Starcraft Data Platform

Status: Work in Progress

This repository is under active development. It defines the database architecture for ingesting, storing, and analyzing Starcraft II replay data.

Overview

The Starcraft Data Platform is designed to:

• Store raw replay data, player information, and event logs in a relational database.
• Provide analytics views and materialized tables for advanced insights.
• Enable machine learning workflows by supporting feature engineering and model outputs.

Planned Features

1. Core Functionality:

   • Relational database schema for storing raw replay data.
   • Analytics views for player statistics and performance trends.
   • ML schema for storing model outputs and feature sets.

2. Technology Stack:

   • PostgreSQL for database storage.
   • SQLAlchemy for ORM models.
   • Docker for containerized deployment.

3. Future Enhancements:

   • Performance optimization for materialized views.
   • Expanded analytics for build orders and map-specific trends.

Current Focus

The current phase of development is focused on:

• Designing the raw schema for replay ingestion.
• Creating analytics views for player statistics.

Getting Started

⚠️ This project is not yet ready for deployment or public use. Setup instructions will be added as development progresses.

Planned Structure

This repository will be organized into the following structure:

• db/: Contains all database-related files.

  • schema.sql: A consolidated SQL file for creating all database schemas, including raw, analytics, and ML schemas.
  • models/: A directory housing SQLAlchemy ORM models.
    • replay.py: ORM models for replay metadata tables (e.g., Info, Player, Team).
    • events.py: ORM models for event data (e.g., UnitBornEvent, SelectionEvent).
    • datapack.py: ORM models for unit and ability metadata (e.g., UnitType, Ability).
  • analytics/: SQL files for analytics views and materialized tables.
    • player_stats.sql: A view aggregating player performance metrics.
    • map_performance.sql: A view tracking win rates by map and race.
    • player_stats_summary.sql: A materialized view summarizing player statistics across replays.

• tests/: Contains unit tests for validating schema and models.

  • test_schema.py: Tests to ensure the database schema is correctly defined.
  • test_models.py: Tests for ORM models to verify accurate data interaction.
  • test_analytics.py: Tests for analytics queries and performance.

• Dockerfile: Configuration for containerizing the project, ensuring consistent environments for development and deployment.

• README.md: This documentation file, providing an overview of the project and its structure.

How to Contribute

Contributions are welcome once the repository reaches a stable state. If you'd like to collaborate, please reach out or open an issue to discuss ideas.

License

[MIT License]