Landsat storage client

.pre-commit-config.yaml

@@ -6,7 +6,7 @@ repos:
       - id: end-of-file-fixer
   - repo: https://github.com/charliermarsh/ruff-pre-commit
     # keep the version here in sync with the version in uv.lock
-    rev: "v0.12.7"
+    rev: "v0.12.9"
     hooks:
       - id: ruff-check
         args: [--fix, --exit-non-zero-on-fix]

CHANGELOG.md

@@ -7,6 +7,14 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+
+- `tilebox-storage`: Added `USGSLandsatStorageClient` to download landsat data from the USGS Landsat S3 bucket.
+- `tilebox-storage`: Storage client now support concurrent downloads of multiple objects, controlled by the
+  `max_concurrent_downloads` parameter.
+- `tilebox-storage`: Added `quicklook` and `download_quicklook` methods to the `CopernicusStorageClient` to download and
+  display preview images for Sentinel data.
+
 ## [0.41.0] - 2025-08-01
 
 ### Added

tilebox-datasets/tilebox/datasets/protobuf_conversion/protobuf_xarray.py

@@ -374,19 +374,16 @@ def _create_field_converter(field: FieldDescriptor) -> _FieldConverter:
     """
     # special handling for enums:
     if field.type == FieldDescriptor.TYPE_ENUM:
-        if field.label == FieldDescriptor.LABEL_REPEATED:
+        if field.is_repeated:  # type: ignore[attr-defined]
             raise NotImplementedError("Repeated enum fields are not supported")
 
         return _EnumFieldConverter(field.name, enum_mapping_from_field_descriptor(field))
 
     field_type = infer_field_type(field)
-    if field.label == FieldDescriptor.LABEL_OPTIONAL:  # simple fields (in proto3 every simple field is optional)
-        return _SimpleFieldConverter(field.name, field_type)
-
-    if field.label == FieldDescriptor.LABEL_REPEATED:
+    if field.is_repeated:  # type: ignore[attr-defined]
         return _ArrayFieldConverter(field.name, field_type)
 
-    raise ValueError(f"Unsupported field type with label {field.label} and type {field.type}")
+    return _SimpleFieldConverter(field.name, field_type)
 
 
 def _combine_dimension_names(array_dimensions: dict[str, int]) -> dict[str, tuple[str, int]]:

tilebox-datasets/tilebox/datasets/protobuf_conversion/to_protobuf.py

@@ -6,7 +6,6 @@
 import numpy as np
 import pandas as pd
 import xarray as xr
-from google.protobuf.descriptor import FieldDescriptor
 from google.protobuf.message import Message
 
 from tilebox.datasets.protobuf_conversion.field_types import (
@@ -80,7 +79,7 @@ def to_messages(  # noqa: C901, PLR0912
         descriptor = field_descriptors_by_name[field_name]
         field_type = infer_field_type(descriptor)
 
-        if descriptor.label == FieldDescriptor.LABEL_REPEATED:
+        if descriptor.is_repeated:
             values = convert_repeated_values_to_proto(values, field_type)
         else:
             values = convert_values_to_proto(values, field_type, filter_none=False)

tilebox-storage/pyproject.toml

@@ -26,8 +26,7 @@ dependencies = [
     "aiofile>=3.8",
     "folium>=0.15",
     "shapely>=2",
-    "boto3>=1.33",
-    "boto3-stubs[essential]>=1.33",
+    "obstore>=0.8.0",
 ]
 
 [dependency-groups]
@@ -37,7 +36,6 @@ dev = [
     "pytest-asyncio>=0.24.0",
     "pytest-cov>=5.0.0",
     "pytest>=8.3.2",
-    "moto>=5",
 ]
 
 [project.urls]

tilebox-storage/tests/storage_data.py

@@ -11,7 +11,12 @@
 
 from hypothesis.strategies import DrawFn, booleans, composite, datetimes, integers, just, one_of, text, uuids
 
-from tilebox.storage.granule import ASFStorageGranule, CopernicusStorageGranule, UmbraStorageGranule
+from tilebox.storage.granule import (
+    ASFStorageGranule,
+    CopernicusStorageGranule,
+    UmbraStorageGranule,
+    USGSLandsatStorageGranule,
+)
 from tilebox.storage.providers import _ASF_URL, StorageURLs
 
 
@@ -46,15 +51,14 @@ def alphanumerical_text(draw: DrawFn, min_size: int = 1, max_size: int = 100) ->
 @composite
 def umbra_granules(draw: DrawFn) -> UmbraStorageGranule:
     """Generate a realistic-looking random Umbra granule."""
-    level = "L0"
     time = draw(datetimes(min_value=datetime(1990, 1, 1), max_value=datetime(2025, 1, 1), timezones=just(None)))
     number = draw(integers(min_value=1, max_value=2))
     text_location = draw(alphanumerical_text(min_size=1, max_size=20))
     granule_id = str(draw(uuids(version=4)))
     granule_name = f"{time:%Y-%m-%d-%H-%M-%S}_UMBRA-{number:02d}"
     location = str(Path(text_location) / granule_id / granule_name)
 
