--- name: cli description: Use the longcelot-sheet-db CLI (lsdb). Use when running lsdb init, generate, sync, validate, seed, mock-users, doctor, status, migrate, migrate-data, drop-table, drop-column, or rename-column commands — or when scaffolding a new project, generating schema files, syncing schemas to Google Sheets (including CI-friendly --token-file), seeding with --skip-existing or --upsert, diagnosing configuration issues, exporting schemas to Prisma/SQL, pushing schema changes to all user sheets, or removing/renaming a table or column safely (schema file + live sheet together). license: MIT metadata: package: longcelot-sheet-db version: "0.1.29" --- # longcelot-sheet-db — CLI Reference (`lsdb`) All commands are available as `lsdb ` (global install) or via: ```bash npx lsdb pnpm dlx lsdb yarn dlx lsdb bunx lsdb ``` --- ## init — Scaffold a new project ```bash npx lsdb init npx lsdb init --integrate # merge into existing project without overwriting ``` **What it creates:** - `lsdb.config.ts` — Project configuration (project name, actors with env var mapping) - `.env` — Environment variable template (one `DEV_*_SHEET_ID` per actor) - `schemas/` — Schemas directory with default admin schemas (`users`, `credentials`, `schema_versions`) Run `init` **once** when first adding the package. Use `--integrate` to add it to an existing project without overwriting existing files. --- ## generate — Interactive schema generator ```bash npx lsdb generate bookings ``` Launches an interactive prompt to define columns (name, type, modifiers). Writes a new schema file to `schemas//`. --- ## sync — Sync schemas to Google Sheets ```bash npx lsdb sync # sync all actor sheets npx lsdb sync --all-users # also push to all registered user sheets npx lsdb sync --all-users --dry-run # preview --all-users changes without writing npx lsdb sync --token-file /tmp/t.json # CI/CD: load tokens from file, skip OAuth prompt ``` **What it does:** 1. Loads all schemas from `schemas/` 2. Validates environment variables 3. **Auth**: reads `.lsdb-tokens.json`, refreshes if stale, prompts for browser auth if absent 4. Calls `syncSchema()` for every schema — creates missing tabs and adds missing headers 5. Prints a per-actor status table: ``` Actor │ Sheet ID │ Tables │ Status ───────────┼────────────────────────────┼──────────┼──────────── admin │ 1ABCyourAdminSheetId │ 3 │ ✅ synced student │ 1DEFyourStudentSheetId │ 5 │ ✅ synced teacher │ (not set) │ 4 │ ⚠ skipped ``` **`--all-users`** — reads all rows from admin `users` table and pushes schema changes to every registered user sheet. Uses exponential backoff (1 s → 32 s) to handle Google Sheets API rate limits. **`--token-file `** — CI/CD-friendly. Load a pre-stored tokens JSON instead of the interactive browser prompt: ```bash # GitHub Actions example echo "$LSDB_TOKENS" > /tmp/tokens.json npx lsdb sync --token-file /tmp/tokens.json ``` > `.lsdb-tokens.json` is written to the project root on first interactive auth. Add it to `.gitignore`. Never commit OAuth tokens. --- ## validate — Validate all schemas ```bash npx lsdb validate ``` Checks all schema files in `schemas/` for: - Duplicate table names within the same actor - Invalid column modifiers - Unknown actor references (actor not listed in `lsdb.config.ts`) - Missing required schema fields (`name`, `actor`, `columns`) Use in CI to catch schema problems before `sync`. --- ## seed — Load initial/test data ```bash npx lsdb seed npx lsdb seed seeds/admin.ts --skip-existing # skip on unique conflict (idempotent) npx lsdb seed seeds/admin.ts --upsert # update on unique conflict npx lsdb seed seeds/users.ts --all-actors # distribute data across all user sheets ``` ### Seed file formats **Static (plain object):** ```typescript // seeds/admin.ts export default { users: [ { email: 'admin@example.com', role: 'admin', status: 'active' }, ], } ``` **Dynamic (function — receives `process.env`):** ```typescript // seeds/admin.ts export default async function(env: NodeJS.ProcessEnv) { return { users: [ { email: env.SUPER_ADMIN_EMAIL, role: 'admin', status: 'active' }, ], } } ``` Use the dynamic form when seed data depends on environment variables or CLI arguments. **Flags:** | Flag | Effect | |---|---| | `--skip-existing` | Skip rows where a unique column already matches — no error, no overwrite | | `--upsert` | Update existing row on unique conflict instead of throwing | | `--all-actors` | Distribute seed data across all registered user sheets (reads from admin `users` table) | --- ## mock-users — Create test user sheets ```bash npx lsdb mock-users npx lsdb mock-users 5 # create 5 mock users ``` Generates mock Google Sheets for development/testing. Lets you inspect what real users see without registering real accounts. --- ## doctor — Diagnostics and health checks ```bash npx lsdb doctor ``` Checks: - All required environment variables are set - Google OAuth credentials are valid - `lsdb.config.ts` structure is correct - `.lsdb-tokens.json` exists and is readable Run `doctor` first when debugging mysterious adapter errors. --- ## status — Show project status ```bash npx lsdb status ``` Displays registered tables, actors, sheet IDs, schema counts, and token info. --- ## migrate — Export schemas to SQL/Prisma ```bash npx lsdb migrate --prisma --output ./prisma # generate schema.prisma