Subgraph stitching query crashes BQ emulator

[hannes-ucsc]

With this patch …

Index: test/docker_container_test_case.py
IDEA additional info:
Subsystem: com.intellij.openapi.diff.impl.patch.CharsetEP
<+>UTF-8
===================================================================
diff --git a/test/docker_container_test_case.py b/test/docker_container_test_case.py
--- a/test/docker_container_test_case.py	(revision 0ed280fc5d28b56d582997e1187d0507aa2c4438)
+++ b/test/docker_container_test_case.py	(date 1705039000741)
@@ -79,7 +79,7 @@
         ports = None if is_sibling else {container_port: ('127.0.0.1', None)}
         container = cls._docker.containers.run(image,
                                                detach=True,
-                                               auto_remove=True,
+                                               auto_remove=False,
                                                ports=ports,
                                                **kwargs)
         try:
Index: test/indexer/test_tdr.py
IDEA additional info:
Subsystem: com.intellij.openapi.diff.impl.patch.CharsetEP
<+>UTF-8
===================================================================
diff --git a/test/indexer/test_tdr.py b/test/indexer/test_tdr.py
--- a/test/indexer/test_tdr.py	(revision 0ed280fc5d28b56d582997e1187d0507aa2c4438)
+++ b/test/indexer/test_tdr.py	(date 1705079815796)
@@ -308,28 +308,17 @@
         with self.assertRaises(RequirementError):
             self._test_fetch_bundle(bundle, load_tables=False)
 
-    @patch('azul.plugins.repository.tdr_hca.Plugin._find_upstream_bundles')
-    def test_subgraph_stitching(self, _mock_find_upstream_bundles):
+    def test_subgraph_stitching(self):
         downstream_uuid = '4426adc5-b3c5-5aab-ab86-51d8ce44dfbe'
         upstream_uuids = [
             'b0c2c714-45ee-4759-a32b-8ccbbcf911d4',
             'bd4939c1-a078-43bd-8477-99ae59ceb555',
         ]
-        # FIXME: Fix the crash in bigquery-emulator and remove the mock
-        #        https://github.com/DataBiosphere/azul/issues/5847
-        _mock_find_upstream_bundles.side_effect = [
-            {SourcedBundleFQID(source=self.source,
-                               uuid=uuid,
-                               version='2020-08-10T21:24:26.174274Z')}
-            for uuid in upstream_uuids
-        ]
         bundle = self._load_canned_bundle(SourcedBundleFQID(source=self.source,
                                                             uuid=downstream_uuid,
                                                             version='2020-08-10T21:24:26.174274Z'))
         assert any(e['is_stitched'] for e in bundle.manifest)
         self._test_fetch_bundle(bundle, load_tables=True)
-        self.assertEqual(_mock_find_upstream_bundles.call_count,
-                         len(upstream_uuids))
 
     def _test_fetch_bundle(self,
                            test_bundle: TDRHCABundle,

… test_subgraph_stitching hangs with the output …

2024-01-11 09:18:34,010    INFO MainThread test.docker_container_test_case: Launched container lucid_hertz from image 465330168186.dkr.ecr.us-east-1.amazonaws.com/ghcr.io/goccy/bigquery-emulator:0.4.4, with container port 9050 mapped to 127.0.0.1:51782 on the host
2024-01-11 09:18:34,970   DEBUG MainThread azul.plugins.repository.tdr_hca: Retrieving 1 entities of type 'links' ...
2024-01-11 09:18:34,970   DEBUG MainThread azul.terra: Query (338 characters total): '\n            SELECT links_id, BYTE_LENGTH(content) AS content_size, JSON_EXTRACT_SCALAR(content, "$.schema_type") AS schema_type, project_id, version, content\n            FROM test_project.snapshot.links\n            WHERE (links_id, version) IN ((\'4426adc5-b3c5-5aab-ab86-51d8ce44dfbe\', TIMESTAMP(\'2020-08-10T21:24:26.174274Z\')))\n        '
2024-01-11 09:18:35,122   DEBUG MainThread azul.terra: Job info: {"job_id": "0c64c459-1f77-4127-9e80-d83a20cb0a74", "stats": {"totalBytesBilled": "3881", "totalBytesProcessed": "3881"}, "query": "\n            SELECT links_id, BYTE_LENGTH(content) AS content_size, JSON_EXTRACT_SCALAR(content, \"$.schema_type\") AS schema_type, project_id, version, content\n            FROM test_project.snapshot.links\n            WHERE (links_id, version) IN (('4426adc5-b3c5-5aab-ab86-51d8ce44dfbe', TIMESTAMP('2020-08-10T21:24:26.174274Z')))\n        "}
2024-01-11 09:18:35,244   DEBUG MainThread azul.plugins.repository.tdr_hca: Retrieved 3 entities of type 'links'
2024-01-11 09:18:35,244    INFO MainThread azul.plugins.repository.tdr_hca: There are 1 dangling inputs in bundle SourcedBundleFQID(uuid='4426adc5-b3c5-5aab-ab86-51d8ce44dfbe', version='2020-08-10T21:24:26.174274Z', source=SourceRef(id='cafebabe-feed-4bad-dead-beaf8badf00d', spec=TDRSourceSpec(prefix=Prefix(common='', partition=2), project='test_project', name='snapshot', is_snapshot=True)))
2024-01-11 09:18:35,244   DEBUG MainThread azul.plugins.repository.tdr_hca: Dangling inputs in bundle SourcedBundleFQID(uuid='4426adc5-b3c5-5aab-ab86-51d8ce44dfbe', version='2020-08-10T21:24:26.174274Z', source=SourceRef(id='cafebabe-feed-4bad-dead-beaf8badf00d', spec=TDRSourceSpec(prefix=Prefix(common='', partition=2), project='test_project', name='snapshot', is_snapshot=True))): {EntityReference(entity_type='sequence_file', entity_id='8f8b9587-237f-4995-9461-c96eac53d615')}
2024-01-11 09:18:35,244   DEBUG MainThread azul.terra: Query (557 characters total): '\n            SELECT links_id, version, JSON_EXTRACT_SCALAR(link_output, "$.output_id") AS output_id\n            FROM test_project.snapshot.links AS links\n                JOIN UNNEST(JSON_EXTRACT_ARRAY(links.content, \'$.links\')) AS content_links\n                    ON JSON_EXTRACT_SCALAR(content_links, \'$.link_type\') = \'process_link\'\n                JOIN UNNEST(JSON_EXTRACT_ARRAY(content_links, \'$.outputs\')) AS link_output\n                    ON JSON_EXTRACT_SCALAR(link_output, "$.output_id") IN UNNEST([\'8f8b9587-237f-4995-9461-c96eac53d615\'])\n        '

