cqfn/aibolit

Leverages a pre-trained machine learning model (pickle-based) to rank detected code patterns by their actual impact on cyclomatic complexity, rather than treating all violations equally. Extracts syntactic metrics from Java AST and uses the model to compute importance scores, enabling developers to prioritize refactoring efforts. Supports multiple output formats (compact, long, XML) with per-pattern line-level reporting and pattern suppression, and runs as a Python CLI tool or Docker container for CI/CD integration.