Plang

Plang Programming Language

Overview

Plang is a natural language programming language that uses pseudo-code syntax and goal-based architecture. Each file represents a "goal" containing steps written in natural language. Use this skill when working with Plang code generation, debugging, or architectural guidance.

Quick Start

For comprehensive syntax and examples, read the appropriate reference files:

• Basic syntax: Read syntax.md for variable handling, database operations, HTTP requests, control flow
• Common patterns: Read patterns.md for proven implementation patterns
• Database schemas: Read database.md for table creation and data modeling

Core Concepts

Goal Structure

Goals are Plang's fundamental unit. Each goal is a file containing steps:

GoalName
- step 1 description
- step 2 description  
- step 3 description

Natural Language Steps

Steps are written in natural language with specific patterns Plang recognizes:

ProcessOrder
- select * from orders where id=%orderId%, return 1, write to %order%
- if %order% is empty then
  - throw "Order not found"
- call goal ValidateOrder
- update orders set status='processed' where id=%orderId%

Key Design Principles

1. SQLite-first: Presume SQLite for all database operations unless specified
2. Goal-based: Break functionality into discrete goals (files)
3. Natural language: Write steps as readable instructions
4. Variable passing: Use %variableName% syntax throughout
5. Error handling: Use on error clauses and validation

When to Generate Plang Code

Generate Plang goals when the user:

• Asks to "create a Plang goal for..."
• Requests database operations (Plang excels at CRUD)
• Needs API integration code
• Wants to build web applications with UI
• Requires LLM integration
• Needs file processing workflows
• Wants event-driven architectures

Code Generation Guidelines

1. Start with Clear Goal Names

CreateUser          ✓ Clear, action-oriented
ProcessPayment      ✓ Verb + noun pattern
GetUserData         ✓ Specific purpose
DoStuff             ✗ Too vague
Main                ✗ Not descriptive

2. Use Proper Variable Syntax

✓ %userId%
✓ %order.id%
✓ %user.email%
✓ %request.body.name%
✓ %Settings.ApiKey%
✓ %now%

✗ userId (missing % delimiters)
✗ %user-id% (use camelCase or underscores)

3. Follow Database Conventions

✓ - select * from users where id=%userId%, return 1, write to %user%
✓ - insert into orders, status='pending', amount=%total%, write to %orderId%
✓ - update products set price=%newPrice% where id=%productId%

✗ - query the users table (too vague)
✗ - get user by id (missing proper syntax)

4. Structure Complex Goals Properly

For complex operations, break into sub-goals with option of parameters:

ProcessOrder
- call goal ValidateOrder
- call goal CalculateTotal id=%contractId%
- call goal CreateTransaction
- call goal SendConfirmation

5. Use Transactions Appropriately

ProcessPayment
- begin transaction "users/%userId%"
- insert into orders ...
- insert into orderItems ...
- call goal ChargeCard
    on error call RollbackOrder
- end transaction

Common Task Patterns

Database CRUD

CreateRecord
- insert into table, field1=%value1%, field2=%value2%, write to %id%

ReadRecord  
- select * from table where id=%id%, return 1, write to %record%

UpdateRecord
- update table set field=%newValue% where id=%id%

DeleteRecord
- delete from table where id=%id%

API Integration

CallExternalApi
- post %Settings.ApiUrl%/endpoint
    headers:
        "X-API-key": "%Settings.ApiKey%"
    data: {
        "field": "%value%"
    }
    on error call HandleApiError
    write to %result%

Form Processing

HandleFormSubmission
- validate %request.body.email% is not empty, "Email required"
- validate %request.body.amount% is not empty and larger than 0
- insert into submissions, email=%request.body.email%, amount=%request.body.amount%
- [ui] render "success.html", navigate

File Processing

ImportCSV
- read data.csv, first row is header, write to %rows%
- foreach %rows% call ProcessRow item=%row%

ProcessRow
- insert into imported_data, column1=%row.header1%, column2=%row.header2%

