/ontology-objecttype - ObjectType Definition Assistant

Version: 3.0.0 | Model: opus

Overview

Aspect	Description
Purpose	Analyze codebases and define ObjectTypes for ODA migration
Output	L1: Candidate summary / L2: Per-class analysis / L3: Full YAML definitions
Reference	`.claude/rules/ontology.md` for naming conventions

Cross-Skill Integration

Skill	Relationship
`/ontology-core`	Validates generated ObjectType definitions
`/ontology-why`	Called for "why?" questions about design decisions
`/ontology-linktype`	(Planned) LinkType definition after ObjectType completion

Commands

Command	Description
`analyze <path>`	Analyze project/file for ObjectType candidates
`resume <session-id>`	Resume previous analysis session
`help`	Show usage information

Detection Patterns

Pattern	Detection Method	Example
Python class	`class ClassName:`	`class Employee:`
SQLAlchemy ORM	`Base` or `declarative_base()`	`class User(Base):`
Django ORM	`models.Model`	`class Article(models.Model):`
Pydantic	`BaseModel`	`class Config(BaseModel):`

Interactive Workflow (Phase 1-4)

Phase 1: Context Clarification

Purpose: Clarify source type and business domain

python

AskUserQuestion(questions=[
    {
        "question": "What is your source for this ObjectType definition?",
        "header": "Source Type",
        "options": [
            {"label": "Existing source code", "description": "Python, Java, TypeScript"},
            {"label": "Database schema", "description": "SQL DDL, ORM models"},
            {"label": "Business requirements", "description": "Requirements documents"},
            {"label": "Manual definition", "description": "New domain model"}
        ]
    }
])

Validation Gate: source_validity

Phase 2: Entity Discovery

Purpose: Extract entities + PK strategy + Property type mapping

Scan source code for entity candidates
Select Primary Key strategy (single_column / composite / composite_hashed)
Map Python types to Foundry DataTypes

Validation Gates: candidate_extraction, pk_determinism

Phase 3: Link Definition

Purpose: Detect relationships and define cardinality

Cardinality	FK Location	Backing Table
ONE_TO_ONE	Either side	No
ONE_TO_MANY	"Many" side	No
MANY_TO_ONE	"Many" side	No
MANY_TO_MANY	-	Yes (required)

Validation Gate: link_integrity

Phase 4: Validation & Output

Purpose: Final validation and YAML generation

Validation Gate: semantic_consistency

Validation Gates

Gate	Phase	Description
`source_validity`	1	Source path accessible, domain context provided
`candidate_extraction`	2	At least one entity with class name and properties
`pk_determinism`	2	PK is STRING, required, follows determinism rules
`link_integrity`	3	M:N has join table, FK references valid targets
`semantic_consistency`	4	API names follow conventions, all types valid

PK Strategy Options

Strategy	Description	Use Case
`single_column`	Use existing unique column	`employee_id`, `user_uuid`
`composite`	Combine columns with separator	`company_id + dept_id`
`composite_hashed`	SHA256 hash of combined columns	Long composite keys

DataType Mapping

Category	Types
Primitive	STRING, INTEGER, LONG, FLOAT, DOUBLE, BOOLEAN, DECIMAL
Temporal	DATE, TIMESTAMP, DATETIME, TIMESERIES
Complex	ARRAY (requires itemType), STRUCT (requires fields), JSON
Spatial	GEOPOINT, GEOSHAPE
Media	MEDIA_REFERENCE, BINARY, MARKDOWN
AI/ML	VECTOR (requires dimension)

Output Format

L1 - Candidate Summary

Found 12 Entity Candidates
  ObjectType candidates (8): Employee, Department, Project...
  Review needed (2): DTO/Mixin patterns
  Excluded (2): Helper/Config classes

L2 - Per-Class Analysis

yaml

entity_name: Employee
pk_strategy: single_column
pk_column: employee_id
properties:
  - api_name: employeeId
    data_type: STRING
    required: true
    is_pk: true

L3 - Full YAML Definition

yaml

# objecttype-Employee.yaml
api_name: Employee
display_name: Employee
primary_key:
  source_columns: [employee_id]
  strategy: single_column
properties:
  - api_name: employeeId
    data_type: STRING
    required: true
links:
  - link_type_name: EmployeeToDepartment
    target_object_type: Department
    cardinality: MANY_TO_ONE

Session State

json

{
  "session_id": "obj-a1b2c3",
  "target_path": "/home/palantir/my-project",
  "current_phase": "phase_2_entity",
  "phase_results": {
    "phase_1_context": { "status": "completed" },
    "phase_2_entity": { "status": "in_progress" },
    "phase_3_link": { "status": "pending" },
    "phase_4_output": { "status": "pending" }
  }
}

Reference System

Local Reference (ontology-definition package)

/home/palantir/park-kyungchan/palantir/Ontology-Definition/
├── ontology_definition/types/    # Type definitions
├── ontology_definition/core/     # Enums, base classes
└── tests/                        # Usage examples

External References (WebSearch/WebFetch)

Trusted sources only:

palantir.com/docs/* (Official documentation)
github.com/palantir/* (Official repositories)
Palantir case studies and conference materials

Error Handling

Error	Recovery
Path not found	Prompt for correct path
No Python files	Show available file types
No classes found	Suggest manual definition
Session expired	Start new or resume with `--resume`

Version History

Version	Change
1.0.0	Initial ObjectType definition assistant
1.1.0	Phase 1-4 interactive workflow, L1/L2/L3 output
3.0.0	Frontmatter normalization, MCP tool inclusion, duplicate removal

End of Skill Definition

Search AI Tools

ontology-objecttype

Install this agent skill to your Project

SKILL.md