Fix copernicus storage client thumbnail URL resolver (#36)

Author: lukasbindreiter

CHANGELOG.md

@@ -7,10 +7,16 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.51.0] - 2026-04-07
+
 ### Changed
 
 - `tilebox-storage`: Replaced `httpx` with `niquests` for ASF HTTP downloads.
 
+### Fixed
+
+- `tilebox-storage`: Fixed an issue with the Copernicus storage client that prevented downloading granules pointing to the Copernicus OData thumbnail endpoint. (All granules ingested from March 2026 onwards).
+
 ## [0.50.1] - 2026-04-01
 
 ### Added
@@ -351,7 +357,8 @@ the first client that does not cache data (since it's already on the local file
 - Released under the [MIT](https://opensource.org/license/mit) license.
 - Released packages: `tilebox-datasets`, `tilebox-workflows`, `tilebox-storage`, `tilebox-grpc`
 
-[Unreleased]: https://github.com/tilebox/tilebox-python/compare/v0.50.0...HEAD
+[Unreleased]: https://github.com/tilebox/tilebox-python/compare/v0.51.0...HEAD
+[0.51.0]: https://github.com/tilebox/tilebox-python/compare/v0.50.1...v0.51.0
 [0.50.1]: https://github.com/tilebox/tilebox-python/compare/v0.50.0...v0.50.1
 [0.50.0]: https://github.com/tilebox/tilebox-python/compare/v0.49.0...v0.50.0
 [0.49.0]: https://github.com/tilebox/tilebox-python/compare/v0.48.0...v0.49.0

tilebox-storage/tilebox/storage/aio.py

@@ -27,6 +27,7 @@
     LocationStorageGranule,
     UmbraStorageGranule,
     USGSLandsatStorageGranule,
+    _is_copernicus_odata_url,
 )
 from tilebox.storage.providers import login
 
@@ -750,6 +751,22 @@ async def _download_quicklook(self, datapoint: xr.Dataset | CopernicusStorageGra
             else Path.cwd() / self._STORAGE_PROVIDER
         )
 
+        if _is_copernicus_odata_url(granule.thumbnail):
+            # the thumbnail is not stored in the S3 bucket, but is accessible via a public URL. So download it
+            # directly.
+            response = await niquests.aget(
+                granule.thumbnail, allow_redirects=True
+            )  # to check if the thumbnail is accessible, raises if not
+            response.raise_for_status()
+            content = response.content
+            if content is None:
+                raise ValueError("Received empty content when downloading quicklook.")
+
+            download_location = (output_folder / granule.granule_name).with_suffix(".jpg")
+            download_location.parent.mkdir(parents=True, exist_ok=True)
+            download_location.write_bytes(content)
+            return download_location
+
         await download_objects(self._store, prefix, [granule.thumbnail], output_folder, show_progress=False)
         return output_folder / granule.thumbnail
 

tilebox-storage/tilebox/storage/granule.py

@@ -99,11 +99,34 @@ def _thumbnail_relative_to_eodata_location(thumbnail_url: str, location: str) ->
         >>> )
         "preview/thumbnail.png"
     """
-
     url_path = thumbnail_url.rsplit("?path=", maxsplit=1)[-1]
     url_path = url_path.removeprefix("/")
     location = location.removeprefix("/eodata/")
-    return str(ObjectPath(url_path).relative_to(location))
+    try:
+        return str(ObjectPath(url_path).relative_to(location))
+    except ValueError:
+        # in case the path couldn't be properly parsed, relative_to will fail. Fall back to the default value then
+        return thumbnail_url
+
+
+def _is_copernicus_odata_url(url: str) -> bool:
+    """
+    Checks whether a thumbnail path is an URL pointing to the Copernicus OData API
+
+    Those URLs don't encode the actual filename/location, so we cannot easily convert them to the S3 Paths.
+    Therefore those thumbnails we'll always download via HTTP
+
+    Example:
+        >>> _is_copernicus_odata_url("https://catalogue.dataspace.copernicus.eu/odata/v1/Assets(822e7592-0a66-41b1-b87d-27eec64c377b)/$value")
+        True
+
+    Args:
+        url: The granule thumbnail URL to check
+
+    Returns:
+        bool: True if the URL is a Copernicus OData API URL, False otherwise
+    """
+    return url.startswith("https://catalogue.dataspace.copernicus.eu/odata/v1/Assets") and url.endswith("/$value")
 
 
 @dataclass
@@ -133,11 +156,9 @@ def from_data(cls, dataset: "xr.Dataset | CopernicusStorageGranule") -> "Coperni
         if "thumbnail" in dataset:
             thumbnail_path = dataset.thumbnail.item().strip()
 
-        thumbnail = (
-            _thumbnail_relative_to_eodata_location(thumbnail_path, location)
-            if isinstance(thumbnail_path, str) and len(thumbnail_path) > 0
-            else None
-        )
+        thumbnail = thumbnail_path if isinstance(thumbnail_path, str) and len(thumbnail_path) > 0 else None
+        if thumbnail is not None and not _is_copernicus_odata_url(thumbnail):
+            thumbnail = _thumbnail_relative_to_eodata_location(thumbnail, location)
 
         return cls(
             time,