Simple lib to parse envs to structs in Go.
A very basic example:
package main
import (
"fmt"
"time"
// if using go modules
"github.com/caarlos0/env/v6"
// if using dep/others
"github.com/caarlos0/env"
)
type config struct {
Home string `env:"HOME"`
Port int `env:"PORT" envDefault:"3000"`
IsProduction bool `env:"PRODUCTION"`
Hosts []string `env:"HOSTS" envSeparator:":"`
Duration time.Duration `env:"DURATION"`
TempFolder string `env:"TEMP_FOLDER" envDefault:"${HOME}/tmp" envExpand:"true"`
}
func main() {
cfg := config{}
if err := env.Parse(&cfg); err != nil {
fmt.Printf("%+v\n", err)
}
fmt.Printf("%+v\n", cfg)
}
You can run it like this:
$ PRODUCTION=true HOSTS="host1:host2:host3" DURATION=1s go run main.go
{Home:/your/home Port:3000 IsProduction:true Hosts:[host1 host2 host3] Duration:1s}
Out of the box all built-in types are supported, plus a few others that are commonly used.
Complete list:
string
bool
int
int8
int16
int32
int64
uint
uint8
uint16
uint32
uint64
float32
float64
string
time.Duration
encoding.TextUnmarshaler
url.URL
Pointers, slices and slices of pointers of those types are also supported.
You can also use/define a custom parser func for any other type you want.
If you set the envDefault
tag for something, this value will be used in the
case of absence of it in the environment.
By default, slice types will split the environment value on ,
; you can change
this behavior by setting the envSeparator
tag.
If you set the envExpand
tag, environment variables (either in ${var}
or
$var
format) in the string will be replaced according with the actual value
of the variable.
Unexported fields are ignored.
If you have a type that is not supported out of the box by the lib, you are able
to use (or define) and pass custom parsers (and their associated reflect.Type
)
to the env.ParseWithFuncs()
function.
In addition to accepting a struct pointer (same as Parse()
), this function
also accepts a map[reflect.Type]env.ParserFunc
.
env
also ships with some pre-built custom parser funcs for common types. You
can check them out here.
If you add a custom parser for, say Foo
, it will also be used to parse
*Foo
and []Foo
types.
This directory contains pre-built, custom parsers that can be used with env.ParseWithFuncs
to facilitate the parsing of envs that are not basic types.
Check the example in the go doc for more info.
The env
tag option required
(e.g., env:"tagKey,required"
) can be added
to ensure that some environment variable is set. In the example above,
an error is returned if the config
struct is changed to:
type config struct {
Home string `env:"HOME"`
Port int `env:"PORT" envDefault:"3000"`
IsProduction bool `env:"PRODUCTION"`
Hosts []string `env:"HOSTS" envSeparator:":"`
SecretKey string `env:"SECRET_KEY,required"`
}
The env
tag option file
(e.g., env:"tagKey,file"
) can be added
to in order to indicate that the value of the variable shall be loaded from a file. The path of that file is given
by the environment variable associated with it
Example below
package main
import (
"fmt"
"time"
"github.com/caarlos0/env"
)
type config struct {
Secret string `env:"SECRET,file"`
Password string `env:"PASSWORD,file" envDefault:"/tmp/password"`
Certificate string `env:"CERTIFICATE,file" envDefault:"${CERTIFICATE_FILE}" envExpand:"true"`
}
func main() {
cfg := config{}
if err := env.Parse(&cfg); err != nil {
fmt.Printf("%+v\n", err)
}
fmt.Printf("%+v\n", cfg)
}
$ echo qwerty > /tmp/secret
$ echo dvorak > /tmp/password
$ echo coleman > /tmp/certificate
$ SECRET=/tmp/secret \
CERTIFICATE_FILE=/tmp/certificate \
go run main.go
{Secret:qwerty Password:dvorak Certificate:coleman}