All notable changes to Retail Scrapers will be documented in this file.
This project follows semantic versioning once published.
- Initial toolkit structure.
- Added a shared Amazon engine for
amazon-de,amazon-gb,amazon-it, andamazon-es. - Added adapters for
boulanger-fr,currys-gb, andelkjop-no. - Added Amazon market-specific delivery sessions, native-currency guards, and URL checks.
- Added JSONL and CSV output.
- Added strict validation for catalog and price runs.
- Added CLI and Python API entry points.
- Added optional batch-first price snapshots with safe detail-page fallback.
- Added Boulanger current-price filtering and Currys 50-item category snapshots.
- Added
auto,batch, anddetailprice strategies with count and coverage gates. - Added explicit channel profiles and
retail-scrape assessfor deterministic URL classification. - Added optional one-page public signal probing with local-network and credential URL guards.
- Added strategy-aware
blank, API, HTML, JSON-LD, and browser adapter scaffolds. - Upgraded the bundled skill for agent-assisted assessment, onboarding, repair, and validation.
- Unified the README, project tour, contributor forms, and scaffold guidance around two quick paths: run a supported channel or assess and onboard a new retailer.
- Merged the English and Chinese README into one paragraph-by-paragraph bilingual quick start.
- Added
retail-scrape setup, an offlinedemo,--version, human-readable environment reports, complete command help, quiet mode, completion summaries, and actionable error categories. - Added a bundled Skill installer and
assess --format agentfor copy-ready agent onboarding. - Added public package links, wheel-level Skill packaging, placeholder-input rejection, and stricter release validation for the first PyPI release.