v1.0.0-beta.3

scrapex v1.0.0-beta.3

Stability and test coverage release with safer result merging and improved LLM/embedding flow ordering.

Highlights

• LLM enhancements now run before embeddings, so summaries/entities are available for embedding inputs.
• Safer merge behavior in extraction context prevents undefined values from overwriting prior results.
• Expanded end-to-end coverage with local HTTP mocks and real-world fixtures.

Improvements

• LLM enhancement + extraction flow ordered ahead of embeddings for better downstream inputs.
• Merge logic filters undefined values to avoid accidental data loss.
• E2E tests now cover:
  • Scraping with redirects, robots.txt handling, and non-HTML responses
  • RSS/Atom parsing and discovery utilities
  • URL utilities (tracking removal, protocol-relative URLs)
  • Markdown parsing behavior
  • LLM HTTP provider and Ollama embeddings via local mocks (no external deps)
• Added realistic HTML/RSS fixtures to mirror production inputs.

Documentation

• Updated embeddings, utilities, RSS parsing, and extractor docs for accuracy.

Installation

npm install scrapex@beta

Notes

• Requires Node.js 20+.

---

Full Changelog: v1.0.0-beta.2...v1.0.0-beta.3