asyncgo

Asyncgo is zero-dependency asynchronous task executor written in pure go, that prioritises speed and ease of use.

Features

• Asynchronous Task Execution: Submit tasks to execute asynchronously and retrieve results.
• No Manual Goroutine Management: Abstracts away the complexity of managing goroutines, and simplifying the code.
• Worker Pool Management: Asyncgo carefully handles worker pool creation & task execution.
• Graceful Shutdown: Guarantees all existing tasks are completed before shutting down the workers
• Task Cancellation: Support for terminating workers.

Usecases

• Asynchronous HTTP Requests for Microservices
• Background Job Execution
• Infinite concurrent polling with worker pool (receiving messages from AWS SQS or similar services)

Documentation

1. Installation

   go get github.com/abhi16180/asyncgo
   

2. Importing

   import "github.com/abhi16180/asyncgo"

Examples

1. Executing multiple functions asynchronously

    package main

    import (
        "github.com/abhi16180/asyncgo"
        "log"
        "time"
    )

    func main() {
        executor := asyncgo.NewExecutor()
        future1 := executor.Submit(func(arg int) (int64, error) {
            time.Sleep(1 * time.Second)
            return int64(arg * arg), nil
        }, 10)
        // first param is function, all remaining params are arguments that needs to be passed for your function
        // if function signature / args do not match, it will result in execution error
        future2 := executor.Submit(func(arg1 int, arg2 int) (int, error) {
            time.Sleep(1 * time.Second)
            return arg1 + arg2, nil
        }, 10, 20)
        // err is execution error, this does not represent error returned by your function
        result1, err := future1.Get()
        if err != nil {
            log.Println(err)
            return
        }
        result2, err := future2.Get()
        if err != nil {
            log.Println(err)
            return
        }
        // result is []interface that contains all the return values including error that is returned by your function
        log.Println(result1, result2)
    }

NOTE:

executor.Submit(function,args..) always swpans new goroutine every time. For large number of tasks, its recommended to use worker pool

2. Executing large number of tasks with fixed sized worker pool

package main

import (
	"context"
	"github.com/abhi16180/asyncgo"
	"github.com/abhi16180/asyncgo/commons"
	"log"
	"time"
)

func main() {
	executor := asyncgo.NewExecutor()
	workerPool := executor.NewFixedWorkerPool(context.Background(), &commons.Options{
		WorkerCount: 100,
		BufferSize:  100,
	})
        // gracefully terminate all workers
	// guarantees every task is executed
	defer workerPool.Shutdown()
	futures := []*asyncgo.Future{}
	for i := 0; i < 1000; i++ {
		future, err := workerPool.Submit(timeConsumingTask)
		if err != nil {
			log.Println("error while submitting task to worker pool")
			continue
		}
		futures = append(futures, future)
	}

	for _, future := range futures {
		result, err := future.Get()
		if err != nil {
			log.Println("error while getting result from future")
			continue
		}
		log.Println(result)
	}
}

func timeConsumingTask() string {
	time.Sleep(2 * time.Second)
	return "success"
}

3. Cancelling worker pool in the middle of execution

package main

import (
	"context"
	"github.com/abhi16180/asyncgo"
	"github.com/abhi16180/asyncgo/commons"
	"log"
	"time"
)

func main() {
	executor := asyncgo.NewExecutor()
	workerPool := executor.NewFixedWorkerPool(context.Background(), &commons.Options{
		WorkerCount: 100,
		BufferSize:  100,
	})

	futures := []*asyncgo.Future{}
	for i := 0; i < 1000; i++ {
		future, err := workerPool.Submit(timeConsumingTask)
		if err != nil {
			log.Println("error while submitting task to worker pool")
			continue
		}
		futures = append(futures, future)
	}
	// terminate worker pool in the middle of task(s) execution
	workerPool.Terminate()
}

func timeConsumingTask() string {
	time.Sleep(2 * time.Second)
	return "success"
}

4. For more use-cases and complex examples check out examples section