Skip to content

Flywheel Core 20.4

20.4.0

FEATURES

Data Transfer

Bulk Export
  • When cancelling a Bulk Export to "My Computer," the in-progress downloads are automatically stopped as well.
  • Usability Improvements
    • Simplified the auto-generated labels for Bulk Export jobs to align with the format used for Bulk Import (e.g., Export via CLI from <date> <time>).
    • The Device name is now displayed instead of the Device ID in the Bulk Export jobs list page for export jobs started by Devices.
    • Ensured consistent styling on both the Bulk Export and Bulk Import pages related to formatting of dates and padding around tables.
  • Error Handling Improvements
    • Improved error handling on the Bulk Export pages for when the backend services are temporarily unavailable.
Bulk Import
  • Conflict resolution actions are now recorded for audit purposes.
    • A new "Import Conflicts Report" (separate from the existing "Import Audit Report") is available for download for any import job where a conflict has been detected.
    • The "Import Conflicts Report" will detail every conflict that was detected for the associated import job along with which action was taken (if any) to resolve each conflict.
  • Usability Improvements
    • Bulk Import rules can be referenced in a template file instead of only being passed in directly on the command line.
      • This enables rules to be more easily stored and re-used across separate Bulk Import jobs.
      • The import rules file supports YAML format and follows the same structure as in-line import rules.
      • A rules file can be specified via the --rules-file /path/to/import/rules/file.yml command-line option.
      • For more information, refer to the new CLI documentation for the import run command.
    • Enhanced the Bulk Import's conflict handling to simplify "File Conflicts" caused by "Unique index error" where the "Keep All" resolution option is not available.
      • Now, this type of conflict will be presented still as a "File Conflict" but caused by "File name already exists" and the "Keep All" will be available for resolving it.
    • Files named .DS_Store are automatically excluded by Bulk Import using the default import rules.
    • The Device name is now displayed instead of the Device ID in the Bulk Import jobs list page for import jobs started by Devices.
    • Ensured consistent styling on both the Bulk Export and Bulk Import pages related to formatting of dates and padding around tables.
  • Error Handling Improvements
    • Improved the resilience and reliability of the malware scanning service for Bulk Import.
    • Improved error handling in Bulk Import for when Amazon S3 returns an internal server error.
    • Improved error messaging in the Add Data dialog for when the list of external storage providers could not be retrieved for any reason.
    • Improved error messaging for when a Bulk Import job cannot be started due to an internal system error.
    • Improved error handling on the Bulk Import pages for when the backend services are temporarily unavailable.
External Storages
  • Improved error handling on the External Storages page for when the backend services are temporarily unavailable.
Legacy Ingest (Cluster)
  • Improved the resiliency of Ingest Cluster by adjusting storage resource usage.

Gear Rules & Reader Tasks

  • Added support for triggering the creation of a reader task via the API using a gear rule, and for submitting annotations to reader tasks through the API.
  • Added support for not including group/project/subject/session/acquisition parent folders when downloading task files as a .zip file.

Reader tasks

  • Usability improvements:
    • Downloading files associated with a task now creates folders with simpler hierarchy and fewer sub-folders.

Viewer

V2 viewer

  • Usability improvements:

    • The viewer in-memory cache can now temporarily expand up to 3GB to allow loading a single large dataset. The previous limit was 1GB which could cause issues for larger datasets.

  • Error handling improvements:

    • Indexing errors due to missing StudyInstanceUID or SeriesInstanceUID are now reported accordingly.
V3 viewer

  • The V3 viewer now supports multi-session viewing for longitudinal assessment of images. Acquisition and analysis files from multiple sessions from the same subject can now be loaded into the viewer for side-by-side comparison.

    • A file browser let users browse all files for a single subject following the Flywheel hierarchy (session/acquisition/file).
    • Sessions and acquisitions are expandable/collapsible for convenient browsing.
    • A search bar in the file browser allows users to quickly locate the files they would like to load.
    • Multiple files can be selected and loaded into the viewer simultaneously.
    • Once loaded in the left-side panel of the viewer, files thumbnails are organized based on the Flywheel hierarchy and sessions can be expanded/collapsed.
    • Segmentation files (RTSTRUCT and SEG files) are made available directly under their reference image series.
    • Each viewport now shows the session and acquisition name of the file being displayed.

  • Zoom and pan can now be synchronized between viewports displaying datasets in the same orientation.

  • Window levels can now be synchronized between viewports displaying datasets of the same image modality.
  • The StudyInstanceUID is no longer appended to the name of DICOM Structured Reports (SR) or DICOM SEG files upon saving annotations, measurements or segmentations for more user-friendly file names.

  • DICOM SEG files are now saved under the same acquisition as their reference image series.

  • Error handling improvements:

  • Usability improvements:

    • The crosshair tool has been moved to the right-side tool bar for improved user experience.
    • Users now see a warning when saving DICOM Structured Reports or DICOM SEG files using the "Save as..." option with the name of an existing report.
    • The viewer now handles zip files containing resource fork file in the _MACOSX directory.

  • Design improvements:

    • UI components have been updated according to Flywheel's design system for a more consistent feel and look across the viewer's interface.

Video Annotation - New in 20.4

Flywheel is introducing an opt-in Video Annotation module with the 20.4.0 release. Flywheel deploys and maintains a private instance of the Computer Vision Annotation Tool (CVAT), and its integration with Flywheel. Flywheel managed video data can be ready to annotate within minutes with only a couple Project settings.

