Valgo is a type-safe, expressive, and extensible validator library for Golang. Valgo is built with generics, so Go 1.18 or higher is required.
Valgo differs from other Golang validation libraries in that the rules are written in functions and not in struct tags. This allows greater flexibility and freedom when it comes to where and how data is validated.
Additionally, Valgo supports customizing and localizing validation messages.
Here is a quick example:
package main
import v "github.com/cohesivestack/valgo"
func main() {
val := v.Is(
v.String("Bob", "full_name").Not().Blank().OfLengthBetween(4, 20),
v.Number(17, "age").GreaterThan(18),
)
if !val.Valid() {
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
}
}
output:
{
"age": [
"Age must be greater than \"18\""
],
"full_name": [
"Full name must have a length between \"4\" and \"20\""
]
}
Valgo is in its early stages, so backward compatibility won't be guaranteed until v1.
Valgo is used in production by Statsignal, but we want community feedback before releasing version 1.
- Valgo
- Quick example
- v0.x.x and backward compatibility
- Table of content
- Getting started
- Using Valgo
Validation
sessionIs(...)
functionValidation.Valid()
functionValidation.IsValid(...)
functionIn(...)
functionInRow(...)
functionCheck(...)
functionAddErrorMessage(...)
function- Merging two
Validation
sessions withValidation.Merge( ... )
New()
function- Custom error message template
- Localizing a validation session with New(...options) function
- Managing common options with Factory
- Custom errors JSON output with Factory
- Validators
- Or Operator in Validators
- Extending Valgo with custom validators
- List of rules by validator type
- Github Code Contribution Guide
- License
Install in your project:
go get github.com/cohesivestack/valgo
Import in your code:
import v github.com/cohesivestack/valgo
Note: You can use any other aliases instead of v
or just reference the
package valgo
directly.
The Validation
session in Valgo is the main structure for validating one or more values. It is called 'Validation' in code.
A validation session will contain one or more Validators, where each Validator
will have the responsibility to validate a value with one or more rules.
There are multiple functions to create a Validation
session, depending on the requirements:
New()
,Is(...)
,In(...)
,InRow(...)
,Check(...)
,AddErrorMessage(...)
Is(...)
is likely to be the most frequently used function in your validations. When Is(...)
is called, the function creates a validation and receives a validator at the same time. In the next section, you will learn more about the Is(...)
function.
The Is(...)
function allows you to pass one or multiple Validator
s, each with their respective values and rules for validation. This creates a Validation
session, which can be used to validate multiple values.
In the following example, we pass multiple Validator
s for the full_name
, age
, and status
values to the Is(...)
function:
val := v.Is(
v.String("Bob", "full_name").Not().Blank().OfLengthBetween(4, 20),
v.Number(17, "age").GreaterThan(18),
v.String("singl", "status").InSlice([]string{"married", "single"})
)
if !val.Valid() {
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
}
output:
```json
{
"age": [
"Age must be greater than \"18\""
],
"full_name": [
"Full name must have a length between \"4\" and \"20\""
],
"status": [
"Status is not valid"
]
}
A Validation
session provide this function, which returns either true
if all their validators are valid or false
if any one of them is invalid.
In the following example, even though the Validator for age
is valid, the Validator
for status
is invalid, making the entire Validator
session invalid.
val := v.Is(
v.Number(21, "age").GreaterThan(18),
v.String("singl", "status").InSlice([]string{"married", "single"}),
)
if !val.Valid() {
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
}
output:
{
"status": [
"Status is not valid"
]
}
This functions allows to check if an specific value in a Validation
session is valid or not. This is very useful for conditional logic.
The following example prints an error message if the age
value is invalid.
val := v.Is(v.Number(16, "age").GreaterThan(18)).
Is(v.String("single", "status").InSlice([]string{"married", "single"}))
if !val.IsValid("age") {
fmt.Println("Warning: someone underage is trying to sign up")
}
output:
Warning: someone underage is trying to sign up
The In(...)
function executes one or more validators in a namespace, so the value names in the error result are prefixed with this namespace. This is useful for validating nested structures.
In the following example we are validating the Person
and the nested Address
structure. We can distinguish the errors of the nested Address
structure in the error results.
type Address struct {
Name string
Street string
}
type Person struct {
Name string
Address Address
}
p := Person{"Bob", Address{"", "1600 Amphitheatre Pkwy"}}
val := v.
Is(v.String(p.Name, "name").OfLengthBetween(4, 20)).
In("address", v.Is(
String(p.Address.Name, "name").Not().Blank(),
String(p.Address.Street, "street").Not().Blank(),
))
if !val.Valid() {
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
}
output:
{
"address.name": [
"Name can't be blank"
],
"name": [
"Name must have a length between \"4\" and \"20\""
]
}
The InRow(...)
function executes one or more validators in a namespace similar to the In(...)
function, but with indexed namespace. So, the value names in the error result are prefixed with this indexed namespace. It is useful for validating nested lists in structures.
