ARCHMAP Devlog:

Caching & GitHub integration

Two foundational pieces are now in place — a disk-backed cache to avoid hammering the API, and a clean GitHub layer for fetching repo trees and file contents.

Caching Layer

Disk-backed analysis cache with 1-hour TTL
Added a file system cache that auto-creates itself on startup. On read, it checks the file’s modification time and invalidates anything older than one hour — so data stays fresh without re-fetching every single time.

Simple key format
Cache keys are dead simple. Both read and write paths are wrapped in try/catch so a corrupted cache file won’t bring down the whole request.

GitHub Integration

Full recursive repo tree fetching
Wired up Octokit to fetch the entire file tree of any public repo in one shot. It auto-resolves the default branch first, grabs the latest commit’s tree SHA, then pulls the full recursive tree. No manual branch specification needed.

File content fetch with base64 decode
Pulls individual files and decodes them from base64 internally — callers just get plain UTF-8 strings back. If a file fails to fetch it returns an empty string instead of throwing, so nothing upstream needs to null-check.

Note: This code isn’t on GitHub yet — the demo isn’t ready and i would rather push everything together once there’s something worth showing end-to-end. Repo goes public when i develop a frontend atleast