notebooks

Imported from xdanny/multiturn-sql-finetuning (notebooks/AGENTS.md). Install upstream with npx skills add xdanny/multiturn-sql-finetuning --skill notebooks. Copyright stays with the author.

Notebooks AGENTS

This subtree owns shareable labs and notebook support code. Notebooks are companion material for a specific public explanation, not the source of truth for benchmark claims.

Primary responsibilities:

• Keep labs small enough for readers to run on CPU by default, with accelerator use treated as optional.
• Make notebook examples explain method shape: direct SQL, structured briefs, semantic-layer state, MEASURE() preservation, and behavior recovery.
• Keep reusable logic in Python support modules instead of large notebook cells.
• Make published HTML labs discoverable without turning every old blog post into a notebook project.

Rules:

• Use notebook code to illustrate the repo contracts; do not let notebook-only results become benchmark evidence.
• Do not expose reference SQL, gold plans, gold metric DSL, expected rows, repair labels, or future turns as production-style model inputs.
• Prefer static, deterministic examples over live endpoint calls.
• If a lab cites a result, point to the manifest, artifact, or generated evidence file that supports it.
• Keep notebook exports synchronized only when the source notebook intentionally changes. Avoid regenerating unrelated notebooks in the same PR.

When editing here, inspect:

• docs/blog/README.md
• docs/research_roadmap.md