sift domains

​

Overview

​

Usage

sift domains

​

Options

​

Behavior

• Name - Domain identifier used with --domain-name
• Description - Purpose and use case
• Entities - Number of defined entity types
• Relations - Number of defined relation types

​

Output Format

┌─────────────────────────────────────────┐
│       Available Domains                 │
├──────────────┬─────────────┬──────┬─────┤
│ Name         │ Description │ Ent  │ Rel │
├──────────────┼─────────────┼──────┼─────┤
│ schema-free  │ Discover... │ 0    │ 0   │
│ general      │ General...  │ 15   │ 20  │
│ osint        │ OSINT...    │ 25   │ 30  │
│ academic     │ Academic... │ 12   │ 18  │
└──────────────┴─────────────┴──────┴─────┘

Usage: sift extract ./docs --domain-name osint
Custom: sift extract ./docs --domain path/to/domain.yaml

​

Bundled Domains

​

schema-free

• No predefined entity or relation types
• LLM discovers schema during extraction
• Best for exploratory analysis
• Saved to discovered_domain.yaml

​

general

• PERSON
• ORGANIZATION
• LOCATION
• DATE
• EVENT

​

osint

• PERSON
• ORGANIZATION
• LOCATION
• IP_ADDRESS
• URL
• PHONE_NUMBER
• EMAIL

​

academic

• AUTHOR
• PUBLICATION
• INSTITUTION
• CONCEPT
• DATASET
• METHOD

​

Usage Examples

​

List domains

sift domains

​

Use a bundled domain

sift extract ./docs --domain-name osint

​

Use schema-free domain

sift extract ./docs --domain-name schema-free

​

Custom Domains

sift extract ./docs --domain path/to/custom-domain.yaml

​

Custom Domain Format

name: my-domain
description: Custom domain for legal documents

entity_types:
  CASE:
    description: Legal case
  STATUTE:
    description: Legal statute or law

relation_types:
  CITES:
    source_types: [CASE]
    target_types: [CASE, STATUTE]

​

Domain Configuration

​

Set Default Domain

domain: osint  # or path/to/domain.yaml

export SIFT_DOMAIN_PATH=osint

​

Priority Order

1. --domain CLI flag
2. SIFT_DOMAIN_PATH environment variable
3. domain in sift.yaml
4. Default: schema-free

​

Domain Selection Guide

​

When to use schema-free

• Exploring new document types
• Unknown domain structure
• Prototyping
• Maximum flexibility

​

When to use general

• Business documents
• News articles
• General reports
• Common entity types

​

When to use osint

• Intelligence gathering
• Security analysis
• Threat research
• Technical indicators

​

When to use academic

• Research papers
• Scientific literature
• Academic citations
• Knowledge synthesis

​

When to use custom

• Domain-specific requirements
• Specialized entity types
• Custom relation schemas
• Strict validation needed

​

Output

​

See Also

• init - Initialize project with domain
• extract - Use domain for extraction