-    return UmbraStorageGranule(time, granule_name, level, location)
+    return UmbraStorageGranule(time, granule_name, location)
 
 
 @composite
@@ -80,5 +84,24 @@ def s5p_granules(draw: DrawFn) -> CopernicusStorageGranule:
     # /eodata/Sentinel-5P/TROPOMI/L2__AER_LH/2024/04/15/S5P_NRTI_L2__AER_LH_20240415T055540_20240415T060040_33707_03_020600_20240415T063447
     location = f"/eodata/Sentinel-5P/{instrument}/{product_type}/{start:%Y}/{start:%m}/{start:%d}/{granule_name.removesuffix('.nc')}"
 
-    file_size = draw(integers(min_value=10_000, max_value=999_999_999))
-    return CopernicusStorageGranule(start, granule_name, location, file_size)
+    return CopernicusStorageGranule(start, granule_name, location)
+
+
+@composite
+def landsat_granules(draw: DrawFn) -> USGSLandsatStorageGranule:
+    """Generate a realistic-looking random USGS Landsat granule."""
+    time = draw(datetimes(min_value=datetime(1990, 1, 1), max_value=datetime(2025, 1, 1), timezones=just(None)))
+    landsat_mission = draw(integers(min_value=1, max_value=9))
+
+    path = draw(integers(min_value=1, max_value=999))
+    row = draw(integers(min_value=1, max_value=999))
+
+    granule_name = f"LC{landsat_mission:02d}_L1GT_{path:03d}{row:03d}_{time:%Y%m%d}_{time:%Y%m%d}_02_T1"
+    location = f"s3://usgs-landsat/collection02/level-1/standard/oli-tirs/{time:%Y}/{path:03d}/{row:03d}/{granule_name}"
+    thumbnail = draw(one_of(just(f"{granule_name}_thumb_small.jpeg"), just(None)))
+    return USGSLandsatStorageGranule(
+        time,
+        granule_name,
+        location,
+        thumbnail,
+    )

tilebox-storage/tests/test_granule.py

@@ -4,8 +4,15 @@
 from hypothesis import given
 from hypothesis.strategies import lists
 
-from tests.storage_data import ers_granules, s5p_granules, umbra_granules
-from tilebox.storage.granule import ASFStorageGranule, CopernicusStorageGranule, UmbraStorageGranule, _asf_download_urls
+from tests.storage_data import ers_granules, landsat_granules, s5p_granules, umbra_granules
+from tilebox.storage.granule import (
+    ASFStorageGranule,
+    CopernicusStorageGranule,
+    UmbraStorageGranule,
+    USGSLandsatStorageGranule,
+    _asf_download_urls,
+    _thumbnail_relative_to_eodata_location,
+)
 
 
 def _asf_granule_to_datapoint(granule: ASFStorageGranule) -> xr.Dataset:
@@ -53,7 +60,6 @@ def _umbra_granule_to_datapoint(granule: UmbraStorageGranule) -> xr.Dataset:
     datapoint = xr.Dataset()
     datapoint.coords["time"] = np.array(granule.time).astype("datetime64[ns]")
     datapoint["granule_name"] = granule.granule_name
-    datapoint["processing_level"] = granule.processing_level
     datapoint["location"] = granule.location
     return datapoint
 
@@ -76,12 +82,51 @@ def test_granule_from_umbra_datapoints(granules: list[UmbraStorageGranule]) -> N
         assert UmbraStorageGranule.from_data(dataset.isel(time=i)) == granules[i]
 
