Rulemorph

A Rust CLI and library to transform CSV/JSON data into JSON using YAML rules.

Features

• Input formats: CSV and JSON with nested record extraction
• Rule-based mapping: Declarative YAML rules with static validation
• Expressions: String ops (concat, replace, trim), numeric ops (+, -, *, /), date formatting
• Lookups: Array lookups from external context data (lookup, lookup_first)
• Conditions: Conditional mapping with comparisons, regex, and logical ops
• DTO generation: Generate type definitions for Rust, TypeScript, Python, Go, Java, Kotlin, Swift
• MCP server: Available as a Model Context Protocol server for AI assistants

Installation

Homebrew (recommended)

brew install vinhphatfsg/tap/rulemorph

Quick Start

Transform user data from an external API response to your schema:

rules.yaml

version: 1
input:
  format: json
  json:
    records_path: "users"
mappings:
  - target: "id"
    source: "user_id"
  - target: "name"
    source: "full_name"
  - target: "email"
    expr:
      op: "concat"
      args:
        - { ref: "input.username" }
        - "@example.com"

input.json

{ "users": [{ "user_id": 1, "full_name": "Alice", "username": "alice" }] }

Run

rulemorph transform -r rules.yaml -i input.json

Output

[{ "id": 1, "name": "Alice", "email": "alice@example.com" }]

Rule Structure

version: 1
input:
  format: json|csv
  json:
    records_path: "path.to.array"  # Optional
mappings:
  - target: "output.field"
    source: "input.field"    # OR value: <literal> OR expr: <expression>
    type: string|int|float|bool
    when: <expression>       # Optional condition

For full rule specification, see docs/rules_spec_en.md (English) or docs/rules_spec_ja.md (Japanese).

DTO Generation

Generate type definitions from your rules:

rulemorph generate -r rules.yaml -l typescript

Output:

export interface Record {
  id: number;
  name: string;
  email: string;
}

Supported languages: rust, typescript, python, go, java, kotlin, swift

Library Usage (Rust)

use rulemorph::{parse_rule_file, transform};

let rule = parse_rule_file(&std::fs::read_to_string("rules.yaml")?)?;
let output = transform(&rule, &std::fs::read_to_string("input.json")?, None)?;

MCP Server

An MCP server (rulemorph-mcp) is included for AI assistant integration:

claude mcp add rulemorph -- rulemorph-mcp