StreamSprout

StreamSprout is a VALORANT esports command desk that combines roster management, scouting analytics, and an LLM-powered assistant in one Node.js + Express experience. The static frontend (HTML/CSS/JS) lives in public/, while the Express API in server/ powers /players, /search-players, /build-team, and /ask.

Note: The Java/Spring artifacts (POM.xml, TensorFlowLlmApplication.java) are experimental and not wired into the Node build. The primary stack is Node.js + Express + static HTML/CSS/JS.

What StreamSprout Delivers

• Roster management: create and search VALORANT player profiles.
• Analytics snapshot: quick visibility into role and agent coverage.
• LLM assistant: ask scouting or strategy questions via /ask.
• Single-stack setup: Express serves the static UI and APIs from one Node process.

Project Layout

• server/ → Node.js + Express API (/players, /search-players, /build-team, /ask).
• public/ → static HTML/CSS/JavaScript frontend.
• valor_db → PostgreSQL schema for the players table.
• src/, POM.xml, TensorFlowLlmApplication.java → experimental Java/Spring artifacts (not part of the Node build).

Local Development

1) Install dependencies

npm install

2) Configure environment variables

Create a .env file in the repo root.

# Server
PORT=5000
LOG_LEVEL=info

# PostgreSQL (either provide DATABASE_URL or the DB_* fields below)
DATABASE_URL=postgres://user:password@localhost:5432/valor_db
DB_USER=your_username
DB_PASSWORD=your_password
DB_HOST=localhost
DB_NAME=valor_db
DB_PORT=5432

# LLM (Amazon Bedrock proxy or compatible endpoint)
BEDROCK_API_KEY=your_bedrock_api_key
BEDROCK_API_ENDPOINT=https://your-bedrock-endpoint

# Optional: return a deterministic response without calling the LLM
LLM_MOCK_RESPONSE="Sample scouting insight"

3) Set up the database

createdb valor_db
psql -d valor_db -f valor_db

4) Run the app

npm start

Then open http://localhost:5000 to load the UI served from public/.

API Routes

• GET /health → health check.
• GET /players → list all players.
• GET /players/:id → get a player by id.
• POST /players → create a player.
• GET /search-players?name= → search by name.
• POST /build-team → filter by role + agent.
• POST /ask → ask the LLM assistant.

Testing

npm test

Integration tests use an in-memory repository and exercise the Express API routes directly.

Contributing

1. Fork the repo
2. Create a feature branch
3. Commit your changes
4. Push the branch and open a PR

License

MIT. See LICENSE.