npx lsdb migrate --sql --output ./migrations # generate CREATE TABLE DDL ``` Exports all table schemas to a target format. Use as a starting point when migrating from Google Sheets to a production database. For the full migration guide see `skills/migrations/SKILL.md`. > Deprecated alias: `lsdb export` still works but emits a deprecation warning. --- ## migrate-data — Generate data migration script ```bash npx lsdb migrate-data npx lsdb migrate-data --table users --output ./scripts npx lsdb migrate-data --all-users # also export every registered user sheet npx lsdb migrate-data --dry-run # preview plan without writing files ``` Generates a `migrate-data.js` script with `insertRow()` stubs. Replace the stubs with your production DB client to transfer data row-by-row. > Deprecated alias: `lsdb export-data` still works but emits a deprecation warning. --- ## drop-table — Delete a table's schema file + Google Sheet tab ```bash npx lsdb drop-table bookings npx lsdb drop-table bookings old_notifications --all-users npx lsdb drop-table # interactive checkbox over every table npx lsdb drop-table --dry-run ``` `sync` never deletes anything on its own (see `skills/migrations/SKILL.md`) — this is the explicit way to remove a table from both the schema file and the live sheet. Deletes the schema file, then the corresponding Google Sheet tab (and, with `--all-users`, the tab in every registered user's personal sheet). Prints a plan and asks for confirmation first (`--yes` to skip). Warns if another table's `ref()` points at the one being dropped. --- ## drop-column — Delete column(s) from a table ```bash npx lsdb drop-column bookings notes npx lsdb drop-column bookings # interactive checkbox over that table's columns npx lsdb drop-column # interactive: pick table, then columns npx lsdb drop-column bookings notes --all-users --yes ``` Removes column(s) from the schema file and deletes the matching column(s) — and all data in them — from the live Google Sheet. Resolves each column's *current* position from the sheet's header row, not the schema file's declared order (`sync` appends new columns at the end, so they can differ). Refuses to drop reserved columns (`_id`, `_created_at`, `_updated_at`, `_deleted_at`) or a table's primary key column — drop the whole table instead if that's really the intent. --- ## rename-column — Rename a column without losing data ```bash npx lsdb rename-column bookings notes remarks npx lsdb rename-column # fully interactive: table, column, new name npx lsdb rename-column bookings notes remarks --all-users ``` Renames the column's key in the schema file and overwrites the Google Sheet **header cell in place** — existing row data is preserved, unlike a drop-and-re-add. This is what prevents the real risk it closes: renaming a column in the schema file alone desyncs every existing row from that point on, and a later `sync`/`auto-sync` has no way to recover data under the old header name. Warns (doesn't block) if another table's `ref()` points at the old name — those `ref()` strings in other schema files aren't rewritten automatically. --- ## lsdb.config.ts structure ```typescript import type { SheetDBConfig } from 'longcelot-sheet-db'; export default { projectName: 'my-app', superAdminEmail: 'admin@example.com', actors: [ { name: 'admin', sheetIdEnv: 'ADMIN_SHEET_ID' }, { name: 'student', sheetIdEnv: 'DEV_STUDENT_SHEET_ID' }, { name: 'teacher', sheetIdEnv: 'DEV_TEACHER_SHEET_ID' }, ], onSchemaMismatch: 'warn', // 'warn' | 'error' | 'auto-sync' } satisfies SheetDBConfig; ``` The `sheetIdEnv` field tells each CLI command which env var holds the sheet ID for that actor. --- ## Common Mistakes - **Running `sync` without `.env` configured** — `sync` fails at OAuth if env vars are missing. Run `doctor` first. - **Committing `.lsdb-tokens.json`** — Contains OAuth refresh tokens. Verify it is in `.gitignore` before any push. - **Not running `sync` after schema changes** — Schema files are the source of truth. New columns don't appear in Sheets until `sync` runs. - **Re-seeding without `--skip-existing`** — Running `seed` twice throws `ValidationError: Unique constraint violation` for any unique column. Use `--skip-existing` for idempotent seeds or `--upsert` to update. - **Forgetting `--token-file` in CI** — Without it, `sync` blocks waiting for interactive input and the CI job hangs. - **Using a dynamic seed file without `export default async function`** — Named exports or non-function defaults are treated as the static plain-object format and must return `Record` directly. - **Expecting `drop-column`/`drop-table` to be additive-safe like `sync`** — They're destructive by design: real Google Sheets columns/tabs (and all data in them) are deleted. Always run with `--dry-run` first if unsure, especially with `--all-users`. - **Renaming a column in the schema file by hand instead of `rename-column`** — A manual rename leaves the live sheet's header on the old name, silently desyncing every existing row. Use `lsdb rename-column` so the schema file and sheet header change together. - **Expecting `rename-column`/`drop-column` to update `ref()` strings in other schema files** — They warn if another table's `ref('table.column')` points at what's being renamed/dropped, but don't rewrite it for you. Update those `ref()` calls manually.