Video Annotation Features
  • Simple Administration: Project Settings are they only customer managed configuration. The CVAT services and integration points are fully deployed and maintained by Flywheel.
  • Automatic Data Coordination: Data is transferred automatically between Flywheel and CVAT. There is no need for users to upload and download video and annotation data.
  • CVAT Annotation Tools: All the manual CVAT annotation tools and shapes are available.
  • Video Formats: Can be used with all the CVAT supported video and still image file formats.
  • Validated: The integration capabilities between Flywheel and CVAT are fully validated by Flywheel.
  • Audit Trail: Annotation files are tied to their owner.
  • Access Logging: A user launching a file in CVAT is captured in the Flywheel Access Log.

FIXES

Audit Trail

  • Resolved an issue where generating an Audit Trail Report on an empty project would result in an error.
  • Resolved an issue where changes to file.info from the OHIF V3 viewer was only visible in the Audit Trail Report, but not visible when inspecting the Audit Trail for the Web UI.

Analyses

  • Resolved an issue where downloading an analysis output file could result in downloading a prior version rather than its current version.

Data Management

  • Resolved an issue where performing a "Move Session to Project" between 2 Smart Copy projects could fail without an opportunity to choose to skip or resolve the label conflict.
  • Resolved an issue where after completing a "Move Session to Subject", any follow up actions on that web page could act on the moves sessions, despite the sessions not showing as checked.

Data Transfer

Bulk Export
  • Fixed an issue with Bulk Export where several of the configuration options for File System type storages (e.g., UID, GID, Permissions) were missing from the Storage Configuration Override section.
  • Fixed an issue with Bulk Export when rerunning a previous Bulk Export job a new timepoint appears to be used instead of the same timepoint from the previous export job.
  • Fixed an issue with Bulk Export when attempting to create an export schedule with insufficient permissions where the error message stated that import_manage permissions are required (should say export_manage).
  • Fixed an issue with Bulk Export via the Web App where if the user searches for a storage destination and then clears out their search term, then the "My Computer" destination cannot be selected.
Bulk Import
  • Fixed an issue with Bulk Import from "My Computer" via the new (BETA) CLI where the import would fail if the Flywheel Core site is backed by Azure Blob Storage.

Gear and Compute

  • Resolved issue with gear jobs failing to run at AWS sites using custom security groups.
  • Enhancement to allow Compute scaling to operate under stricter AWS Service Control Policies that block public IP addresses. This update enhances security and simplifies deployment in restricted cloud configurations.
  • Fixes a transport issue when uploading large (>5GB) files to FW site for customers using the self-service HPC Client
  • Resolved an issue with the communication between individual gear compute resources and the compute monitoring system that was causing premature shutdown of gear resources leading to random gear failures.
  • Resolved issue causing failed gear execution due to deployment script specifying a read-only mount, when it requires a read/write mount. This impacted some on-premises sites.
  • Resolved an issue where the provenance tab would not launch for certain subjects, causing the page to spin indefinitely and throw 500 errors. The fix ensures the provenance tab loads correctly for all subjects. This impacts multiprovider sites that were in existence from 2020 or earlier.

Data Views

  • Resolved issue where the record count shown in UI was not being updated correctly when columns are removed from the data view.
  • When data view results are saved to the project, they are now saved as a .csv file.

Reader Tasks

  • Resolved an issue where the Tasks Web UI was missing the label and filtering capability for the Project column.
  • Resolved an issue where launching the viewer from a reader task could lead to that viewer, and other viewers launching multiple times over themselves. This only affected a minority of Flywheel site configurations.

SDK Usage Report

  • Resolved an issue where the Usage Report methods get_daily_usage_range_report() and get_daily_usage_report() would not show 0 for center_storage_bytes and group_storage_bytes after the final purging of soft deleted data for a deleted project.

Smart Copy

  • Resolved an issue where Tags and Notes were not excluded from a Smart Copy operation when the Smart Copy was configured to do so.

Viewer

V2 viewer
  • Fixed an issue where the viewer would open older versions of a file instead of the latest version.
  • Fixed an issue where the viewer would crash when trying to open a zip file containing copies of the same DICOM file.
  • Fixed an issue where the RGB channel selection menu would sometimes be unresponsive.
  • Fixed an issue where queries made to IO-Proxy for resources that do not exist used to return an error 500 (internal server error) instead of an error 404 (not found).
  • Fixed an issue where the Viewer would not handle private tags with a value of undefined.
V3 viewer
  • Fixed an issue where the window levels for PET images were reset to incorrect values when loading a DICOM SEG file.
  • Fixed an issue where the "Save" option was enabled when editing segmentation files from analyses, even though analysis files are non-editable.
  • Fixed an issue where the threshold values for the threshold brush tool were reset when switching tools.
  • Fixed an issue where units for measurements were incorrectly saved as mm in SR reports measurements made on images missing PixelSpacing information.
  • Fixed an issue where single-frame data missing the PhotometricInterpretation (0028,0004) would not open in the viewer.

Web UI

  • Resolved an issue for Flywheel-hosted SaaS Tenants where Tenant information could be missing and refreshing the web browser could lead to errors.
  • Added a performance optimization to improve initial page load time after user login or browser refresh.
  • Fixed a minor issue in the Flywheel Core web app where the user's auth token was being repeatedly refreshed unnecessarily.