Extract Information Node

Overview #

The Extract Information node is a specialized AI node designed to extract specific information from unstructured text using advanced AI models. It allows users to define custom extraction fields and receive structured data output.

Key Features #

• Custom field definition
• Support for both single values and lists
• Intelligent pattern recognition
• Flexible data type handling
• Format preservation
• Context-aware extraction

Node Configuration #

Basic Setup #

1. Node Addition:

   Workflow Builder → AI Actions → Extract Information

2. Required Fields:

• Node Name (optional, for identification)
• Input Content (text to analyze)
• Extraction Fields (what to extract)

Field Configuration #

Field Properties #

{
    "name": "field_name",
    "description": "What to extract",
    "type": "string|number|boolean|array",
    "isList": true|false
}

Supported Data Types #

• string: Text content
• number: Numerical values
• boolean: True/false values
• array: Lists of items

Usage Examples #

1. Basic Contact Information #

Field Configuration:
- Name: "email"
  Description: "Extract email addresses"
  Type: string

- Name: "phone"
  Description: "Extract phone numbers"
  Type: string

- Name: "addresses"
  Description: "Extract physical addresses"
  Type: array
  IsList: true

2. Product Information #

Field Configuration:
- Name: "product_name"
  Description: "Extract product name"
  Type: string

- Name: "price"
  Description: "Extract price in numbers"
  Type: number

- Name: "features"
  Description: "Extract product features"
  Type: array
  IsList: true

3. Document Analysis #

Field Configuration:
- Name: "dates"
  Description: "Extract all dates mentioned"
  Type: array
  IsList: true

- Name: "names"
  Description: "Extract person names"
  Type: array
  IsList: true

- Name: "key_points"
  Description: "Extract main points"
  Type: array
  IsList: true

Working Process #

1. Input Processing #

graph TD
    A[Input Text] --> B[Text Preprocessing]
    B --> C[Context Analysis]
    C --> D[Field Mapping]
    D --> E[Extraction Process]

2. Extraction Flow #

1. Text Analysis

• Content parsing
• Structure identification
• Pattern recognition

1. Field Matching

• Pattern matching
• Context evaluation
• Type validation

1. Data Extraction

• Value extraction
• Type conversion
• Format validation

1. Output Formatting

• Data structuring
• Type enforcement
• List processing

Output Format #

Standard Output Structure #

{
    "field_name1": "extracted_value",
    "field_name2": 123,
    "field_name3": ["item1", "item2", "item3"],
    "field_name4": true
}

Sample Response #

{
    "email": "john.doe@example.com",
    "phone": "+1-555-123-4567",
    "addresses": [
        "123 Main St, City, State 12345",
        "456 Side Ave, Town, State 67890"
    ]
}

Best Practices #

Field Definition #

1. Clear Descriptions

• Be specific about what to extract
• Include format requirements
• Specify any constraints

1. Appropriate Types

• Use correct data types
• Consider list vs single value
• Match expected format

1. Naming Conventions

• Use descriptive names
• Maintain consistency
• Avoid special characters

Input Preparation #

1. Text Formatting

• Clean input text
• Remove irrelevant content
• Maintain structure

1. Content Organization

• Group related information
• Maintain context
• Preserve relationships

Error Handling #

Common Issues and Solutions #

Issue	Cause	Solution
No Data Extracted	Unclear description	Improve field description
Wrong Data Type	Type mismatch	Verify field type configuration
Missing Values	Content not found	Check input text coverage
Invalid Format	Format mismatch	Specify format requirements

Error Messages #

Error Types:
- FIELD_NOT_FOUND: Required field not found in text
- TYPE_MISMATCH: Extracted data doesn't match specified type
- FORMAT_ERROR: Data format validation failed
- EXTRACTION_FAILED: General extraction failure

Performance Optimization #

Best Practices #

1. Input Optimization

• Limit text length
• Remove unnecessary content
• Maintain relevant context

1. Field Configuration

• Limit number of fields
• Use specific descriptions
• Optimize field types

1. Processing Efficiency

• Group similar extractions
• Use appropriate models
• Cache common patterns

Integration Examples #

1. Form Processing #

Workflow:
Form Submission → Extract Information → Database Storage
Fields:
- Personal Information
- Contact Details
- Requirements

2. Document Analysis #

Workflow:
Document Upload → Text Extraction → Extract Information → Report Generation
Fields:
- Key Terms
- Important Dates
- Action Items

3. Email Processing #

Workflow:
Email Receipt → Extract Information → CRM Update
Fields:
- Customer Details
- Order Information
- Support Requirements

Troubleshooting Guide #

Diagnostic Steps #

1. Verify input text quality
2. Check field configurations
3. Validate data types
4. Review extraction patterns
5. Check model responses

Common Solutions #

1. No Data Extracted

• Improve field descriptions
• Check input text
• Verify field names

1. Wrong Data

• Review field types
• Check format specifications
• Validate input content

1. Performance Issues

• Optimize input length
• Reduce field count
• Improve descriptions

Additional Resources #

Documentation #

• Field configuration guide
• Data type reference
• Pattern matching guide
• Best practices guide

Support #

• Community forums
• Technical support
• Usage examples
• FAQ section

Remember to regularly test your extraction configurations and validate the output to ensure accurate and reliable data extraction.