Skip to content

ymow/PitLane

Repository files navigation

PitLane: Omni-Graph Intelligence Engine

PitLane is a Universal Domain Intelligence Engine — an LLM-driven, graph-based CMS that reconstructs complex industries in real time. Formula 1 is the first pilot scenario. The same architecture can be switched to MotoGP, cinema, or finance by swapping a YAML config file.

"From F1 Paddock to Hollywood Red Carpet: One Engine, Infinite Intelligence."


Why PitLane is Different

Traditional news platforms report what happened. PitLane answers why — by fusing real-time telemetry, paddock social signals, and 70+ years of historical context into a single LLM-driven War Room.

Capability Traditional CMS PitLane
News aggregation
Multi-language ✅ 10+ languages
Live race telemetry ✅ Sub-second sync
Causal inference ("why Honda left") ✅ RAG + LLM
Paddock social monitoring (IG Stories) ✅ 2000+ accounts
Entity relationship graph ✅ Driver → WAG → Legend
Cross-domain template ✅ F1 → Movies → Finance

Four Core Pillars

🏎️ Real-time Race Sync — When lights go out, every car's telemetry is mirrored in the database. WebSocket push to frontend. Anomaly detection triggers automatic incident reports.

🌍 Global Intelligence — Claude AI translates with F1-specific terminology across 10+ languages. 2000+ paddock accounts monitored across X, Instagram, and TikTok.

🧠 LLM Contextual Engine — Historical RAG over regulations, Wikipedia, and F1 books. Causal inference links breaking news to decades of context. Auto-generates weekly deep-dive articles.

🛠️ Universal CMS Template — Domain-agnostic Entity + Relationship graph. Pluggable skill system. Switch domains by loading a new YAML config.


Technology Stack

Layer Technology
Backend Django 5.x · Celery · Redis · PostgreSQL
Frontend React 19 · Vike SSR · Tailwind CSS v4 · Base UI
AI Anthropic Claude API (translation, inference)
NER spaCy · SimHash
Live Data OpenF1 (real-time) · Jolpica · FastF1 (analysis)
Vector DB pgvector (Phase 3)
Dev Workflow Spec-driven · Linear.app

Roadmap

Phase Focus Status
Phase 1 F1 2026 Pilot — news pipeline, live telemetry, social registry, 2026 grid 🟡 In Progress
Phase 2 Graph & Abstraction — universal Entity model, paddock people graph, causal graph ⚪ Planned
Phase 3 Intelligence — social ingestion, historical RAG, live race state machine ⚪ Planned
Phase 4 Template & Expand — YAML domain config, skill interface, second domain pilot ⚪ Planned

See SPEC.md for the full issue backlog and technical architecture.


Quick Start

1. Backend

cd backend
pip install -r requirements.txt
python manage.py migrate
python manage.py runserver

2. Frontend

cd frontend
npm install
npm run dev

Database Seeding (2026 Season)

Run these commands in order to fully populate a fresh database with the 2026 F1 season data.

Step 1 — Base data (teams, drivers, news categories, RSS sources)

cd backend
python manage.py seed_data

Creates 10 teams, 20 drivers, 8 news categories, and all RSS feed sources.

Step 2 — 2026 grid sync (Audi rebrand + RACE contracts)

python manage.py seed_2026_grid
  • Renames the Kick Sauber entry → Audi F1 Team (code AUD, color #BB0000)
  • Creates 20 DriverContract(role=RACE) rows linking each driver to their 2026 team

The 2026 driver lineup:

Team Drivers
Red Bull Racing Max Verstappen, Liam Lawson
Scuderia Ferrari Charles Leclerc, Lewis Hamilton
McLaren F1 Team Lando Norris, Oscar Piastri
Mercedes-AMG Petronas George Russell, Andrea Kimi Antonelli
Aston Martin Aramco Fernando Alonso, Lance Stroll
Alpine F1 Team Pierre Gasly, Jack Doohan
Williams Racing Alexander Albon, Carlos Sainz
Racing Bulls Yuki Tsunoda, Isack Hadjar
MoneyGram Haas F1 Team Oliver Bearman, Esteban Ocon
Audi F1 Team Nico Hülkenberg, Gabriel Bortoleto

This command is idempotent — safe to run multiple times.

Step 3 — Paddock staff contracts (Team Principals, engineers, etc.)

python manage.py seed_staff_contracts \
  --file data/staff_contracts_2026.json \
  --season 2026

Creates 37 staff contract entries: 10 Team Principals, 10 Technical Directors, 14 Race Engineers, and other paddock roles.

Note: Must run Step 2 first. The staff data references team code AUD, which is created by the Audi rebrand in Step 2.

Step 4 — 2026 race calendar (24 races)

python manage.py seed_2026_schedule

Populates the full 2026 race calendar with circuit data.

Step 5 — Paddock social handles (optional)

python manage.py seed_social_handles \
  --file data/paddock_handles_2026.json

Seeds 2000+ social media handles for drivers, team principals, and paddock staff across X/Twitter, Instagram, and TikTok.

Full reset (all steps at once)

cd backend
python manage.py seed_data && \
python manage.py seed_2026_grid && \
python manage.py seed_staff_contracts --file data/staff_contracts_2026.json --season 2026 && \
python manage.py seed_2026_schedule && \
python manage.py seed_social_handles --file data/paddock_handles_2026.json

API Overview

Base URL: http://localhost:8000/api/v1/

Endpoint Description
GET /teams/ All 10 teams with color and logo
GET /drivers/ All 20 drivers
GET /contracts/?role=RACE&is_active=true 2026 race grid with team colors
GET /contracts/?role=TEAM_PRINCIPAL&is_active=true All team principals
GET /social-handles/ Paddock social registry
GET /articles/ News articles (supports ?lang=zh-TW)
GET /f1/standings/ Championship standings
GET /f1/schedule/ Race calendar

Write endpoints (POST, PATCH, DELETE) require admin authentication.


Built for the 2026 F1 season. Designed for every domain beyond it.

About

Global F1 News Platform

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors