Leetcode Dashboard

A static LeetCode progress dashboard for WUCG. This project has been migrated from a Django / Celery / PostgreSQL deployment to a purely frontend architecture that needs no backend servers to run.

Architecture

Google Sheets (member roster)
        │
        ▼
GitHub Actions  ──►  scripts/collect.py  ──►  data/*.json  (committed)
  (daily cron)                                    │
                                                  ▼
                                         Next.js static export
                                                  │
                                                  ▼
                                          GitHub Pages site

Concern	Where it lives
Scheduler	.github/workflows/collect.yml (daily cron)
Collector	scripts/collect.py + scripts/leetcode_api.py
Persistent storage	data/ directory in this repo (JSON, git-tracked)
Schema validation	scripts/validator.py + scripts/schemas/*.json
Frontend	web/ (Next.js 14 App Router, static export)
Hosting	GitHub Pages (.github/workflows/deploy.yml)

There is no database, no long-running process, and no server to operate. The repository itself is the source of truth — each collector run commits a new snapshot of data/, and that push triggers a Pages redeploy.

Repository layout

.
├── .github/workflows/
│   ├── collect.yml         # daily cron → runs collector → commits data
│   └── deploy.yml          # builds web/ and publishes to GitHub Pages
├── scripts/
│   ├── collect.py          # orchestrator invoked by the cron workflow
│   ├── leetcode_api.py     # LeetCode GraphQL client (US + CN)
│   ├── google_sheet.py     # Google Sheets v4 roster reader
│   ├── validator.py        # JSON Schema validator for data/ artifacts
│   ├── parse_problems.py   # legacy text→CSV helper
│   ├── requirements.txt    # requests + jsonschema
│   └── schemas/
│       ├── user.schema.json
│       ├── submission.schema.json
│       └── summary.schema.json
├── data/
│   ├── summary.json        # top-level roster snapshot
│   └── users/<slug>.json   # per-member records (recent AC, totals, profile)
├── web/                    # Next.js frontend
│   ├── app/                # App Router pages (roster + per-user)
│   ├── lib/                # data loaders + shared types
│   ├── next.config.mjs     # static export + Pages base-path
│   └── package.json
├── AGENTS.md               # conventions for coding agents / contributors
└── README.md               # (this file)

The original Django / Celery source trees (check/, member/, main/, static/, Dockerfile, docker-compose.yml, requirements.txt, manage.py, entrypoint.sh, nginx/) are kept in place for reference. They are not used by the new pipeline and can be removed once the migration has been verified in production.

How a day looks

1. 07:15 UTC — collect.yml cron fires.
2. scripts/collect.py reads the roster from Google Sheets, then calls the LeetCode GraphQL API once per unique leetcode_username.
3. Responses are normalized into the schemas in scripts/schemas/ and written to data/users/<slug>.json + data/summary.json.
4. scripts/validator.py re-validates everything. If any file fails, the job exits non-zero and no commit happens.
5. If data/ has changed, the workflow commits the diff with chore(data): daily collector refresh YYYY-MM-DD and pushes to main.
6. That push triggers deploy.yml, which runs npm run build in web/ and uploads web/out/ to GitHub Pages.

Required repository secrets

Set these under Settings → Secrets and variables → Actions:

Secret	Purpose
GOOGLE_SHEET_ID	Spreadsheet id of the member roster.
GOOGLE_API_KEY	API key with Google Sheets read access.

Enable GitHub Pages under Settings → Pages, selecting GitHub Actions as the source.

Roster sheet schema

The collector reads columns A:Z from the first worksheet, skipping the header row. Columns (0-indexed):

Col	Field	Notes
0	register_date	MM/DD/YYYY HH:MM:SS
1	leetcode_username	required
2	server_region	contains "cn" → CN, else US
3	problems_per_week	integer, blank for free-form
4	start_date	YYYY-MM-DD
5	expire_date	YYYY-MM-DD, blank for no expiry
6	scheduled_problems	whitespace-separated problem codes
7	email	identity key
8	mode	e.g. Normal, Free
9	display_name	optional, falls back to username

Local development

Collector

python -m venv .venv
source .venv/bin/activate        # Windows: .venv\Scripts\activate
pip install -r scripts/requirements.txt

export GOOGLE_SHEET_ID=...
export GOOGLE_API_KEY=...
python scripts/collect.py --dry-run     # fetch + validate, no writes
python scripts/collect.py               # write to data/
python scripts/validator.py             # re-validate everything

Frontend

cd web
npm install
npm run dev                             # http://localhost:3000
npm run build                           # static export to web/out/

web/lib/data.ts reads from ../data at build time. A brand-new clone with an empty data/ will still build — the roster page just renders an empty state until the collector has run once.

License

See LICENSE.