Merge branch 'main' into added_tutorial

docs/api/datamodules.rst

@@ -124,6 +124,11 @@ GID-15
 
 .. autoclass:: GID15DataModule
 
+HySpecNet-11k
+^^^^^^^^^^^^^
+
+.. autoclass:: HySpecNet11kDataModule
+
 Inria Aerial Image Labeling
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
@@ -213,6 +218,11 @@ SustainBench Crop Yield
 
 .. autoclass:: SustainBenchCropYieldDataModule
 
+TreeSatAI
+^^^^^^^^^
+
+.. autoclass:: TreeSatAIDataModule
+
 Tropical Cyclone
 ^^^^^^^^^^^^^^^^
 

docs/api/datasets.rst

@@ -312,6 +312,11 @@ GID-15
 
 .. autoclass:: GID15
 
+HySpecNet-11k
+^^^^^^^^^^^^^
+
+.. autoclass:: HySpecNet11k
+
 IDTReeS
 ^^^^^^^
 
@@ -469,6 +474,11 @@ SustainBench Crop Yield
 
 .. autoclass:: SustainBenchCropYield
 
+TreeSatAI
+^^^^^^^^^
+
+.. autoclass:: TreeSatAI
+
 Tropical Cyclone
 ^^^^^^^^^^^^^^^^
 

docs/api/datasets/non_geo_datasets.csv

@@ -21,6 +21,7 @@ Dataset,Task,Source,License,# Samples,# Classes,Size (px),Resolution (m),Bands
 `Forest Damage`_,OD,Drone imagery,"CDLA-Permissive-1.0","1,543",4,"1,500x1,500",,RGB
 `GeoNRW`_,S,Aerial,"CC-BY-4.0","7,783",11,"1,000x1,000",1,"RGB, DEM"
 `GID-15`_,S,Gaofen-2,-,150,15,"6,800x7,200",3,RGB
+`HySpecNet-11k`_,-,EnMAP,CC0-1.0,11k,-,128,30,HSI
 `IDTReeS`_,"OD,C",Aerial,"CC-BY-4.0",591,33,200x200,0.1--1,RGB
 `Inria Aerial Image Labeling`_,S,Aerial,-,360,2,"5,000x5,000",0.3,RGB
 `LandCover.ai`_,S,Aerial,"CC-BY-NC-SA-4.0","10,674",5,512x512,0.25--0.5,RGB
@@ -52,6 +53,7 @@ Dataset,Task,Source,License,# Samples,# Classes,Size (px),Resolution (m),Bands
 `SSL4EO-L Benchmark`_,S,Lansat & CDL,"CC0-1.0",25K,134,264x264,30,MSI
 `SSL4EO-L Benchmark`_,S,Lansat & NLCD,"CC0-1.0",25K,17,264x264,30,MSI
 `SustainBench Crop Yield`_,R,MODIS,"CC-BY-SA-4.0",11k,-,32x32,-,MSI
+`TreeSatAI`_,"C, R, S","Aerial, Sentinel-1/2",CC-BY-4.0,50K,"12, 15, 20","6, 20, 304","0.2, 10","CIR, MSI, SAR"
 `Tropical Cyclone`_,R,GOES 8--16,"CC-BY-4.0","108,110",-,256x256,4K--8K,MSI
 `UC Merced`_,C,USGS National Map,"public domain","2,100",21,256x256,0.3,RGB
 `USAVars`_,R,NAIP Aerial,"CC-BY-4.0",100K,-,-,4,"RGB, NIR"

pyproject.toml

@@ -61,9 +61,9 @@ dependencies = [
     # pyproj 3.3+ required for Python 3.10 wheels
     "pyproj>=3.3",
     # rasterio 1.3+ required for Python 3.10 wheels
-    # rasterio 1.4+ no longer supports merging WarpedVRT objects
+    # rasterio 1.4.0-1.4.2 lack support for merging WarpedVRT objects
     # https://github.com/rasterio/rasterio/issues/3196
-    "rasterio>=1.3,<1.4",
+    "rasterio>=1.3,!=1.4.0,!=1.4.1,!=1.4.2",
     # rtree 1+ required for Python 3.10 wheels
     "rtree>=1",
     # segmentation-models-pytorch 0.2+ required for smp.losses module

requirements/required.txt

@@ -5,14 +5,14 @@ setuptools==75.6.0
 einops==0.8.0
 fiona==1.10.1
 kornia==0.7.4
-lightly==1.5.14
+lightly==1.5.15
 lightning[pytorch-extra]==2.4.0
-matplotlib==3.9.2
+matplotlib==3.9.3
 numpy==2.1.3
 pandas==2.2.3
 pillow==11.0.0
 pyproj==3.7.0
-rasterio==1.3.11
+rasterio==1.4.3
 rtree==1.3.0
 segmentation-models-pytorch==0.3.4
 shapely==2.0.6

requirements/style.txt

@@ -1,3 +1,3 @@
 # style
 mypy==1.13.0
-ruff==0.8.0
+ruff==0.8.1

requirements/tests.txt

@@ -1,4 +1,4 @@
 # tests
 nbmake==1.5.4
-pytest==8.3.3
+pytest==8.3.4
 pytest-cov==6.0.0

tests/conf/hyspecnet_byol.yaml

@@ -0,0 +1,11 @@
+model:
+  class_path: BYOLTask
+  init_args:
+    model: 'resnet18'
+    in_channels: 202
+data:
+  class_path: HySpecNet11kDataModule
+  init_args:
+    batch_size: 2
+  dict_kwargs:
+    root: 'tests/data/hyspecnet'

tests/conf/hyspecnet_moco.yaml

@@ -0,0 +1,11 @@
+model:
+  class_path: MoCoTask
+  init_args:
+    model: 'resnet18'
+    in_channels: 202
+data:
+  class_path: HySpecNet11kDataModule
+  init_args:
+    batch_size: 2
+  dict_kwargs:
+    root: 'tests/data/hyspecnet'

tests/conf/hyspecnet_simclr.yaml

@@ -0,0 +1,11 @@
+model:
+  class_path: SimCLRTask
+  init_args:
+    model: 'resnet18'
+    in_channels: 202
+data:
+  class_path: HySpecNet11kDataModule
+  init_args:
+    batch_size: 2
+  dict_kwargs:
+    root: 'tests/data/hyspecnet'

tests/conf/treesatai.yaml

@@ -0,0 +1,13 @@
+model:
+  class_path: MultiLabelClassificationTask
+  init_args:
+    model: 'resnet18'
+    in_channels: 19
+    num_classes: 15
+    loss: 'bce'
+data:
+  class_path: TreeSatAIDataModule
+  init_args:
+    batch_size: 1
+  dict_kwargs:
+    root: 'tests/data/treesatai'

tests/data/hyspecnet/data.py

@@ -0,0 +1,61 @@
+#!/usr/bin/env python3
+
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License.
+
+import os
+import shutil
+
+import numpy as np
+import rasterio
+from rasterio import Affine
+from rasterio.crs import CRS
+
+SIZE = 32
+DTYPE = 'int16'
+
+np.random.seed(0)
+
+# Tile name purposefully shortened to avoid Windows git filename length limit.
+tiles = ['ENMAP01_20221103T162438Z']
+patches = ['Y01460273_X05670694', 'Y01460273_X06950822']
+
+profile = {
+    'driver': 'GTiff',
+    'dtype': DTYPE,
+    'nodata': -32768.0,
+    'width': SIZE,
+    'height': SIZE,
+    'count': 224,
+    'crs': CRS.from_epsg(32618),
+    'transform': Affine(30.0, 0.0, 691845.0, 0.0, -30.0, 4561935.0),
+    'blockysize': 3,
+    'tiled': False,
+    'compress': 'deflate',
+    'interleave': 'band',
+}
+
+root = 'hyspecnet-11k'
+path = os.path.join(root, 'splits', 'easy')
+os.makedirs(path, exist_ok=True)
+for tile in tiles:
+    for patch in patches:
+        # Split CSV
+        path = os.path.join(tile, f'{tile}-{patch}', f'{tile}-{patch}-DATA.npy')
+        for split in ['train', 'val', 'test']:
+            with open(os.path.join(root, 'splits', 'easy', f'{split}.csv'), 'a+') as f:
+                f.write(f'{path}\n')
+
+        # Spectral image
+        path = os.path.join(root, 'patches', path)
+        os.makedirs(os.path.dirname(path), exist_ok=True)
+        path = path.replace('DATA.npy', 'SPECTRAL_IMAGE.TIF')
+        Z = np.random.randint(
+            np.iinfo(DTYPE).min, np.iinfo(DTYPE).max, size=(SIZE, SIZE), dtype=DTYPE
+        )
+        with rasterio.open(path, 'w', **profile) as src:
+            for i in range(1, profile['count'] + 1):
+                src.write(Z, i)
+
+shutil.make_archive(f'{root}-01', 'gztar', '.', os.path.join(root, 'patches'))
+shutil.make_archive(f'{root}-splits', 'gztar', '.', os.path.join(root, 'splits'))

