Gear-Based Workflow Automation
Overview
Gear-based workflows enable you to automate data processing in Flywheel using containerized analysis tools called gears. By configuring gear rules, batch operations, and job priorities, you can create efficient, automated pipelines that process data as it arrives in your projects.
Key Concepts
Gears
Gears are containerized applications that perform specific data processing tasks. They can convert file formats, extract metadata, run analysis algorithms, and generate quality control reports.
Gear Rules
Gear rules automatically trigger gears to run when files matching specific criteria are uploaded or modified. This eliminates manual gear execution and ensures consistent processing across your datasets.
Learn how to create and manage gear rules
Batch Processing
Run gears on multiple datasets simultaneously to improve efficiency and reduce processing time. Batch operations allow you to select multiple sessions or acquisitions and execute the same gear on all of them at once.
Job Prioritization
Control the execution order of gear jobs by setting priority levels (critical, high, medium, low). Prioritization helps ensure time-sensitive analyses run first while less urgent jobs wait in the queue.
Learn about job prioritization
Jobs Log
Monitor and manage all gear executions through the Jobs Log. View job status, inputs, outputs, configuration, logs, and perform actions like canceling or retrying jobs.
Common Workflows
Automated DICOM Processing
- Configure gear rules to automatically run when DICOM files are uploaded
- Chain gears together using file tags (e.g., File Metadata Importer → File Classifier → dcm2niix)
- Monitor job progress in the Jobs Log
Quality Control Pipeline
- Set up gear rules to run MRIQC on specific file types and classifications
- Use job priorities to ensure QC gears run before analysis gears
- Review QC outputs in the Analysis tab
Batch Reprocessing
- Select multiple sessions or acquisitions that need reprocessing
- Use batch run to execute gears on all selected data
- Set appropriate job priority based on urgency
- Monitor progress in the Jobs Log
Best Practices
- Test gear rules on a small dataset before enabling them project-wide
- Use exception criteria in gear rules to prevent infinite loops (e.g., exclude files that already have output tags)
- Set appropriate job priorities to balance throughput and latency
- Monitor the Jobs Log regularly to identify and resolve gear failures
- Use batch processing for large-scale reprocessing operations