DFS API Service

High-performance FastAPI microservice for Daily Fantasy Sports optimization and data endpoints.

Overview

This service replaces Django Lambda for performance-critical endpoints:

• Optimizer: 25-30s → 5-10s (60-80% improvement)
• Zero cold starts: Always-on containers
• 4 vCPUs: Better than Lambda's 1.3 vCPUs

Quick Start

Local Development

# Install dependencies
pip install -r requirements.txt

# Run locally
uvicorn app.main:app --reload --port 8000

# Test optimizer
curl -X POST http://localhost:8000/api/v2/optimize \
  -H "Content-Type: application/json" \
  -d @test_data.json

Deploy to Fly.io

# Install Fly CLI
curl -L https://fly.io/install.sh | sh

# Login
fly auth login

# Deploy
fly launch --copy-config --yes

# Check status
fly status

# View logs
fly logs

API Endpoints

Optimizer

POST /api/v2/optimize

Generate optimized DFS lineups using Linear Programming.

{
  "players": [...],
  "options": {...},
  "game": {...},
  "exposure": {...},
  "num_solutions": 25
}

GET /api/v2/health

Health check endpoint.

Architecture

• FastAPI: High-performance async web framework
• PuLP: Linear Programming optimization
• Fly.io: Global edge deployment
• SQLAlchemy: Database ORM (Phase 2)

Performance

Metric	Before (Lambda)	After (Fly.io)
Optimizer (25 lineups)	25-30s	5-10s
Cold starts	2-5s	<100ms
CPU	1.3 vCPU	4 vCPU

Migration from Django

See FASTAPI_MIGRATION_PLAN.md in the main dailyfantasyserver repo for complete migration instructions.