tests/data/hyspecnet/hyspecnet-11k/splits/easy/test.csv

@@ -0,0 +1,2 @@
+ENMAP01_20221103T162438Z/ENMAP01_20221103T162438Z-Y01460273_X05670694/ENMAP01_20221103T162438Z-Y01460273_X05670694-DATA.npy
+ENMAP01_20221103T162438Z/ENMAP01_20221103T162438Z-Y01460273_X06950822/ENMAP01_20221103T162438Z-Y01460273_X06950822-DATA.npy

tests/data/hyspecnet/hyspecnet-11k/splits/easy/train.csv

@@ -0,0 +1,2 @@
+ENMAP01_20221103T162438Z/ENMAP01_20221103T162438Z-Y01460273_X05670694/ENMAP01_20221103T162438Z-Y01460273_X05670694-DATA.npy
+ENMAP01_20221103T162438Z/ENMAP01_20221103T162438Z-Y01460273_X06950822/ENMAP01_20221103T162438Z-Y01460273_X06950822-DATA.npy

tests/data/hyspecnet/hyspecnet-11k/splits/easy/val.csv

@@ -0,0 +1,2 @@
+ENMAP01_20221103T162438Z/ENMAP01_20221103T162438Z-Y01460273_X05670694/ENMAP01_20221103T162438Z-Y01460273_X05670694-DATA.npy
+ENMAP01_20221103T162438Z/ENMAP01_20221103T162438Z-Y01460273_X06950822/ENMAP01_20221103T162438Z-Y01460273_X06950822-DATA.npy

tests/data/treesatai/data.py

@@ -0,0 +1,129 @@
+#!/usr/bin/env python3
+
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License.
+
+import glob
+import json
+import os
+import random
+import shutil
+import zipfile
+
+import numpy as np
+import rasterio
+from rasterio import Affine
+from rasterio.crs import CRS
+
+SIZE = 32
+
+random.seed(0)
+np.random.seed(0)
+
+classes = (
+    'Abies',
+    'Acer',
+    'Alnus',
+    'Betula',
+    'Cleared',
+    'Fagus',
+    'Fraxinus',
+    'Larix',
+    'Picea',
+    'Pinus',
+    'Populus',
+    'Prunus',
+    'Pseudotsuga',
+    'Quercus',
+    'Tilia',
+)
+
+species = (
+    'Acer_pseudoplatanus',
+    'Alnus_spec',
+    'Fagus_sylvatica',
+    'Picea_abies',
+    'Pseudotsuga_menziesii',
+    'Quercus_petraea',
+    'Quercus_rubra',
+)
+
+profile = {
+    'aerial': {
+        'driver': 'GTiff',
+        'dtype': 'uint8',
+        'nodata': None,
+        'width': SIZE,
+        'height': SIZE,
+        'count': 4,
+        'crs': CRS.from_epsg(25832),
+        'transform': Affine(
+            0.19999999999977022, 0.0, 552245.4, 0.0, -0.19999999999938728, 5728215.0
+        ),
+    },
+    's1': {
+        'driver': 'GTiff',
+        'dtype': 'float32',
+        'nodata': -9999.0,
+        'width': SIZE // 16,
+        'height': SIZE // 16,
+        'count': 3,
+        'crs': CRS.from_epsg(32632),
+        'transform': Affine(10.0, 0.0, 552245.0, 0.0, -10.0, 5728215.0),
+    },
+    's2': {
+        'driver': 'GTiff',
+        'dtype': 'uint16',
+        'nodata': None,
+        'width': SIZE // 16,
+        'height': SIZE // 16,
+        'count': 12,
+        'crs': CRS.from_epsg(32632),
+        'transform': Affine(10.0, 0.0, 552241.6565, 0.0, -10.0, 5728211.6251),
+    },
+}
+
+multi_labels = {}
+for split in ['train', 'test']:
+    with open(f'{split}_filenames.lst') as f:
+        for filename in f:
+            filename = filename.strip()
+            for sensor in ['aerial', 's1', 's2']:
+                kwargs = profile[sensor]
+                directory = os.path.join(sensor, '60m')
+                os.makedirs(directory, exist_ok=True)
+                if 'int' in kwargs['dtype']:
+                    Z = np.random.randint(
+                        np.iinfo(kwargs['dtype']).min,
+                        np.iinfo(kwargs['dtype']).max,
+                        size=(kwargs['height'], kwargs['width']),
+                        dtype=kwargs['dtype'],
+                    )
+                else:
+                    Z = np.random.rand(kwargs['height'], kwargs['width'])
+
+                path = os.path.join(directory, filename)
+                with rasterio.open(path, 'w', **kwargs) as src:
+                    for i in range(1, kwargs['count'] + 1):
+                        src.write(Z, i)
+
+            k = random.randrange(1, 4)
+            labels = random.choices(classes, k=k)
+            pcts = np.random.rand(k)
+            pcts /= np.sum(pcts)
+            multi_labels[filename] = list(map(list, zip(labels, map(float, pcts))))
+
+os.makedirs('labels', exist_ok=True)
+path = os.path.join('labels', 'TreeSatBA_v9_60m_multi_labels.json')
+with open(path, 'w') as f:
+    json.dump(multi_labels, f)
+
+for sensor in ['s1', 's2', 'labels']:
+    shutil.make_archive(sensor, 'zip', '.', sensor)
+
+for spec in species:
+    path = f'aerial_60m_{spec}.zip'.lower()
+    with zipfile.ZipFile(path, 'w') as f:
+        for path in glob.iglob(os.path.join('aerial', '60m', f'{spec}_*.tif')):
+            filename = os.path.split(path)[-1]
+            f.write(path, arcname=filename)

