Release Date:
Summary
Stack-612 is a quality-and-reliability release centered on the Curator data grid. Large curation projects now load faster and stay responsive, validation is stricter and more accurate, and Curator can now handle masked age values used to protect participant privacy.
Beyond Curator, this release fixes several issues people were running into elsewhere in Synapse: the table query builder's "is all of" filter, markdown plot widgets, and turning off faceting on table columns. It also adds participant-count obfuscation to Cohort Builder 2.0, modernizes how data access requests are stored behind the scenes, and continues security and code-quality hardening work.
In total, 14 tickets were completed across the web client and platform, covering curation performance, data validation, table search, privacy, security, and engineering infrastructure.
No breaking changes or required user actions in this release. A few fixes change results people may notice (for example, "is all of" queries now return fewer, correct rows); these are called out below.
Affected Users & Systems
|
Who/What |
Impact |
|---|---|
|
Data contributors and curators using Curator |
Faster, more stable grids for large datasets; stricter integer validation; support for masked age values. |
|
Researchers searching Synapse tables |
The query builder's "is all of" filter now returns only rows matching every selected value. |
|
Project and wiki authors |
Plot widgets in wiki/markdown pages render again. |
|
Table and view owners |
Faceting can be turned off on a column even when a facet sort order was set. |
|
ELITE portal / Cohort Builder 2.0 users |
Participant counts are obfuscated to better protect participant privacy. |
|
Data access requesters and the Access and Compliance Team (ACT) |
No visible change expected; data access requests use a modernized storage format. |
|
Engineering and security teams |
Legacy code removed, additional code scanning, and faster linting and testing. |
Deprecated or Breaking Changes
There are no deprecated features or breaking changes in this release, and no action is required from users.
New Features
Age masking support in Curator New
Curator now correctly handles schema fields that accept either a number or a specific set of text values, such as an age field that allows a numeric age or a masked value like ">= 90" or "Unknown". Previously, entering a normal numeric age was saved as text and flagged as invalid, showing contributors confusing red cells and forcing data managers to fix values by hand afterward. Numeric ages are now saved and validated as numbers, while the approved masked values remain valid.
Who benefits: Data contributors and data managers working with human-subjects data that must mask ages to protect participant identity.
Participant count obfuscation in Cohort Builder 2.0 New
Cohort Builder 2.0 now applies an obfuscation method to the participant counts it displays. This reduces the risk that small counts could be used to identify individual participants, while still giving researchers a useful sense of cohort size when exploring data.
Who benefits: Researchers exploring cohorts in the ELITE portal, and the participants whose privacy this protects.
Fixes & Improvements
Faster, more stable Curator grids for large datasets
We made several performance improvements to the Curator grid. Large grids now load with less memory use, edits and updates from collaborators are processed more efficiently, and the browser no longer crashes when a grid has a very large number of validation errors. Together, these changes make it practical to curate much larger datasets in the browser.
Who benefits: Curators and contributors working with large metadata spreadsheets in Curator.
Curator flags decimals entered in whole-number fields
When a field was defined as a whole number with an allowed range, Curator previously rounded a decimal entry down without warning (for example, a cell viability of 0.85 silently became 0). Curator now shows a validation error instead, so mistakes like entering a fraction instead of a percentage are caught at entry time rather than saved as wrong values.
Who benefits: Contributors entering numeric metadata, and data managers relying on that data being accurate.
Query builder "is all of" filter now works correctly
In the table query builder, choosing "is all of" for a list column (for example, a funding agency column) was behaving the same as "is any of", returning rows that matched only one of the selected values. The filter now returns only rows that contain every selected value. Users who relied on this filter may notice fewer results, and those results are now the correct ones.
Who benefits: Researchers and portal users filtering Synapse tables by multiple values.
Plot widgets in wiki pages render again
Plots embedded in wiki and markdown pages had stopped displaying. They now render as expected again.
Who benefits: Project owners and anyone viewing wiki pages that include charts.
Turning off faceting on a table column
Table owners could not turn off faceting for a column if a facet sort order had been set; the schema editor showed an error instead. The sort setting is now cleared automatically when faceting is turned off, so the change saves without errors.
Who benefits: Owners and editors of Synapse tables and views.
Modernized storage for data access requests
Data access requests, renewals, and submissions are now stored in a newer, more maintainable format. This is groundwork for upcoming schema-based access requirements. Requesters and reviewers should see no difference, except that the list of accessors on a submission may now appear in a different, consistent order.
Who benefits: Data access requesters and the Access and Compliance Team (indirectly, through more reliable future features).
Security hardening
We removed leftover code from the previous provenance graph viewer, which closes a reported security finding, and added automated security and code-quality scanning to the Synapse web client's review process. There is no user-facing change.
Who benefits: All Synapse users, through a more secure platform; security and compliance teams.
Engineering tooling and testing improvements
Developers now catch more code-quality issues on their own machines before code is submitted, and we investigated ways to speed up automated end-to-end testing of portals. These are internal improvements with no user-facing impact, but they help the team ship fixes faster and more safely.
Who benefits: The engineering team, and indirectly all users through faster, higher-quality releases.