Ship the sample orders in backend/brewgo.db - #7

Summary

• Commit backend/brewgo.db with the sample order history: 300 synthetic orders and 40 customers, from 24 September to 23 October 2026. A fresh clone now has orders in SQLite, so a coding agent can build on real data right away, without the Python setup or npm run seed:orders.
• .gitignore adds one exception, !backend/brewgo.db. All other *.db, *.db-shm and *.db-wal files stay ignored. The database uses the normal journal mode, so it is one file.
• The README no longer says that the database is ignored and not committed. npm run seed:orders stays for databases set by BREWGO_DB.
• The README now requires Node.js 22.13+. spec.md reads SQLite with the built-in node:sqlite, which works without a flag from Node.js 22.13.0 (doc,lib,src,test: unflag sqlite module nodejs/node#55890). Python 3.10+ stays, because the app's API needs it.

The Braid Quickstart uses this: the reader clones BrewGo, creates feat/admin-analytics, and asks the agent to follow spec.md.

Test plan

• Fresh clone of this branch, then git switch -c feat/admin-analytics: backend/brewgo.db is present.
• Read the orders the way spec.md does (node:sqlite, readOnly: true, SELECT payload FROM orders ORDER BY id) with Node.js 24: 300 orders (273 completed, 17 cancelled, 10 refunded), all totals are integer cents, and the SHA-256 hash of brewgo.db does not change.
• git status stays clean after the read.
• npm run test:api: 24 passed. The tests use temporary databases and do not change the committed file.

🤖 Generated with Claude Code