[kevinmarete]

Addresses

https://broadworkbench.atlassian.net/browse/DT-4378

Security risk: low — two new admin-only endpoints that rerun an existing, reviewed sync; no change to who can read or write studies.

Summary

StudyAssetRecordSync.syncStudy (#3213) had no caller. This adds the trigger the plan's Rollback steps need, and DT-4331 will use the bulk call to convert existing studies instead of a SQL changeset.

• POST /api/dataset/study/{studyId}/assets/sync: runs syncStudy for one study in its own transaction, then reindexes the study's datasets. syncStudy's 404 (no study) and 400 (a biospecimen in row-id form with no sampleId) pass through.
• POST /api/dataset/study/assets/sync?since=<ISO-8601 instant>: every study, or each study with create_date or update_date at or after since (new studies have a null update_date). One transaction per study in id order. Streams one result per line as application/x-ndjson, then a summary line; a failing study is reported and the run continues, and the run also continues if the caller disconnects.
• The plan's Rollback section names both calls.

Differences from the design in the ticket comment:

• The single-study call lives in StudyAssetResource. That class owns api/dataset/study/{studyId}/assets, and Jersey picks the root class with the longer literal prefix and does not fall back, so the route on StudyResource answered 404.
• The reindex covers only datasets already in the index (indexed_date IS NOT NULL), not indexStudy, which indexes every dataset of the study and would re-add one an admin removed. The result reports index as UPDATED, NOT_INDEXED or FAILED; a bulk request that answers 200 with "errors": true counts as FAILED.

Testing

• StudyResourceTest, StudyAssetResourceTest, DatasetServiceTest, StudyDAOTest, StudyAssetRecordSyncTest: admin-only on both calls, error pass-through, a malformed since, the loop continuing past a failed study and a dropped client, reindex outcomes, the since query (inclusive, new and edited studies), and a sync that rolls back when a later asset type fails.
• mvn validate passes the OpenAPI spec.

---

Have you read CONTRIBUTING.md lately? If not, do that first.

• Label PR with a Jira ticket number and include a link to the ticket
• Label PR with a security risk modifier [no, low, medium, high]
• PR describes scope of changes
• Get a minimum of one thumbs worth of review, preferably two if enough team members are available
• Get PO sign-off for all non-trivial UI or workflow changes
• Verify all tests go green
• Test this change deployed correctly and works on dev environment after deployment

🤖 Generated with Claude Code

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota. Try requesting a review at a lower effort level.

An organization owner can choose the repository organization as the billing source for reviews requested by members with a paid Copilot license. Learn about Copilot code review billing.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
99.1% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[rushtong]

Looks good, only low findings 👍🏽

• Low: The bulk resync runs on one request thread. A run over all studies can take longer than the ingress timeout. The summary then goes only to the server log.
• Low: If a study is deleted during the run, the summary counts it as failed with a 404.
• Low: The since filter finds only the edits that change study.update_date.