Scout

Competitive intelligence for Japan's post-rotation Pokemon TCG format. Tracks City League and Champions League tournament results, generates meta tier lists, buy lists, and trend analysis.

Live at scout.trainerlab.io

Stack

• Data Pipeline: Python 3.12+, SQLite (WAL mode)
• Frontend: Next.js 16, Tailwind CSS, Recharts
• Hosting: Vercel (static export)
• Data Sources: LimitlessTCG (City Leagues), pokemon-card.com (Champions League)

Project Structure

trainerlab-scout/
  config.py              # Rotation rules, tier thresholds, placement weights
  db.py                  # SQLite schema and connection helpers
  cli.py                 # CLI entrypoint (scrape, compute, export)
  analysis/
    archetype.py         # Sprite-based archetype detection
    meta.py              # Meta snapshot computation and tier assignment
    buylist.py           # Priority-scored buy list generation
  reports/
    json_export.py       # Static JSON export for Next.js
  scraper/               # Limitless/Labs and pokemon-card.com scrapers
  tests/                 # Python test suite (pytest)
  web/                   # Next.js frontend (deployed to Vercel)
    app/
      [format]/          # Format dashboard and detail pages
      champions/         # Champions League decklists
      lib/
        types.ts         # TypeScript type definitions
        data.ts          # Static JSON data loaders
        utils.ts         # Formatting utilities
      components/        # Shared UI components
    public/
      data/              # Exported JSON fetched from GCS at build time
      images/            # Committed static card/Oak assets; sprites are in /data/images

Data Pipeline

Scrape (Limitless/pokemon-card.com)
  -> SQLite (tournaments, placements, decklist_cards, cl_*)
  -> compute_meta_snapshot (tier assignment)
  -> json_export (static JSON to web/public/data/)
  -> GCS data tarball + web/data-manifest.json
  -> Next.js build (static HTML)
  -> Vercel deploy

Key Commands

# Scrape latest City League results
uv run scout --format ninja-spinner scrape-jp

# Compute meta snapshot
uv run scout --format ninja-spinner meta

# Export all data for the web frontend
uv run scout --format ninja-spinner export-web --strict

# Validate exported data and database integrity
uv run scout --format ninja-spinner validate

Frontend Development

cd web
npm ci
npm run dev       # Dev server on localhost:3000
npm run build     # Production static export
npm test          # Run vitest

Testing

# Python tests
uv run pytest tests/ -v

# Frontend checks
cd web && npx tsc --noEmit && npx eslint . --quiet && npm test

Format Context

Nihil Zero is Japan's post-rotation format covering Temporal Forces (SV5) through Perfect Order, plus Mega Evolution sets (ME01-02.5). Japan plays this format months ahead of the international release, which goes live on April 10, 2026.

Meta Analysis

• Weighted scoring: Tournament placements weighted by finish position (1st = 3.0x through top 16 = 1.2x, others = 1.0x)
• Tier assignment: S (15%+), A (8%+), B (3%+), C (1%+), Rogue (<1%)
• Archetype detection: Sprite filename derivation from Limitless URLs, with content-based fallback for decklist-only sources
• Trend analysis: Early vs late period comparison with per-archetype breakdowns