Readable rules.
A preserved foundation.

What is RuleScript?

RuleScript is the original domain-specific language proposed for expressing data transformations and validations in readable text. Its core idea separates the declaration of a rule from the environment that executes it.

The examples use quoted field names, explicit operation keywords and optional error messages. They are useful as a language reference; edge cases and actual support require a defined runtime.

Syntax & examples

1. Transform a record

rename "price_usd" to "price";
split "full_name" into "first_name", "last_name" by " ";
map "country_code" [US: "United States", CA: "Canada", MX: "Mexico"];

2. Validate values

validate "price": number, min 0, max 10000, on error "Price must be valid";
validate "country": required, in ["United States", "Canada", "Mexico"], on error "Invalid country";

3. Combine explicit steps

Apply the intended rename and mapping before checking the resulting field values. Preserve the original data and make the transformation order explicit.

rename "price_usd" to "price";
validate "price": number, min 0, max 10000, on error "Price must be valid";

4. Conditional validation

validate "discount": number, on error "Discount must be numeric", if "price" > 100;

The historical example omitted its final semicolon; it is normalized here. The order of optional clauses and the behavior of missing fields remain runtime specification questions.

Transformations

rename "price_usd" to "price";

split "full_name" into "first_name", "last_name" by " ";

merge "first_name", "last_name" into "full_name" with " ";

map "country_code" [US: "United States", CA: "Canada", MX: "Mexico"];

convert "quantity" to integer;

calculate "total" as "price * quantity";

normalize "name" to lowercase;

replace "status" null with "N/A";

round "price" to 2 decimal places;

anonymize "email" with "****@domain.com";

drop "temp_field";

impute "age" missing with average;

reorder columns "id", "name", "email";

filter "age" > 18;

Validation functions

validate "name": required, on error "Name is required";

validate "age": number, on error "Age must be numeric";

validate "price": number, min 0, max 10000, on error "Price must be valid";

validate "country": in ["United States", "Canada", "Mexico"], on error "Invalid country";

validate "discount": > 0, on error "Discount must be positive";
validate "price": >= 0, on error "Price must be non-negative";

validate "age": < 18, on error "Age must be below 18";
validate "discount": <= 50, on error "Discount must be 50% or less";

validate "status": equal "Active", on error "Status must be active";
validate "status": not in ["Inactive", "Suspended"], on error "Invalid status";

Historical workflows

CLI

The original design used .rs source files and .rsx compiled artifacts. These are historical command shapes; no installer is provided.

rg-compile rules.rs
rg-run rules.rsx input.csv output.csv

API

The documentation proposed CSV batches and single JSON records. The routes below are historical sketches, not working endpoints. Multipart uploads must carry an actual file part, not the JSON filename placeholder shown on the former site.

CSV payload

POST /api/v1/validate/my_rules.rsx
Content-Type: multipart/form-data; boundary=example

--example
Content-Disposition: form-data; name="file"; filename="input.csv"
Content-Type: text/csv

price,country
25,US
--example--

Single JSON record

POST /api/v1/validate/my_rules.rsx
Content-Type: application/json

{"record":{"full_name":"Sample Person","price":-50,"country_code":"US"}}

{"status":"error","errors":[{"field":"price","error":"Price must be between 0 and 10000"}]}

Batch processing checks a complete delivery; record validation checks an event or submission. A future implementation must specify error handling and consistency across both interfaces.

Pipelines

rg-transform step1.rsx input.csv \
  | rg-transform step2.rsx \
  | rg-validate rules.rsx --output output.csv

This preserves the original pipeline concept. A runtime would need a defined stream format and exit-code behavior before this can be used reliably.

Private deployment

On-premises execution was a deployment direction on the old site. It is not an available enterprise offering announced here, and does not by itself establish compliance with any regulation.

Try two checks in your browser.

This small browser demonstration checks a numeric price from 0 to 10000 and a country code from US, CA or MX. It is hand-written JavaScript for these two checks, not a RuleScript parser or the future engine. Your input stays in this page.

What changes in the migration?

• RuleScript remains the language name. The .rs and .rsx examples are preserved as documentation; changing the website does not change their execution semantics.
• Curly quotes are normalized for copying. The conditional example now includes a closing semicolon.
• The 14 transformations and 7 validation groups are preserved. CLI and API sketches are labeled with their status.
• Masking is described accurately; it is not presented as a guarantee of anonymization.