refactor(data): migrate data layout to raw/processed/canonical (path-only) - #13

Merged

xianyu9n merged 7 commits into

refactor/data-layout-migration

Aug 19, 2026

Commits on Aug 19, 2026

refactor(data): migrate data layout to raw/processed/canonical (path-only, no content/logic change)
曾立宏
committed
abdd52c
docs(sft): canonical dataset path in exporter module docstring (data/canonical/<ds>)
曾立宏
committed
4aa3a00
docs(data-layout): update normative docs + stage3c migration note for raw/processed/canonical layout
曾立宏
committed
d973cf2
fix(tests/analysis): real-data test paths to processed/canonical layout; analysis docstring; restore skipped data tests
曾立宏
committed
74081f5
docs(data): add 'how to use data/processed' runbook (canonical->parquet->train/eval) + cross-link from new-dataset guide
曾立宏
committed
4c7a310
chore(security): scrub machine-specific/env references from committed docs (plan B forward-fix)
曾立宏
committed
2286281
fix(data): legacy corpus discovery key after layout migration + review notes
曾立宏
committed
8316dbc