gotq

golang distribute task queue, base on redis, lightweight and easy use, inspired by python's celery

Feature

use redis as broker
customize worker numbers
support async task or block for wait reply
worker will discard outdate sync task

Installation

go get github.com/shidenggui/gotq

Overview

you can see example in gotq/example

start worker

go run example/worker/worker.go

dispatch task

block for result

go run example/server/server.go

dispatch async task

go run example/server/main.go -m delay

QuickStart

define task

//config broker
var cfg = config.Config{
	Broker: &config.BrokerCfg{
		Host:     "127.0.0.1",
		Port:     6379,
		Password: "",
		DB:       0,
	},
}

var App = gotq.New(&cfg)

// taskSender, one task function one sender
var AddSender = App.Register(Add)

type AddArgs struct {
	X int64
	Y int64
}

type AddResult struct {
	Sum int64
}

func Add(argsInter map[string]interface{}) map[string]interface{} {
	args := new(AddArgs)

	//recommended use mapstructure to bind map to struct
	ms.Decode(argsInter, args)
    
    // do soemthing you want

    //recommeded use github.com/fatih/strtuts to convert map to struct
    res := &AddResult{
        Sum: args.X + args.Y,
    }

    /* also can simple use map[string]interface{}
	res := make(map[string]interface{})
	res["sum"] = args.X + args.Y
    /*

	return structs.Map(res)
}

start workers

workerNum := int64(500)
App.WorkerStart(workerNum)

dispatch task

Async task dispatch

args := &AddArgs{3, 4}
err := AddSender.Delay(args)

Sync task dispatch, block for wait result

args := &AddArgs{3, 4}
timeout := int64(5)
res, err := AddSender.Request(args, timeout)
...
//recommended use mapstructure to bind map to struct
rep := new(AddReply)
ms.Decode(res, rep)