Package sector-rotation research toolkit v0.1

[gigichengnc]

Summary

This PR turns the 2026 reconstruction into an installable sector-rotation research toolkit while preserving the historical audit and frozen research evidence.

The project was conceived and initiated by the repository author for an HKSI competition in December 2025. That competition context is historical background only; this repository is independent and is not affiliated with or endorsed by HKSI or RRG Research.

What changes

• add root-level Python package metadata for sector-rotation-research;
• expose the generic public CLI sector-rotation;
• add sector-rotation fetch for versioned adjusted-close data + provenance/hash capture;
• add sector-rotation forecast for one-command fresh 1/2/4/8-week RRG-style prospective forecasts;
• expose a small public Python surface while keeping research-only/frozen workflows separate;
• add offline end-to-end workflow tests, including repeated-fetch archive collision coverage and CLI-name regression coverage;
• extend CI to build a wheel, install it into a clean virtual environment, and smoke-test the installed CLI/package;
• restructure the README around project ownership/origin, original question, 2025→2026 reconstruction, usable outputs, and research results;
• add a beginner-facing usage path that explains which repository files matter for normal use and that users do not need to run individual .py files;
• update badges and package metadata to the renamed Sector-Rotation-Forecasting repository;
• clarify third-party naming/affiliation and the current no-open-source-license rights boundary in NOTICE.md;
• remove stale migration-stage bundle/sanitization files.

Research boundaries retained

• the historical 2025 evidence remains a curated audited subset;
• the implementation remains RRG-style / transparent approximation, not a claim of proprietary JdK equivalence;
• the frozen final predictive holdout is a recorded one-time evaluation artifact, not a benchmark to reopen while tuning;
• the prospective CLI forecasts RRG-style coordinates/quadrants, not expected returns, calibrated probabilities, or buy/sell signals;
• the first pre-declared top-3 trading mapping remains a negative result and is not tuned away.

Validation

The latest pull-request GitHub Actions run is green. The workflow currently verifies:

• package installation + 51 passing pytest tests;
• CLI smoke test (sector-rotation version → 0.1.0);
• wheel build for sector-rotation-research from repository root;
• clean virtual-environment wheel installation;
• installed-package CLI/import smoke test.

Merge intent

This PR is Ready for review. If approved for integration, prefer Squash merge so the implementation-step commits on this private preparation branch become one clean main-branch commit.

No PyPI publishing or repository visibility change is included in this PR.