Merge pull request #1107 from dandi/gh-980

get_content_url(): If a HEAD fails, return the failing URL
dandi · Aug 24, 2022 · 99fa604 · 99fa604
2 parents 647b351 + f5a6693
commit 99fa604
Show file tree

Hide file tree

Showing 2 changed files with 26 additions and 13 deletions.
diff --git a/dandi/dandiapi.py b/dandi/dandiapi.py
@@ -1341,19 +1341,22 @@ def get_content_url(
             raise NotFoundError(
                 "No matching URL found in asset's contentUrl metadata field"
             )
-        if follow_redirects is True:
-            url = self.client.request(
-                "HEAD", url, json_resp=False, allow_redirects=True
-            ).url
-        elif follow_redirects:
-            for _ in range(follow_redirects):
-                r = self.client.request(
-                    "HEAD", url, json_resp=False, allow_redirects=False
-                )
-                if "Location" in r.headers:
-                    url = r.headers["Location"]
-                else:
-                    break
+        try:
+            if follow_redirects is True:
+                url = self.client.request(
+                    "HEAD", url, json_resp=False, allow_redirects=True
+                ).url
+            elif follow_redirects:
+                for _ in range(follow_redirects):
+                    r = self.client.request(
+                        "HEAD", url, json_resp=False, allow_redirects=False
+                    )
+                    if "Location" in r.headers:
+                        url = r.headers["Location"]
+                    else:
+                        break
+        except requests.HTTPError as e:
+            url = e.request.url
         if strip_query:
             url = urlunparse(urlparse(url)._replace(query=""))
         return url

diff --git a/dandi/tests/test_dandiapi.py b/dandi/tests/test_dandiapi.py
@@ -250,6 +250,16 @@ def test_get_content_url_follow_one_redirects_strip_query() -> None:
         )
 
 
+def test_get_content_url_follow_redirects_zarr(zarr_dandiset: SampleDandiset) -> None:
+    asset = zarr_dandiset.dandiset.get_asset_by_path("sample.zarr")
+    assert isinstance(asset, RemoteZarrAsset)
+    url = asset.get_content_url(follow_redirects=True, strip_query=True)
+    assert re.fullmatch(
+        f"http://localhost:9000/dandi-dandisets/zarr/{asset.zarr}/*",
+        url,
+    )
+
+
 def test_remote_asset_json_dict(text_dandiset: SampleDandiset) -> None:
     asset = text_dandiset.dandiset.get_asset_by_path("file.txt")
     assert asset.json_dict() == {