Expand description
Verificar - Synthetic Data Factory for Domain-Specific Code Intelligence
Verificar is a unified combinatorial test generation and synthetic data factory
that serves multiple transpiler projects (depyler, bashrs, ruchy, decy). It generates
verified (source, target, correctness) tuples at scale, creating training data
for domain-specific code intelligence models.
§Architecture
┌─────────────────────────────────────────────────────────────┐
│ VERIFICAR CORE │
├─────────────────────────────────────────────────────────────┤
│ Grammar → Generator → Mutator → Oracle │
│ Definitions Engine Engine Verification │
└─────────────────────────────────────────────────────────────┘§Quick Start
use verificar::generator::{Generator, SamplingStrategy};
use verificar::Language;
// Create a generator for Python
let generator = Generator::new(Language::Python);
// Generate test cases using coverage-guided sampling
let strategy = SamplingStrategy::CoverageGuided {
coverage_map: None,
max_depth: 3,
seed: 42,
};
let test_cases = generator.generate(strategy, 100);§Modules
grammar- Language grammar definitions (tree-sitter, pest PEGs)generator- Combinatorial program generation enginemutator- AST mutation operators (AOR, ROR, LOR, BSR, etc.)oracle- Verification oracle (sandbox execution, I/O diffing)data- Data pipeline (Parquet output)ml- ML model training (bug prediction, embeddings)
Re-exports§
Modules§
- data
- Data pipeline for storing verified test cases
- error
- Error types for Verificar
- generator
- Combinatorial program generation engine
- grammar
- Grammar definitions for source languages
- ml
- ML model training pipeline
- mutator
- AST mutation operators
- oracle
- Verification oracle for transpilation correctness
- prelude
- Prelude module for convenient imports
- transpiler
- Transpiler integration
Enums§
- Language
- Supported source languages for generation