Debugging Plang Code

Common Issues

Variable not found:

• Ensure variables are written to before reading: write to %varName%
• Check variable scope (variables from sub-goals need to be returned)

Database errors:

• Verify table exists and columns match
• Check data types match column definitions
• Ensure foreign key relationships are valid

Goal not found:

• Check file path is correct (relative paths supported)
• Verify goal name matches filename (case-sensitive)

HTTP errors:

• Add error handlers: on error status code = 404, call HandleNotFound
• Check API authentication headers
• Verify URL and request body format

Debug Output

- write out "Debug: %variableName%"
- write out to system "System log: %value%"
- write out to user log "User-visible message"

Architecture Patterns

Microservice-Style Goals

Organize by feature domain:

/user/
  Create.goal
  Login.goal
  Update.goal
/order/
  Create.goal
  Process.goal
  Cancel.goal
/payment/
  Charge.goal
  Refund.goal

Event-Driven Pattern

Events
- before each goal(including private) in /admin/.*, call CheckAdmin
- on app error, call goal HandleError
- before each goal in /api/.*, call RateLimitCheck

Multi-Datasource Pattern

/ Setup.goal - 'data' datasource created automatically
Setup
- create table.....

/ Analytics.goal
Analytics
- create datasource "analytics"
- create table.....

/ UserSetup.goal
UserSetup
- create datasource "users/%userId%"
- create table.....

Query
- select * from main.orders o join analytics.analysis a on a.orderId=o.id
    datasource: "data", "analytics"
    write to %results%

Note: For sqlite, attaching two datasource, the first datasource uses main. prefix

Reference Files

This skill includes comprehensive reference documentation:

references/syntax.md

Complete syntax reference covering:

• Variable handling and operations
• Database operations (SELECT, INSERT, UPDATE, DELETE)
• HTTP operations (GET, POST with error handling)
• File operations (read, write, CSV, Excel, JSON)
• Control flow (if/then, loops, foreach)
• UI operations (rendering, DOM manipulation)
• LLM integration
• Events and goal control

references/patterns.md

Proven implementation patterns:

• Database patterns (transactions, upsert, pagination)
• API integration patterns (auth, retry, pagination)
• Error handling patterns
• Payment processing patterns
• UI patterns (forms, modals, progressive enhancement)
• LLM processing patterns
• File processing patterns
• Authentication patterns
• Email patterns
• Testing patterns

references/database.md

Database schema patterns:

• Table creation syntax
• Index patterns
• Column modifications
• Common schemas (users, orders, transactions, analytics)
• Data integrity patterns
• Multi-database queries

Best Practices

1. Read references first: Before generating code, read the relevant reference file
2. Use transactions: For multi-step database operations, wrap in transactions
3. Validate inputs: Always validate user input and required parameters
4. Handle errors: Add error handlers for external calls (APIs, file operations)
5. Use meaningful names: Goal names and variables should be self-documenting
6. Break down complexity: Split complex logic into multiple goals
7. Leverage caching: Use cache for expensive operations (set cache, get cache)
8. Test incrementally: Build and test goals one at a time

Example: Complete Feature Implementation

Here's a complete example showing best practices:

CreateOrder
- validate %user.id% is not empty, "User must be logged in"
- validate %cart% is not empty, "Cart cannot be empty"
- begin transaction "users/%user.id%"
- insert into orders, status='pending', 
    amount=%cartTotal%, created=%now%
    write to %orderId%
- foreach %cart% call CreateOrderItem item=%cartItem%
- call goal ProcessPayment
    on error call HandlePaymentError
- update orders set status='paid' where id=%orderId%
- end transaction
- call goal SendOrderConfirmation
- [ui] render "order_success.html", navigate

CreateOrderItem
- insert into orderItems, orderId=%orderId%, 
    productId=%cartItem.productId%, 
    quantity=%cartItem.quantity%,
    price=%cartItem.price%

This example demonstrates:

• Input validation
• Transaction management
• Error handling
• Goal composition
• Database operations
• UI rendering