This repository provides a tool to (conceptually) clone the games, categories, levels, variables, players, runs, and leaderboards endpoints. It maintains a running tail in order to keep database up to date. Additionally, it downloads Twitch streams from active speedrunners.
Goals for this project:
- Fast initial sync time and API performance
- Modular components
- High availability support
- (long term) Support multiple storage backends
Potential uses for this project may be:
- Custom clients or tools which use Speedrun.com API, which require additional processing or management from a intermediate endpoint
- Bulk analysis of Speedrun.com data without
This tool was originally designed for my Speedrun Browser for Android app.
With Docker Compose: docker-compose up
In order to record data from Speedrun.com, some extra tools are required:
- Redis v5+ for state tracking and low-level indexing
- A backing database. Currently supported:
Additionally, minimum system requirements:
- 4GB RAM
- 5GB free disk space for Speedrun.com data
- Dual core CPU
It takes about 12-24 hours for the initial sync from Speedrun.com API (as of Jun 2020). This speed is mostly constant due to request limits on the Speedrun.com API.
There are a few easy ways to to start the tool.
If you have Docker Compose, you can bring up everything with one command:
docker-compose up
Access the generated API at http://localhost:3500
A Helm chart which includes a regular MongoDB and Redis is provided in the repository.
There is no official Helm package deployment at this time. To use, clone this repo and run the following command:
helm upgrade --install speedrun-browser ./helm/full
For more information and available configured values, checkout the helm/full/values.yaml file.
For any of the above options, many custom configuration options are supported. A full documented list is avilable in the source code lib/config.ts