# Session 08 — Materials

## `liangsui.7z` — the corpus

A 7-zip archive of **29 sources on Tang–Song fiscal history**, centered on the Two-Tax Law (兩稅法 *liǎng shuì fǎ*, 780 CE). The corpus mixes:

- Modern Chinese-language scholarly articles (PDFs) — including 陈明光, 刘玉峰, 翁俊雄, and other historians.
- A few cleaned Markdown conversions of the same articles for easier ingestion.

Extract with:

```bash
# macOS
brew install p7zip          # one-time setup
7z x liangsui.7z -o<destination>

# Windows
# Install 7-Zip (https://www.7-zip.org), then right-click → 7-Zip → Extract here
```

After extraction, drop the resulting folder of files into your wiki's `raw/` layer.

## The two prompts

The session walks you through **two Codex prompts** in sequence. Both are reproduced in `session_08.html` / `session_08.qmd`. Briefly:

1. **Scaffold prompt** — builds `AGENTS.md`, the `derived/text/` extraction pipeline, the wiki folder structure (`sources/, concepts/, people/, institutions/, debates/`), Obsidian Bases views, page templates, and 29 source stubs. Does *not* ingest content.

2. **Ingest prompt** — has Codex read each extracted source directly (no script — the work is interpretive) and fill the 29 source stubs with page-level notes plus expand the concept graph.

Read the qmd; copy the prompts; paste into Codex; approve the file writes.

## Optional follow-on

After the two main prompts there is a third optional one (in the qmd) that renames source files from `src-001.md` to article-title filenames and rewrites every Obsidian link. Run it if you want readable filenames in the sidebar; skip it if `src-NNN` is fine.

## Sources / acknowledgements

The corpus is for workshop teaching only. Each PDF retains its original copyright; redistribution outside the workshop is not authorized.