Add support for Pangaea datasets

J535D165 · Aug 24, 2023 · df3c867 · df3c867
1 parent b5cf8ae
commit df3c867
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 1 deletion.
diff --git a/datahugger/api.py b/datahugger/api.py
@@ -15,6 +15,7 @@
 from datahugger.services import HuggingFaceDataset
 from datahugger.services import MendeleyDataset
 from datahugger.services import OSFDataset
+from datahugger.services import PangaeaDataset
 from datahugger.services import ZenodoDataset
 from datahugger.utils import _is_doi
 from datahugger.utils import _is_url
@@ -53,7 +54,7 @@
     "get.iedadata.org": DataOneDataset,
     "usap-dc.org": DataOneDataset,
     "iys.hakai.org": DataOneDataset,
-    "doi.pangaea.de": DataOneDataset,
+    "doi.pangaea.de": PangaeaDataset,
     "rvdata.us": DataOneDataset,
     "sead-published.ncsa.illinois.edu": DataOneDataset,
     # DataVerse repositories (extracted from re3data)

diff --git a/datahugger/services.py b/datahugger/services.py
@@ -210,6 +210,12 @@ def files(self):
         return self._files
 
 
+class PangaeaDataset(DataOneDataset):
+    """Downloader for PangaeaDataset repository."""
+
+    REGEXP_ID = r"doi\.pangaea\.de/(.*)"
+
+
 class DSpaceDataset(DatasetDownloader, DatasetResult):
     """Downloader for DSpaceDataset repositories."""