 
+@pytest.mark.parametrize(
+    ("thumbnail_url", "location", "expected"),
+    [
+        (
+            "https://catalogue.dataspace.copernicus.eu/get-object?path=/Sentinel-1/SAR/EW_GRDM_1S/2025/08/07/S1A_EW_GRDM_1SDH_20250807T111242_20250807T111346_060429_078305_DB6A.SAFE/preview/thumbnail.png",
+            "/eodata/Sentinel-1/SAR/EW_GRDM_1S/2025/08/07/S1A_EW_GRDM_1SDH_20250807T111242_20250807T111346_060429_078305_DB6A.SAFE",
+            "preview/thumbnail.png",
+        ),
+        (
+            "https://catalogue.dataspace.copernicus.eu/get-object?path=/Sentinel-2/MSI/L1C/2025/08/07/S2B_MSIL1C_20250807T004159_N0511_R045_T08XNR_20250807T004945.SAFE/S2B_MSIL1C_20250807T004159_N0511_R045_T08XNR_20250807T004945-ql.jpg",
+            "/eodata/Sentinel-2/MSI/L1C/2025/08/07/S2B_MSIL1C_20250807T004159_N0511_R045_T08XNR_20250807T004945.SAFE",
+            "S2B_MSIL1C_20250807T004159_N0511_R045_T08XNR_20250807T004945-ql.jpg",
+        ),
+        (
+            "https://catalogue.dataspace.copernicus.eu/get-object?path=/Sentinel-3/OLCI/OL_2_LFR___/2025/08/07/S3A_OL_2_LFR____20250807T011653_20250807T011953_20250807T033036_0179_129_074_1620_PS1_O_NR_003.SEN3/quicklook.jpg",
+            "/eodata/Sentinel-3/OLCI/OL_2_LFR___/2025/08/07/S3A_OL_2_LFR____20250807T011653_20250807T011953_20250807T033036_0179_129_074_1620_PS1_O_NR_003.SEN3",
+            "quicklook.jpg",
+        ),
+        (
+            "https://catalogue.dataspace.copernicus.eu/get-object?path=/Sentinel-3/SLSTR/SL_1_RBT___/2025/08/07/S3B_SL_1_RBT____20250807T002314_20250807T002614_20250807T025411_0179_109_316_0720_ESA_O_NR_004.SEN3/quicklook.jpg",
+            "/eodata/Sentinel-3/SLSTR/SL_1_RBT___/2025/08/07/S3B_SL_1_RBT____20250807T002314_20250807T002614_20250807T025411_0179_109_316_0720_ESA_O_NR_004.SEN3",
+            "quicklook.jpg",
+        ),
+        (
+            "https://catalogue.dataspace.copernicus.eu/get-object?path=/Sentinel-3/SYNERGY/SY_2_VG1___/2025/08/04/S3A_SY_2_VG1____20250804T000000_20250804T235959_20250806T202029_AUSTRALASIA_______PS1_O_NT_002.SEN3/quicklook.jpg",
+            "/eodata/Sentinel-3/SYNERGY/SY_2_VG1___/2025/08/04/S3A_SY_2_VG1____20250804T000000_20250804T235959_20250806T202029_AUSTRALASIA_______PS1_O_NT_002.SEN3",
+            "quicklook.jpg",
+        ),
+    ],
+)
+def test_thumbnail_relative_to_eodata_location(thumbnail_url: str, location: str, expected: str) -> None:
+    assert (
+        _thumbnail_relative_to_eodata_location(
+            thumbnail_url,
+            location,
+        )
+        == expected
+    )
+
+
 def _copernicus_granule_to_datapoint(granule: CopernicusStorageGranule) -> xr.Dataset:
     datapoint = xr.Dataset()
     datapoint.coords["time"] = np.array(granule.time).astype("datetime64[ns]")
     datapoint["granule_name"] = granule.granule_name
     datapoint["location"] = granule.location
-    datapoint["file_size"] = granule.file_size
     return datapoint
 
 
@@ -101,3 +146,31 @@ def test_granule_from_copernicus_datapoints(granules: list[CopernicusStorageGran
 
     for i in range(len(granules)):  # converting a dataset with a time dimension of 1 should still work though
         assert CopernicusStorageGranule.from_data(dataset.isel(time=i)) == granules[i]
+
+
+def _landsat_granule_to_datapoint(granule: USGSLandsatStorageGranule) -> xr.Dataset:
+    datapoint = xr.Dataset()
+    datapoint.coords["time"] = np.array(granule.time).astype("datetime64[ns]")
+    datapoint["granule_name"] = granule.granule_name
+    datapoint["location"] = granule.location
+    if granule.thumbnail is not None:
+        datapoint["thumbnail"] = f"{granule.location}/{granule.thumbnail}"
+    return datapoint
+
+
+@given(landsat_granules())
+def test_granule_from_landsat_datapoint(granule: USGSLandsatStorageGranule) -> None:
+    datapoint = _landsat_granule_to_datapoint(granule)
+    assert USGSLandsatStorageGranule.from_data(datapoint) == granule
+    assert USGSLandsatStorageGranule.from_data(USGSLandsatStorageGranule.from_data(datapoint)) == granule
+
+
+@given(lists(landsat_granules(), min_size=2, max_size=5))
+def test_granule_from_landsat_datapoints(granules: list[USGSLandsatStorageGranule]) -> None:
+    datapoints = [_landsat_granule_to_datapoint(granule) for granule in granules]
+    dataset = xr.concat(datapoints, dim="time")
+    with pytest.raises(ValueError, match=".*more than one granule.*"):
+        USGSLandsatStorageGranule.from_data(dataset)
+
+    for i in range(len(granules)):  # converting a dataset with a time dimension of 1 should still work though
+        assert USGSLandsatStorageGranule.from_data(dataset.isel(time=i)) == granules[i]

tilebox-storage/tests/test_providers.py

@@ -5,7 +5,7 @@
 from tilebox.storage.providers import _asf_login
 
 
-@pytest.mark.anyio
+@pytest.mark.asyncio
 async def test_asf_login(httpx_mock: HTTPXMock) -> None:
     httpx_mock.add_response(headers={"Set-Cookie": "logged_in=yes"})
 
@@ -15,8 +15,10 @@ async def test_asf_login(httpx_mock: HTTPXMock) -> None:
     assert isinstance(client.auth, BasicAuth)
     assert client.cookies["logged_in"] == "yes"
 
+    await client.aclose()
 
-@pytest.mark.anyio
+
+@pytest.mark.asyncio
 async def test_asf_login_invalid_auth(httpx_mock: HTTPXMock) -> None:
     httpx_mock.add_response(401)
     with pytest.raises(ValueError, match="Invalid username or password."):