Redesign AI-ChainGraph around chain-first discovery

[judefluen-coder]

Summary

• Make the AI industry atlas the default entry point, with upstream, core, and downstream discovery before company search.
• Replace consumer-facing scores and review controls with relationship basis, factual explanations, source links, and verification dates.
• Add directed Dagre layout, published-graph filtering, responsive mobile tabs, and a refreshed investor research workspace.
• Extend the canonical model with industry stages, bilingual fields, relationship metadata, and stricter snapshot validation.
• Reframe the project as an open research map with English documentation, MIT/CC BY 4.0 boundaries, contribution rules, and a new README screenshot.

Product decisions

• Readers only see maintainer-published relationships and never need to review or correct data.
• Named supplier/customer relationships require an explicit public source.
• Search remains a shortcut; industry-chain discovery is the primary journey.
• The current repository still uses fictional demo data and does not claim full real-market coverage.

Validation

• npm run smoke
• npm run validate:example
• npm run validate:tabular
• npm run build
• VITE_BASE_PATH=/ai-chaingraph/ npm run build
• npm audit --audit-level=high (0 vulnerabilities)
• Playwright at 1440x900, 1024x768, and 390x844
• Browser console: 0 errors, 0 warnings
• Private-data path and staged-diff secret scans

Follow-up milestones

Real A-share/US data ingestion, weekly automated data PRs, and the complete bilingual UI remain the next roadmap phases.