MTLExecutionContext

@MainActor final class MTLExecutionContext

Overview

The execution context maintains all runtime state during template execution, including variable bindings, writer stack, indentation, protected areas, and model access. It coordinates between MTL templates and the underlying AQL expression evaluator.

Overview

The execution context provides:

• Variable scoping: Stack-based variable scopes with push/pop operations

• Expression evaluation: Integration with AQL for expression evaluation

• Indentation management: Stack-based indentation tracking

• Writer management: Stack of writers for nested file blocks

• Protected areas: Preservation of user code sections

• Trace support: Model-to-text traceability links

• Model access: Registration and access to input models

Variable Scoping

Variables are managed in a stack of scopes:

context.setVariable("x", value: 10)     // Global scope
context.pushScope()
context.setVariable("x", value: 20)     // Local scope (shadows global)
await context.getVariable("x")          // Returns 20
context.popScope()
await context.getVariable("x")          // Returns 10

Writer Stack

File blocks create new writers that are pushed onto a stack:

// Main output (stdout)
context.write("Top level")

// Open file block - pushes new writer
try await context.openFile(url: "output.txt", mode: .create, charset: "UTF-8")
context.write("File content")  // Goes to output.txt
try await context.closeFile()  // Pops writer, finalizes file

context.write("Top level again")  // Back to stdout

Example Usage

let module = MTLModule(...)
let strategy = MTLInMemoryStrategy()
let context = MTLExecutionContext(
    module: module,
    generationStrategy: strategy
)

// Register models
await context.registerModel("input", resource: inputModel)

// Execute template
context.setVariable("model", value: modelRoot)
let template = module.templates["generateClass"]!
// ... execute template body ...

// Finalize
try await context.finalize()