Module compilation 

Graph compilation and caching infrastructure.

This module provides ahead-of-time graph optimization and compilation capabilities:

• CompiledGraph: Optimized, executable representation of computation graphs
• GraphCompiler: Applies optimization passes and produces compiled graphs
• CompilationCache: Caches compiled graphs to avoid recompilation
• TlCompilableExecutor: Trait for executors that support graph compilation

Example

use tensorlogic_infer::compilation::{GraphCompiler, CompilationConfig, OptimizationLevel};
use tensorlogic_infer::DummyExecutor;
use tensorlogic_ir::EinsumGraph;

let mut compiler = GraphCompiler::new(CompilationConfig {
    optimization_level: OptimizationLevel::Aggressive,
    ..Default::default()
});

let graph = EinsumGraph::new();
let compiled = compiler.compile(&graph).unwrap();

Structs

CacheStats

Statistics for the compilation cache.

CompilationCache

Cache for compiled graphs.

CompilationConfig

Configuration for graph compilation.

CompilationKey

Cache key for compiled graphs.

CompilationStats

Statistics about the compilation process.

CompiledGraph

Compiled representation of a computation graph.

GraphCompiler

Graph compiler that applies optimization passes.

Enums

OptimizationLevel

Optimization level for graph compilation.

Traits

TlCompilableExecutor

Trait for executors that support graph compilation.