api: add get file extension from key

* closes zenodo/rdm-project#547

invenio_records_resources/records/api.py

@@ -10,10 +10,12 @@
 """Records API."""
 
 import mimetypes
+import os
 from contextlib import contextmanager
 
 from invenio_db import db
 from invenio_files_rest.models import FileInstance, ObjectVersion
+from invenio_files_rest.utils import FALLBACK_MIMETYPE
 from invenio_records.api import Record as RecordBase
 from invenio_records.dumpers import SearchDumper
 from invenio_records.systemfields import DictField, SystemFieldsMixin
@@ -177,8 +179,20 @@ def dumps(self):
     @property
     def ext(self):
         """File extension."""
-        ext = mimetypes.guess_extension(self.object_model.mimetype)
-        return ext[1:] if ext else None
+        # The ``ext`` property is used to in search to aggregate file types, and we want e.g. both ``.jpeg`` and
+        # ``.jpg`` to be aggregated under ``.jpg``
+        ext_guessed = mimetypes.guess_extension(self.object_model.mimetype)
+
+        # Check if a valid extension is guessed and it's not the default mimetype
+        if ext_guessed is not None and ext_guessed != FALLBACK_MIMETYPE:
+            return ext_guessed[1:]
+
+        # Support non-standard file extensions that cannot be guessed
+        _, ext = os.path.splitext(self.key)
+        if ext and len(ext) <= 5:
+            return ext.lower()
+
+        return ext_guessed[1:]
 
     def __getattr__(self, name):
         """Override to get attributes from ObjectVersion and FileInstance."""