… while the container segfaults with …

2024-01-11 09:18:35 2024-01-11T17:18:35.253Z    INFO    contentdata/repository.go:167           {"query": "\n            SELECT links_id, version, JSON_EXTRACT_SCALAR(link_output, \"$.output_id\") AS output_id\n            FROM test_project.snapshot.links AS links\n                JOIN UNNEST(JSON_EXTRACT_ARRAY(links.content, '$.links')) AS content_links\n                    ON JSON_EXTRACT_SCALAR(content_links, '$.link_type') = 'process_link'\n                JOIN UNNEST(JSON_EXTRACT_ARRAY(content_links, '$.outputs')) AS link_output\n                    ON JSON_EXTRACT_SCALAR(link_output, \"$.output_id\") IN UNNEST(['8f8b9587-237f-4995-9461-c96eac53d615'])\n        ", "values": []}
2024-01-11 09:18:35 panic: runtime error: invalid memory address or nil pointer dereference
2024-01-11 09:18:35 [signal SIGSEGV: segmentation violation code=0x1 addr=0x70 pc=0x154abf4]
2024-01-11 09:18:35 
2024-01-11 09:18:35 goroutine 2206 [running]:
2024-01-11 09:18:35 github.com/goccy/go-zetasqlite/internal.RegisterFunctions.func2({0x2ab1660?, 0x4e2f540?})
2024-01-11 09:18:35     /go/pkg/mod/github.com/goccy/[email protected]/internal/function_register.go:430 +0x34
2024-01-11 09:18:35 reflect.Value.call({0x2b311e0?, 0x2ef1378?, 0xc000e98340?}, {0x2e58683, 0x4}, {0xc000c8d008, 0x1, 0x4775fa?})
2024-01-11 09:18:35     /usr/local/go/src/reflect/value.go:596 +0xce7
2024-01-11 09:18:35 reflect.Value.Call({0x2b311e0?, 0x2ef1378?, 0x1?}, {0xc000c8d008?, 0x1?, 0x14b3320?})
2024-01-11 09:18:35     /usr/local/go/src/reflect/value.go:380 +0xb9
2024-01-11 09:18:35 github.com/mattn/go-sqlite3.(*functionInfo).Call(0xc000579000, 0x0?, {0x7fff88104a68?, 0x0?, 0x0?})
2024-01-11 09:18:35     /go/pkg/mod/github.com/mattn/[email protected]/sqlite3.go:407 +0x75
2024-01-11 09:18:35 github.com/mattn/go-sqlite3.callbackTrampoline(0x0?, 0x1, 0x7fff88104a68)
2024-01-11 09:18:35     /go/pkg/mod/github.com/mattn/[email protected]/callback.go:39 +0x59
2024-01-11 09:18:35 github.com/mattn/go-sqlite3._Cfunc__sqlite3_step_internal(0x7fff880f2db8)
2024-01-11 09:18:35     _cgo_gotypes.go:367 +0x47
2024-01-11 09:18:35 github.com/mattn/go-sqlite3.(*SQLiteRows).nextSyncLocked.func1(0xc000bac158?)
2024-01-11 09:18:35     /go/pkg/mod/github.com/mattn/[email protected]/sqlite3.go:2186 +0x45
2024-01-11 09:18:35 github.com/mattn/go-sqlite3.(*SQLiteRows).nextSyncLocked(0xc001234f60, {0xc000d446f0, 0x3, 0x30dd5d8?})
2024-01-11 09:18:35     /go/pkg/mod/github.com/mattn/[email protected]/sqlite3.go:2186 +0x37
2024-01-11 09:18:35 github.com/mattn/go-sqlite3.(*SQLiteRows).Next.func1()
2024-01-11 09:18:35     /go/pkg/mod/github.com/mattn/[email protected]/sqlite3.go:2167 +0x2c
2024-01-11 09:18:35 created by github.com/mattn/go-sqlite3.(*SQLiteRows).Next in goroutine 2097
2024-01-11 09:18:35     /go/pkg/mod/github.com/mattn/[email protected]/sqlite3.go:2166 +0x189

[hannes-ucsc]

I filed a PR against upstream. Let's see if it gets picked up.

[achave11-ucsc]

@hannes-ucsc: "PR was not picked up immediately so we may need to go with forking upstream longer term."

[hannes-ucsc]

Someone on the PR asked for a test case.