Codebase Context

Purpose

Rhedyn is a small static site generator/asset pipeline. It reads a config file, runs tasks (optionally in parallel per step), and writes outputs to opts.outDir with caching based on file hashes and state access.

Entry Point and Flow

src/index.js is the CLI entrypoint. It loads config (user or defaults), then iterates through steps in tasks, running each task group in parallel.
src/lib.js owns config loading, task expansion (file/state), caching, and the task runner that calls action functions.

Configuration (current shape)

src/defaults.js exports tasks and opts used by src/index.js.
Tasks are objects (or arrays of objects for parallelism) with:
- jobConfig for file/state selection, expansion, output paths, and caching
- actionConfig for action-specific options
- action function and a name
Global config under opts controls output/cache dirs, logging, and site meta.

Task Expansion

File-based tasks expand input files via glob patterns (readFilesByGlob), driven by jobConfig.inputFiles.
State-based tasks expand via selectors into meta.resources, with optional pagination and auto-built output paths from jobConfig settings.
Tasks can opt out of expansion (expand: false) to run once with aggregated inputs.

Caching

src/cache.js stores cache files per job. Cache keys are based on job IDs, file hashes, and accessed state paths.
createTrackedObject wraps state/config so accessed properties are tracked and used as cache dependencies.

Actions

src/actions/ contains built-ins:
- compileSass, renderMarkdownToHtml, renderTemplate, renderMarkdownWithTemplate, renderIndex, optimiseSvg, copy, imageToWebP, generateFavicons, generateTaxonomy.
Actions receive { config, jobConfig, meta } and return { detail, paths, deps, ref }. The config value maps to actionConfig.
Returned ref results are exposed via meta.resources for later tasks.

Utilities

src/util/* provides filesystem helpers, path manipulation, and simple object/state utilities.
src/logging.js formats log output and implements log levels.