tests/data/treesatai/labels/TreeSatBA_v9_60m_multi_labels.json

@@ -0,0 +1 @@
+{"Picea_abies_3_46636_WEFL_NLF.tif": [["Prunus", 0.20692122963708826], ["Fraxinus", 0.7930787703629117]], "Pseudotsuga_menziesii_1_339575_BI_NLF.tif": [["Tilia", 0.4243067837573989], ["Larix", 0.5756932162426011]], "Quercus_rubra_1_92184_WEFL_NLF.tif": [["Tilia", 0.5816157697641007], ["Fagus", 0.4183842302358993]], "Fagus_sylvatica_9_29995_WEFL_NLF.tif": [["Larix", 1.0]], "Quercus_petraea_5_80549_WEFL_NLF.tif": [["Alnus", 0.5749721529276662], ["Acer", 0.4250278470723338]], "Acer_pseudoplatanus_3_5758_WEFL_NLF.tif": [["Tilia", 0.8430361090251272], ["Larix", 0.1569638909748729]], "Alnus_spec._5_13114_WEFL_NLF.tif": [["Pseudotsuga", 0.17881149698366108], ["Quercus", 0.38732907538618866], ["Cleared", 0.4338594276301503]], "Quercus_petraea_2_84375_WEFL_NLF.tif": [["Acer", 0.3909090505343164], ["Pseudotsuga", 0.2628926194326892], ["Cleared", 0.34619833003299444]], "Picea_abies_2_46896_WEFL_NLF.tif": [["Acer", 0.4953810312272686], ["Fraxinus", 0.0006659055704136941], ["Pinus", 0.5039530632023177]], "Acer_pseudoplatanus_4_6058_WEFL_NLF.tif": [["Tilia", 1.0]]}

tests/data/treesatai/test_filenames.lst

@@ -0,0 +1 @@
+Acer_pseudoplatanus_4_6058_WEFL_NLF.tif

tests/data/treesatai/train_filenames.lst

@@ -0,0 +1,9 @@
+Picea_abies_3_46636_WEFL_NLF.tif
+Pseudotsuga_menziesii_1_339575_BI_NLF.tif
+Quercus_rubra_1_92184_WEFL_NLF.tif
+Fagus_sylvatica_9_29995_WEFL_NLF.tif
+Quercus_petraea_5_80549_WEFL_NLF.tif
+Acer_pseudoplatanus_3_5758_WEFL_NLF.tif
+Alnus_spec._5_13114_WEFL_NLF.tif
+Quercus_petraea_2_84375_WEFL_NLF.tif
+Picea_abies_2_46896_WEFL_NLF.tif