Logic Steps

The logic section is an ordered list of steps that execute sequentially. Each step is one of three types: SQL, Lark, or Expr.

logic:
  - sql: ...       # query your database
  - lark: ...      # transform with Starlark
  - expr: ...      # lightweight expression

Every step has an omit field that controls whether its result appears in the response:

omit	Result
false (default)	Included in response
true	Available to later steps but not returned

---

SQL Step

Runs a SQL query against the configured database.

- sql: |
    SELECT * FROM users ORDER BY id DESC
  omit: true

Parameter Binding

Use ?:field_name to bind values from a previous step or from the request input:

- sql: |
    INSERT INTO users VALUES(?:id, '?:name', '?:email', now('UTC'))
  omit: true

String fields require quotes around the placeholder. Numeric fields do not:

# Numeric — no quotes
?:id

# String — single quotes required
'?:name'
'?:email'

Values are resolved from the most recent step result that contains that field name.

---

Lark Step (Starlark)

Runs Starlark code — a deterministic, sandboxed subset of Python. Use Lark for loops, custom functions, and complex transformations.

- lark: enhance_data
  omit: true
  code: |
    load("math", "upby10")
    load("strings", "slugify")

    def mask_email(email):
        parts = email.split("@")
        return parts + "***@" + parts[1]

    result = []
    for row in step_0:
        result.append({
            "id": upby10(row["id"]),
            "name": slugify(row["name"]),
            "email": mask_email(row["email"]),
        })

The final value assigned to result becomes the step's output.

Built-in Modules

load("math", "upby10")       # Math utilities
load("strings", "slugify")   # String utilities

Accessing Previous Steps

step_0   # result of first step
step_1   # result of second step

Accessing Request Input

input.field_name   # field from request body or query params

---

Expr Step

A lightweight expression evaluator for simple transformations. Use Expr when you don't need loops or custom functions.

- expr: build_response
  code: |
    {
      "version": input.version,
      "total_users": len(step_0),
      "has_data": len(step_0) > 0
    }

The expression must be a single value — an object, array, or scalar. The result becomes the step output.

Available Variables

Variable	Description
input	Request body or query parameters
step_N	Result of step N

Built-in Functions

Function	Description
len()	Length of array or string

Expr vs Lark

Use Case	Use
Build a response object	expr
Count rows, check length	expr
Loop over rows	lark
Define custom functions	lark
Complex multi-step transforms	lark