Allow parsing GRD files in the Planetary Computer

Author: alexamici

tests/test_10_esa_safe.py

@@ -159,7 +159,7 @@ def test_parse_annotation_filename() -> None:
         "s1b-iw1-slc-vv-20210401t052624-20210401t052649-026269-032297-004.xml"
     )
 
-    assert res == ("iw1", "vv", "20210401t052624")
+    assert res == ("", "iw1", "vv", "20210401t052624")
 
     with pytest.raises(ValueError):
         esa_safe.parse_annotation_filename("")

xarray_sentinel/esa_safe.py

@@ -94,9 +94,9 @@ def findall(
     return values
 
 
-def parse_annotation_filename(name: str) -> T.Tuple[str, str, str]:
+def parse_annotation_filename(name: str) -> T.Tuple[str, str, str, str]:
     match = re.match(
-        r"[a-z-]*s1[ab]-([^-]*)-[^-]*-([^-]*)-([\dt]*)-", os.path.basename(name)
+        r"([a-z-]*)s1[ab]-([^-]*)-[^-]*-([^-]*)-([\dt]*)-", os.path.basename(name)
     )
     if match is None:
         raise ValueError(f"cannot parse name {name!r}")
@@ -106,7 +106,7 @@ def parse_annotation_filename(name: str) -> T.Tuple[str, str, str]:
 @functools.lru_cache
 def parse_manifest_sentinel1(
     manifest_path: PathOrFileType,
-) -> T.Tuple[T.Dict[str, T.Any], T.Dict[str, T.Tuple[str, str, str, str]]]:
+) -> T.Tuple[T.Dict[str, T.Any], T.Dict[str, T.Tuple[str, str, str, str, str]]]:
     # We use ElementTree because we didn't find a XSD definition for the manifest
     manifest = ElementTree.parse(manifest_path)
 
@@ -155,12 +155,10 @@ def parse_manifest_sentinel1(
         if location_tag is not None:
             file_href = location_tag.attrib["href"]
             try:
-                swath, polarization, start = parse_annotation_filename(
-                    os.path.basename(file_href)
-                )
+                description = parse_annotation_filename(os.path.basename(file_href))
             except ValueError:
                 continue
             file_type = file_tag.attrib["repID"]
-            files[file_href] = (file_type, swath, polarization, start)
+            files[file_href] = (file_type,) + description
 
     return attributes, files

xarray_sentinel/sentinel1.py

@@ -353,13 +353,13 @@ def open_azimuth_fm_rate_dataset(annotation: esa_safe.PathOrFileType) -> xr.Data
 
 
 def find_available_groups(
-    product_files: T.Dict[str, T.Tuple[str, str, str, str]],
+    product_files: T.Dict[str, T.Tuple[str, str, str, str, str]],
     product_path: str,
     check_files_exist: bool = False,
     fs: fsspec.AbstractFileSystem = fsspec.filesystem("file"),
 ) -> T.Dict[str, T.List[str]]:
     groups: T.Dict[str, T.List[str]] = {}
-    for path, (type, swath, polarization, _) in product_files.items():
+    for path, (type, _, swath, polarization, _) in product_files.items():
         swath_pol_group = f"{swath}/{polarization}".upper()
         abspath = os.path.join(product_path, os.path.normpath(path))
         if check_files_exist:
@@ -419,6 +419,7 @@ def open_pol_dataset(
     }
     encoding = {}
     swap_dims = {}
+    chunks: T.Union[None, T.Dict[str, int]] = None
 
     azimuth_time = pd.date_range(
         start=first_azimuth_time,
@@ -452,6 +453,7 @@ def open_pol_dataset(
             import dask  # noqa
 
             encoding["preferred_chunks"] = {"line": lines_per_burst}
+            chunks = {}
         except ModuleNotFoundError:
             pass
 
@@ -486,7 +488,7 @@ def open_pol_dataset(
     #   the with is needed to avoid polluting stderr when the try block fails
     with contextlib.redirect_stderr(open("/dev/null", "w")):
         try:
-            arr = xr.open_dataarray(fs.open(measurement), engine="rasterio")  # type: ignore
+            arr = xr.open_dataarray(fs.open(measurement), engine="rasterio", chunks=chunks)  # type: ignore
         except AttributeError:
             arr = xr.open_dataarray(measurement, engine="rasterio")  # type: ignore
 
@@ -669,14 +671,15 @@ def compute_burst_centres(
 
 
 def do_override_product_files(
-    template: str, product_files: T.Dict[str, T.Tuple[str, str, str, str]]
-) -> T.Dict[str, T.Tuple[str, str, str, str]]:
+    template: str, product_files: T.Dict[str, T.Tuple[str, str, str, str, str]]
+) -> T.Dict[str, T.Tuple[str, str, str, str, str]]:
     overridden_product_files = {}
-    for path, (type, swath, polarization, date) in product_files.items():
+    for path, description in product_files.items():
+        type, prefix, swath, polarization, date = description
         ext = os.path.splitext(path)[1]
         dirname = os.path.dirname(path)
         overridden_path = template.format(**locals())
-        overridden_product_files[overridden_path] = (type, swath, polarization, date)
+        overridden_product_files[overridden_path] = description
     return overridden_product_files