmallHolderCashews(bands=('foo', 'bar'))
 
     def test_plot(self, dataset: BeninSmallHolderCashews) -> None: