Add configurable LRU caching for import resolver

[usefuulness]

Summary

• replace the import resolver's map cache with an LRU cache and expose a configurable maximum size
• add a VS Code command plus configuration hooks to clear the cache when template roots or settings change
• extend documentation and tests to cover cache eviction behaviour and new settings

Testing

• npm test
• npx jest src/static-analyzer/tests/import-resolver.test.ts --runInBand

---

https://chatgpt.com/codex/tasks/task_e_68c865e213988328b57fcd7f518c3fcb