Input Templating System

ℹ️ Note: Workflow nodes generate output fields at runtime. The templating system manages these dynamic values and their relationships between nodes.

Block Definition Model

Block Structure

interface Block {
  // Input field definitions
  inputs: {
    [key: string]: {
      type: string;
      required: boolean;
      description: string;
    };
  };
  
  // Output field definitions
  outputs: {
    [key: string]: {
      type: string;
      description: string;
    };
  };
}

Frontend Implementation

Workflow Hierarchy

The system uses a graph-based structure to manage node relationships:

graph TD
  A[Parent Node] --> B[Child Node 1]
  A --> C[Child Node 2]
  B --> D[Grandchild Node]
  C --> D

1. Field Discovery

   • Performs BFS (Breadth-First Search) on ascendant nodes

   • Collects available output fields

   • Maps fields to their source nodes

2. Template Generation

   • Creates template strings for each field

   • Format: {{node_id-value}}

   • Example: {{node-123-output_field}}

Backend Processing

Kafka Workflow Task

interface WorkflowTask {
  // Maps template keys to their runtime values
  templateValues: {
    [key: string]: any;
  };
  
  // Node execution status
  nodeStatus: {
    [nodeId: string]: {
      status: 'pending' | 'running' | 'completed' | 'failed';
      output?: any;
    };
  };
}

Template Resolution Process

1. Value Collection

// Example template map
const templateValues = {
  'node-123-output': 'processed value',
  'node-456-result': 42
};

1. Template Replacement

// Input with templates
const input = {
  field1: '{{node-123-output}}',
  field2: '{{node-456-result}}'
};

// Resolved input
const resolved = {
  field1: 'processed value',
  field2: 42
};

Implementation Example

Sample Workflow

// Node definition
const textProcessingNode = {
  id: 'text-processor',
  inputs: {
    text: {
      type: 'string',
      required: true,
      description: 'Text to process'
    }
  },
  outputs: {
    processedText: {
      type: 'string',
      description: 'Processed text result'
    },
    wordCount: {
      type: 'number',
      description: 'Number of words in text'
    }
  }
};

// Template usage
const childNode = {
  inputs: {
    content: '{{text-processor-processedText}}',
    metrics: '{{text-processor-wordCount}}'
  }
};

⚠️ Warning: Remember to handle cases where template values might not be available yet, as nodes execute asynchronously.

Best Practices

Tips for Template Usage

1. Unique Node IDs

   • Ensure node IDs are unique across the workflow

   • Use descriptive IDs for better debugging

2. Error Handling

   • Validate template syntax before execution

   • Handle missing or undefined values gracefully

3. Performance

   • Cache template resolution results when possible

   • Minimize template string parsing operations