lazymap

lazymap

Golang thread-safe LazyMap implementation with additional features:

1. Concurrency control
2. Capacity control (if defined, it automatically cleans least recently used elements)
3. Expiration (if defined, it automatically deletes expired elements)

The main purpose of LazyMap at webtor.io was to reduce requests between services by introducing intermediate caching-layer.

But what is LazyMap anyway?

Unlike ordinary Map, LazyMap evaluates value on demand if there is no value already exists for a specific key.

Example

package main

import (
	"io/ioutil"
	"log"
	"net/http"
	"time"

	"github.com/webtor-io/lazymap"
)

type ResponseMap struct {
	lazymap.LazyMap
	cl *http.Client
}

func NewResponseMap(cl *http.Client) *ResponseMap {
	return &ResponseMap{
		cl: cl,
		LazyMap: lazymap.New(&lazymap.Config{
			Concurrency: 10,
			Expire:      600 * time.Second,
			ErrorExpire: 30 * time.Second,
			Capacity:    1000,
		}),
	}
}
func (s *ResponseMap) get(u string) (string, error) {
	res, err := s.cl.Get(u)
	if err != nil {
		return "", err
	}
	b := res.Body
	defer b.Close()
	r, err := ioutil.ReadAll(b)
	if err != nil {
		return "", err
	}
	return string(r), nil
}

func (s *ResponseMap) Get(u string) (string, error) {
	v, err := s.LazyMap.Get(u, func() (interface{}, error) {
		return s.get(u)
	})
	if err != nil {
		return "", err
	}
	return v.(string), nil
}

func main() {
	rm := NewResponseMap(http.DefaultClient)
	res, err := rm.Get("https://example.org")
	if err != nil {
		log.Fatal(err)
	}
	log.Println(res)
}