Skip to content

Latest commit

 

History

History
70 lines (51 loc) · 2.89 KB

README.md

File metadata and controls

70 lines (51 loc) · 2.89 KB
Raw

game-node-sync-hltb

The GameNode sync system for HLTB.

Usage

This service is dependent on game-node-server, since the RabbitMQ instance is configured there. Have it set-up before proceeding.

Installation

Pre-requisites

  • At least go 1.23.1 in your machine.
  • A redis instance running in port 9112 You can also change the redis addr (host:port) used by setting the REDIS_ADDR env variable.
  • A rabbitmq instance running in port localhost:5672
    You can also change the RabbitMQ URL (default: amqp://localhost:5672) by setting the REDIS_URL env variable.
    If possible, have game-node-server set up before this.

Starting

After installing the project dependencies (go mod download), simply use these commands to start each service:

# Starts the 'listener' process
go run cmd/listener/main.go

# Starts the 'processor' process
go run cmd/processor/main.go

Using the Docker image

You can also use this service by pulling and running the lamarcke/game-node-sync-hltb image.

# Starts the 'listener' process with the 'hltb-listener' container name
docker run --name hltb-listener lamarcke/game-node-sync-hltb /app/listener

# Starts the 'processor' process with the 'hltb-processor' container name
docker run --name hltb-processor lamarcke/game-node-sync-hltb /app/processor

See the docker-compose.prod.yml for a compose file example.

Implementation

This Go app has two main entrypoints:

  • listener
  • processor

Both are available at the cmd directory.

listener is responsible for receiving requests for playtime info updates, and creates new tasks based on said requests. processor is a asynq worker pool. It handles the actual processing of requests, and publishes the results accordingly.

RabbitMQ is used as the message broker for this application, and it's instance is available as a dependency in our main api.

The actual processing is quite simple:
update request received -> handled as task -> parse game name to improve results -> search for best possible match -> publishes result

The process then continues on our main api, where data is made available to users.

Disclaimer

GameNode is not affiliated with HowLongToBeat or it's partner websites/services in any way.

We're doing our best to keep the requests WAY below a possible block threshold by having a single worker processing requests and artificial timeouts in place. Requests to HLTB are only actually done every 4-8 seconds, and only when strictly necessary.

Data received from HLTB is also credited everytime it's shown in our client apps.

If you are a HLTB affiliate and wish to contact us, please feel free to email us here: [email protected]