All Star Dashboard

Internal dashboard for All Star Turlock. Three views:

1. Reorder Calls — last year's Printavo orders whose anniversary is 21–45 days out, filtered to orders that were $200+, from recurring customers, or tagged as events in the job name.
2. Weekly Leads — a fresh list of prospects to call each Monday, pulled from Google Places (schools, gyms, leagues) plus local school-district directories within 25 miles of Turlock, CA.
3. Competition — track competitor pricing, products, promos, and social posts. Configurable scrape targets per competitor; daily cron refresh.

Tech stack

• Next.js 16 (App Router, Turbopack) + React 19
• TypeScript + Tailwind v4
• Neon Postgres (serverless HTTP driver)
• Deploys to Vercel; Vercel Cron drives the sync jobs.

One-time setup

1. Install deps

   npm install

2. Provision a database. Create a Neon project and copy the pooled connection URL. Save it as DATABASE_URL in .env.local (see .env.example).

3. Apply the schema

   npm run db:push

4. Fill in the rest of .env.local:

   • DASHBOARD_PASSWORD — the shared password you'll type at the login screen.
   • SESSION_SECRET — any long random string (session cookie HMAC).
   • CRON_SECRET — Vercel will send this as a Bearer token to /api/cron/*.
   • PRINTAVO_EMAIL / PRINTAVO_TOKEN — from Printavo → My Account → API.
   • GOOGLE_PLACES_API_KEY — Google Cloud Console, enable “Places API (New)”.

5. Run locally

   npm run dev

   Visit http://localhost:3000 and sign in with DASHBOARD_PASSWORD.

6. First-time data load. After signing in, click:

   • Sync Printavo now on the Reorders page — pulls ~13 months of invoices.
   • Sync leads now on the Leads page — populates Google Places + schools.
   • Add a competitor on the Competition page, configure scrape targets, and click Scrape now.

Deploy to Vercel

1. Push this repo / subtree to GitHub (see “Extract to its own repo” below).
2. Import in Vercel. Set all env vars from .env.example under Project → Settings → Environment Variables.
3. Vercel auto-detects vercel.json and registers three cron jobs:
   • /api/cron/printavo-sync — daily 08:00 UTC
   • /api/cron/leads-sync — Mondays 09:00 UTC
   • /api/cron/competitors-sync — daily 06:00 UTC
4. Once deployed, run the schema push against the production DB by running npm run db:push locally with DATABASE_URL pointed at prod.

Swapping in Microsoft SSO later

The proxy at src/proxy.ts only checks that getSession() returns a valid session. To switch to Microsoft:

1. Install a provider like next-auth with the Azure AD provider.
2. Replace src/app/login/page.tsx and src/app/api/auth/* with next-auth's routes.
3. Keep getSession() shape-compatible ({ sub, exp }) — the rest of the app doesn't need to change.

Extract to its own repo

This app currently lives in the dashboard/ subdirectory of caghassi/all-star-inventory-app. To publish it as its own repo:

# from the monorepo root
git subtree split --prefix=dashboard -b allstar-dashboard-split
cd /tmp && git clone /path/to/all-star-inventory-app allstar-dashboard
cd allstar-dashboard
git checkout allstar-dashboard-split
git remote remove origin
git remote add origin git@github.com:caghassi/allstar-dashboard.git
git push -u origin main

Project layout

dashboard/
  src/
    proxy.ts                  # Next.js 16 middleware (session gate)
    app/
      api/
        auth/{login,logout}/  # Password login + logout
        cron/                 # Sync jobs - authed by CRON_SECRET
          printavo-sync/
          leads-sync/
          competitors-sync/
      login/                  # /login page
      reorders/               # Reorder call list
      leads/                  # Weekly call list
      competitors/            # Competition tracking
    components/Shell.tsx
    lib/
      auth.ts                 # Session cookie + password check
      db.ts                   # Neon SQL tagged template
      config.ts               # Geo, keywords, thresholds
      printavo.ts             # Printavo GraphQL client
      reorder-sync.ts         # Pulls invoices + rebuilds reorder queue
      places.ts               # Google Places (New) client
      schools.ts              # Local school-district scraper
      leads-sync.ts           # Upserts leads + rotates weekly queue
      competitor-sync.ts      # Per-competitor scrape runner
      cron-auth.ts            # Bearer-token / session check
    db/schema.sql             # Postgres DDL
  scripts/db-push.mjs         # Applies schema.sql to $DATABASE_URL
  vercel.json                 # Cron schedule

How the reorder rule works

A Printavo invoice qualifies for the reorder queue if ALL of these are true:

• Its due_date + 365 days falls within [today+21, today+45] days.
• AT LEAST ONE of:
  • order_total_cents >= 20000 ($200)
  • The customer has 2+ historical orders (recurring)
  • The job name contains an event keyword: tournament, tourney, camp, league, season, classic, invitational, championship, showcase, meet, relay

Tune the constants in src/lib/config.ts.

Notes on the Printavo API

• Authentication uses email + token HTTP headers.
• The invoices connection supports cursor pagination via first / after.
• We pull 13 months on every sync to stay fully caught up without relying on webhooks. A single sync costs roughly N/50 GraphQL calls where N is your yearly invoice count.

Notes on Google Places

• Uses the new Places API (places.googleapis.com/v1), not the legacy one.
• Field masks keep billing minimal — we only request fields we store.
• Each Text Search call is billed per response page. The weekly sync makes roughly PLACES_TYPES.length + PLACES_KEYWORDS.length searches, each up to 3 pages of 20 results.