From 924805f0146b81d42ece5c6c3e426cc05983823b Mon Sep 17 00:00:00 2001
From: Noa Aviel Dove <nadove@ucsc.edu>
Date: Fri, 22 Mar 2024 18:54:58 -0700
Subject: [PATCH] fixup! Add JSONL-based verbatim manifest format (#6028)

---
 src/azul/plugins/__init__.py         |  5 ++++-
 src/azul/service/manifest_service.py |  2 +-
 test/service/test_manifest.py        | 10 +++++-----
 3 files changed, 10 insertions(+), 7 deletions(-)

diff --git a/src/azul/plugins/__init__.py b/src/azul/plugins/__init__.py
index 44de233ad2..212f4d6733 100644
--- a/src/azul/plugins/__init__.py
+++ b/src/azul/plugins/__init__.py
@@ -409,7 +409,10 @@ def source_id_field(self) -> str:
     def implicit_hub_type(self) -> str:
         """
         The type of entities that do not explicitly track their hubs in replica
-        documents.
+        documents in order to avoid a large list of hub references in the
+        replica document, and to avoid contention when updating that list during
+        indexing. Note that this is not a type of hub entities, but rather the
+        type of replica entities that have implicit hubs.
         """
         raise NotImplementedError
 
diff --git a/src/azul/service/manifest_service.py b/src/azul/service/manifest_service.py
index 8f5f496f20..b777765516 100644
--- a/src/azul/service/manifest_service.py
+++ b/src/azul/service/manifest_service.py
@@ -2096,7 +2096,7 @@ def _join_replicas(self, keys: Iterable[ReplicaKeys]) -> Iterable[Hit]:
         return request.scan()
 
     def create_file(self) -> tuple[str, Optional[str]]:
-        fd, path = mkstemp()
+        fd, path = mkstemp(suffix=f'.{self.file_name_extension()}')
         os.close(fd)
         with open(path, 'w') as f:
             for replica in self._all_replicas():
diff --git a/test/service/test_manifest.py b/test/service/test_manifest.py
index f971fa9a1b..4e7736e960 100644
--- a/test/service/test_manifest.py
+++ b/test/service/test_manifest.py
@@ -1278,7 +1278,7 @@ def test_manifest_content_disposition_header(self):
     @manifest_test
     def test_verbatim_jsonl_manifest(self):
         bundle = self._load_canned_bundle(one(self.bundles()))
-        expected_contents = [
+        expected = [
             bundle.metadata_files[d]
             for d in [
                 'cell_suspension_0.json',
@@ -1290,14 +1290,14 @@ def test_verbatim_jsonl_manifest(self):
         ]
         response = self._get_manifest(ManifestFormat.verbatim_jsonl, {})
         self.assertEqual(200, response.status_code)
-        response_contents = list(map(json.loads, response.content.decode().splitlines()))
+        response = list(map(json.loads, response.content.decode().splitlines()))
 
         def sort_key(hca_doc: JSON) -> str:
             return hca_doc['provenance']['document_id']
 
-        expected_contents.sort(key=sort_key)
-        response_contents.sort(key=sort_key)
-        self.assertEqual(expected_contents, response_contents)
+        expected.sort(key=sort_key)
+        response.sort(key=sort_key)
+        self.assertEqual(expected, response)
 
 
 class TestManifestCache(ManifestTestCase):