Submitting New Workflows¶
The Submit Panel is the main interface for creating and submitting computational jobs in the system. It provides a comprehensive workflow that guides users through model selection, parameter configuration, and input data specification. To submit a new workflow:
Navigate to the submission page
Select the workflow type from the available options
Fill in the required parameters in the form - Forms are dynamically generated based on the workflow type - Required fields are marked with an asterisk (*) - Help text provides guidance for each parameter
Click “Submit” to start the workflow
You’ll be redirected to the dashboard where you can monitor the task’s progress
The submission forms use React JSON Schema Form, which provides:
Validation: Immediate feedback on invalid inputs
Conditional Fields: Fields that appear based on other selections
File Uploads: For workflows that require input files
Advanced Controls: Sliders, dropdowns, and other interactive elements
Special Fields: for complex objects, e.g. molecular data
The Submit Panel consists of several key sections organized in a logical workflow:
Model Selection (CascadingNodeSelector)
Job Configuration (Custom Name & Category)
Parameter Settings (ParametersAccordion)
Input Data Configuration (Input Mappings)
Template Management (TemplateSelector)
Job Submission
Node Selector¶
The CascadingNodeSelector is the primary interface for browsing and selecting computational models from the available node registry.
Hierarchical model browser with favorites and tree structure¶
- Hierarchical Model Organization
Models are organized in a tree structure based on their function mapping, making it easy to navigate through different categories and modules.
Example of hierarchical model organization which mirrors the directory structure of your project¶
- Favorites System
Users can mark frequently used models as favorites for quick access:
Star icon checkbox for marking/unmarking favorites
Dedicated “Favorites” section at the top of the tree
Persistent favorite status across sessions
Once a model is selected, the ParametersAccordion and Input sections dynamically update to reflect the chosen model’s requirements.
Parameter Settings¶
The ParametersAccordion section provides comprehensive job parameter configuration options.
Complete parameters configuration interface (for a slurm-queue job)¶
- Resource Selection
Configure computational resources for job execution:
Dropdown selection from allowed resources
Default option available if no specific resource needed
Dynamic loading of available resources from the system
- Queue Configuration
Specify execution queue for job scheduling:
Selection from allowed queues list
When the SLURM queue is selected, additional SLURM parameters become available
- SLURM Parameters (when using “slurm-queue”)
Additional SLURM-specific configuration options become available when the SLURM queue is selected, including:
Job allocation parameters
Resource requirements
Time limits
Node specifications
- Execution Options
Control job execution behavior:
Force Rerun: Force execution even if cached results exist
Recompute Artifacts: Regenerate all output artifacts regardless of cache status
- Custom Name & Category Editor
Provides job identification and organization:
- Custom Name Field
User-defined name for the job
Defaults to “submitted-job”
Automatically saved to localStorage per operation
- Category Field
Organizational category for the job
Defaults to “submission”
Persistent storage with operation-specific recall
Model Input Section¶
The Input section handles data configuration for each input mapping required by the selected model. For each input mapping defined by the selected model:
Individual input mapping configuration card¶
- Dynamic Form Generation
Automatically generated forms based on input mapping schema
Real-time validation and data binding
Memoized form components for optimal performance
- Template Management
Load existing input templates
Save current configuration as template
Update existing templates on submission
Template-specific parameter inheritance