@productivities/document-sources

URL → document-source router. Given an arbitrary URL, tells you what kind of document it points at (PDF, EPUB, DOCX, Google Doc, YouTube video, Markdown, LaTeX, reStructuredText, Jupyter notebook, audio, video) and, where useful, rewrites it to a fetchable canonical URL (e.g. arXiv abstract → ar5iv HTML, GitHub blob → raw, Google Doc → text export).

Pure functions, zero runtime dependencies, runs anywhere — browser, MV3 service worker, Node, edge runtimes.

Install

npm install @productivities/document-sources

Usage

import {
  documentSourceFromUrl,
  isPdfUrl,
  isYouTubeUrl,
  parseArxivId,
  arxivHtmlCandidates,
  googleDocTextCandidates,
} from '@productivities/document-sources';

documentSourceFromUrl('https://arxiv.org/pdf/1706.03762');
// → { kind: 'pdf', sourceType: 'pdf', label: 'PDF' }

documentSourceFromUrl('https://docs.google.com/document/d/abc123/edit');
// → { kind: 'document', sourceType: 'google-doc', label: 'Google Doc' }

documentSourceFromUrl('https://www.gutenberg.org/ebooks/1342.epub3.images');
// → { kind: 'document', sourceType: 'epub', label: 'EPUB' }

arxivHtmlCandidates('https://arxiv.org/pdf/1706.03762');
// → [{ url: 'https://ar5iv.labs.arxiv.org/html/1706.03762', label: 'ar5iv HTML' }]

API

• documentSourceFromUrl(url) — primary entry point. Returns { kind, sourceType, label, url? } or null.
• isPdfUrl(url), isYouTubeUrl(url) — predicates.
• parseArxivId(url), parseGoogleDoc(url), parseYouTubeUrl(url) — site-specific extractors.
• Candidate generators (URL → fetchable canonical URLs): arxivHtmlCandidates, googleDocTextCandidates, wordDocxCandidates, epubCandidates, markdownCandidates, textCandidates, notebookCandidates, latexCandidates, rstCandidates, mediaCandidates.

License

MIT