Skip to content

Flywheel Core 21.1

21.1.0

FEATURES

LDAP Sync

The Flywheel Core API and LDAP Sync service have been updated to support matching users on non-email formatted User IDs via a new Flywheel user field.

Note

LDAP Sync is a legacy Flywheel service, and supported only on sites where it is already in use.

  • Flywheel user records gain a new (optional) ldap_sync_id field.
    • It can be set and viewed with the Web UI and existing user management APIs.
    • It must be unique for each user on the site.
    • Once set, it cannot be changed.
    • While set, LDAP Sync treats it as the primary matching Flywheel User key
    • While unset, LDAP Sync falls back to pre-existing user matching criteria.
    • It is set automatically for users added by LDAP Sync.
  • LDAP Sync treats sAMAccountName as the primary matching LDAP User key. This is configurable.
  • Audit Trail reports include ldap_sync_id in addition to the Flywheel User ID.

Viewers

V3 Viewer

  • The 2D annotations have been significantly improved in this release to provide a better user experience with annotations workflows:

    • Annotations are now saved as "measurements sets" in the Flywheel database, and a summary can be viewed at the file level in Flywheel.
    • Multiple measurements sets can be loaded or created for the same image file to allow comparison between annotations from different users.
    • Annotations can now be created on multiple image files simultaneously in the viewer.
    • Annotations can now be grouped under the same label and area is aggregated for each label.
    • Option to automatically assign new annotation to the active label.
    • Colors can be assigned to each annotations label.
    • Undo/redo functionalities are now available to quickly correct any mistake.

  • Improvements in 3D segmentations capabilities:

    • Undo/redo functionalities now allow to experiment with segmentation and quickly correct or adjust any mistake.
    • New option to automatically turn the brush tool into an eraser when the mouse cursor is outside a segmented region.
    • Metrics are now available for each region of interest.

FIXES

Compute and Analytics

  • Fixes a race condition in the timeout and retry logic for collecting gear logs and execution health status, introduced with the compute reliability improvements in version 20.6.

Data Transfer

  • Significantly improved speed of Bulk Export and Import jobs to and from "My Computer" by removing performance bottlenecks.
  • Fixed an issue where Bulk Export or Bulk Import jobs may fail unexpectedly on sites with an extremely large number (>400) of External Storage registrations.
    • This issue was caused by the system performing status checks for all storages at the same time, resulting in the system running out of memory.
    • Storage checks now run less frequently and at slightly different times (random jitter) to prevent concurrent execution. Additionally status checks for File System storages are now performed sequentially.
Bulk Export
  • Fixed an issue with Bulk Export to "My Computer" where files downloads are significantly delayed if unzipping is disabled.
  • Fixed an issue where Bulk Export to "My Computer" with unzipping enabled fails due to the system running out of memory when processing ZIP files containing large (>200 MB) individual files.
  • Fixed an issue with Bulk Export to "My Computer" where downloads would fail for large projects (>40 GB) due to resource constraints on users' local machines.
    • The solution includes improved memory usage, additional indexes, and other refactoring.
    • As a side effect of this fix, downloads now transfer the data directly to disk instead of first pre-zipping files before transfer.
    • As an extra benefit, this fix also enables better support for resuming interrupted transfers.
  • Fixed an issue where Bulk Export to "My Computer" would fail for larger files (>300 MB) due to HTTP 416 errors related to multi-part range requests.
    • Added additional logic to fall back to single-part downloads when multipart range requests fail to improve reliability.
Bulk Import
  • Fixed an issue with where Bulk Import jobs fail early due to the malware scanner failing to start up.
  • Fixed an issue where Bulk Import jobs may fail in certain circumstances due to running out of disk space due to an accumulation of temporary files.
  • Fixed an issue with Bulk Import for sites hosted in Azure where the completed files popup incorrectly displayed "0 B completed" instead of showing the actual uploaded file size.
    • This issue only affected Flywheel Core sites hosted in Azure.
  • Fixed an issue with Bulk Import with metadata where classification and modality information was not imported from metadata sidecar files.
External Storages
  • Revised the error message displayed when creation of an External Storage registration fails for due object storage bucket names exceeding the 63-character limit.
    • The error now clearly indicates the character limit and advises users to move path components to the Prefix field instead.

JupyterLab

  • Fixes the formatting of items in the Launch JupyterLab Server menu.

Viewers

V3 Viewer
  • Fixed an issue where switching between a common layout and an advanced layout would create a duplicate of segmentations in the segmentation panel.
  • Fixed an issue where the perimeter value from the freehand tool was incorrect.
  • Fixed an issue where segmentations saved on Philips MR data could not be re-opened.
  • Fixed an issue where a user would become unable to draw ROIs after switching between some of the advanced layouts.
  • Fixed an issue where measurements would incorrectly be moved to the same slice upon saving in MPR mode for multi-frame data.

Web UI

  • Fixes an issue where some Site Admin pages could rendered for authenticated users without the Site Admin role. The pages were not fully functional, and could not result in a user gaining unauthorized access to information, or administrative actions.