@neurowire/core

The format authority for Neurowire: the canonical feed model, the serializers, and the merge logic. Pure TypeScript, no network and no DOM, with zod as the only runtime dependency.

Install

npm install @neurowire/core

What it gives you

• Canonical model (NeurowireFeed, NeurowireEntry, Mesh) plus the zod schemas (NeurowireFeedSchema, MeshSchema, ...). Every parser produces this model and every serializer consumes it. It carries list metadata only, not full article bodies.
• Serializers for four output formats, dispatched by serialize(feed, format):
  • atom: Atom 1.0 (application/atom+xml)
  • json: JSON Feed 1.1 (application/feed+json)
  • md: Markdown digest
  • nwf: Neurowire Feed, a compact line-oriented format (interned authors, tags and sources, relative links, delta timestamps)
• nwf round-trip and validation: fromNwf parses it back to the model, and validateNwf returns line-numbered diagnostics.
• Merge: mergeFeeds combines many feeds into one, tagged by source, deduped, newest first.

HTML is deliberately not a core format. It lives in @neurowire/web so core stays format-pure and dependency-light.

Usage

import { serialize, mergeFeeds, validateNwf, type NeurowireFeed } from '@neurowire/core'

const feed: NeurowireFeed = {
  id: 'https://example.com/',
  title: 'Example',
  updated: new Date().toISOString(),
  entries: [{ id: '1', title: 'Hello', link: 'https://example.com/hello' }],
}

const atom = serialize(feed, 'atom')
const nwf = serialize(feed, 'nwf')

const result = validateNwf(nwf)
if (!result.valid) console.error(result.errors)

FORMATS, MEDIA_TYPES, and EXTENSIONS enumerate the supported formats and their content types and file extensions.

License

Apache-2.0