$ show zazoomauro/node-dependency-injection
This repository is a Node.js/TypeScript library for dependency injection (DI), heavily inspired by Symfony's DI container. It provides a service container that manages object construction, autowiring (automatic dependency resolution from constructor parameter types), and service configuration via YAML, JSON, JavaScript, and XML files. The library supports advanced features such as compiler passes, service decoration, tagged and keyed services, conditional service registration, lazy loading, and graph export. It includes a CLI tool for generating configurations and inspecting the container. The architecture follows a modular design: a ContainerBuilder manages definitions and compilation, loaders parse configuration files, and a compiler runs passes in ordered phases (optimization, removal, decoration). Autowiring uses TypeScript's AST (via @typescript-eslint/typescript-estree) to scan source files and register services automatically. The library also provides validation, graph export in multiple formats (JSON, Mermaid, DOT), and a condition evaluation system for environment-aware service registration. The codebase is well-structured with clear separation of concerns, extensive tests (both JavaScript and TypeScript), and demonstrates several design patterns (compiler pass, strategy for loaders/dumpers, reference objects, condition evaluation). However, the project appears unmaintained: the last CHANGELOG entry is from 2018, dependencies are outdated (e.g., commander v8, chalk v4, TypeScript 4.6.4), and the package.json version is 0.0.0. Despite its age, the core concepts remain relevant and the code quality is solid for a library of its kind.
The LLM identified 2 files worth highlighting for reuse:
Two-phase condition evaluator that first checks environment variables and custom functions, then checks service existence. Demonstrates a reusable pattern for conditional configuration loading that can be adapted to any system needing pre- and post-processing checks. While internally dependent on Definition and Reference, the core logic for two-phase evaluation is isolated and can be extracted with minimal adaptation.
Exports the service dependency graph in JSON, Mermaid, and DOT formats. The core logic maps a set of nodes and edges to multiple output representations, demonstrating a strategy pattern for graph serialization that can be reused with any graph data structure. Although it imports internal types like Reference, the graph generation algorithm itself is independent and could be extracted with moderate effort.