In the following example we validate the Person
and the nested list Addresses
. The error results can distinguish the errors of the nested list Addresses
.
type Address struct {
Name string
Street string
}
type Person struct {
Name string
Addresses []Address
}
p := Person{
"Bob",
[]Address{
{"", "1600 Amphitheatre Pkwy"},
{"Home", ""},
},
}
val := v.Is(String(p.Name, "name").OfLengthBetween(4, 20))
for i, a := range p.Addresses {
val.InRow("addresses", i, v.Is(
v.String(a.Name, "name").Not().Blank(),
v.String(a.Street, "street").Not().Blank(),
))
}
if !val.Valid() {
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
}
output:
{
"addresses[0].name": [
"Name can't be blank"
],
"addresses[1].street": [
"Street can't be blank"
],
"name": [
"Name must have a length between \"4\" and \"20\""
]
}
The Check(...)
function, similar to the Is(...)
function, however with Check(...)
the Rules of the Validator parameter are not short-circuited, which means that regardless of whether a previous rule was valid, all rules are checked.
This example shows two rules that fail due to the empty value in the full_name
Validator
, and since the Validator
is not short-circuited, both error messages are added to the error result.
val := v.Check(v.String("", "full_name").Not().Blank().OfLengthBetween(4, 20))
if !val.Valid() {
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
}
output:
{
"full_name": [
"Full name can't be blank",
"Full name must have a length between \"4\" and \"20\""
]
}
The AddErrorMessage
function allows to add an error message to a Validation session without executing a field validator. This function takes in two arguments: name
, which is the name of the field for which the error message is being added, and message
, which is the error message being added to the session.
When an error message is added using this function, the Validation session is marked as invalid, indicating that at least one validation error has occurred.
One use case for the AddErrorMessage
function is to add a general error message for the validation of an entity structure. As shown in the example below, if you have an entity structure for an address
and need to validate multiple fields within it, such as the city
and street
, you can use AddErrorMessage
to include a general error message for the entire address
in case any of the fields fail validation.
type Address struct {
City string
Street string
}
a := Address{"", "1600 Amphitheatre Pkwy"}
val := v.Is(
v.String(a.city, "city").Not().Blank(),
v.String(a.Street, "street").Not().Blank(),
)
if !val.Valid() {
v.AddErrorMessage("address", "The address is wrong!")
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
}
output:
{
"address": [
"The address is wrong!"
],
"city": [
"City can't be blank"
]
}
It's worth noting that there may be other use cases for this function as well.
Using Merge(...)
you can merge two Validation
sessions. When two validations are merged, errors with the same value name will be merged. It is useful for reusing validation logic.
The following example merges the Validation
session returned by the validatePreStatus
function. Since both Validation
sessions validate a value with the name status
, the error returned will return two error messages, and without duplicate the Not().Blank()
error message rule.
type Record struct {
Name string
Status string
}
validatePreStatus := func(status string) *Validation {
regex, _ := regexp.Compile("pre-.+")
return v.
Is(v.String(status, "status").Not().Blank().MatchingTo(regex))
}
r := Record{"Classified", ""}
val := v.Is(
v.String(r.Name, "name").Not().Blank(),
v.String(r.Status, "status").Not().Blank(),
)
val.Merge(validatePreStatus(r.Status))
if !val.Valid() {
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
}
output:
{
"status": [
"Status can't be blank",
"Status must match to \"pre-.+\""
]
}
This function allows you to create a new Validation
session without a Validator
. This is useful for conditional validation or reusing validation logic.
The function accepts an optional parameter of type [Options] struct, which allows you to specify options such as the specific locale code and locale to use, and a custom JSON marshaler for errors.
The following example conditionally adds a Validator
rule for the month_day
value.
month := 5
monthDay := 11
val := v.New()
if month == 6 {
val.Is(v.Number(monthDay, "month_day").LessOrEqualTo(10))
}
if val.Valid() {
fmt.Println("The validation passes")
}
output:
The validation passes
As we mentioned above, you can pass the Options type to the New() function, in order to specify additional options when creating a new Validation session, such as the specific locale code and locale to use, and a custom JSON marshaler for errors. More information about the Options parameter in the following sections.
Customizing the default Valgo error messages is possible through the New()
function as it's explained in the Localizing a validation session with New section, however the Valgo validators allow to customize the template of a specific template validator rule. Below is an example illustrating this with the String empty validator rule.
val := v.Is(v.String("", "address_field", "Address").Not().Empty("{{title}} must not be empty. Please provide the value in the input {{name}}."))
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
output:
{
"address": [
"Address must not be empty. Please provide the value in the input address_field."
]
}
Valgo has localized error messages. The error messages are currently available in English (default), Spanish and German. However, it is possible to set error messages for any locale by passing the Options parameter to the New()
function. Using this parameter, you can also customize the existing Valgo locale messages.
There are two options for localization: localeCode
and locale
. Below, we list the different ways to customize localization with these two parameters.
-
Changing the validation session's locale In the following example, we are setting the Spanish locale:
// Creating the new validation session with other locale val := v.New(v.Options{ LocaleCode: "es" }) // Testing the output val := val.Check(v.String(" ", "nombre").Not().Blank()) out, _ := json.MarshalIndent(val.Error(), "", " ") fmt.Println(string(out))
output:
{ "name": [ "Nombre no puede estar en blanco" ] }
If the specified locale does not exist, Valgo's default English locale will be used. If you wish to change the default locale, you should use a Factory function, which is explained in the Factory section.
-
Changing the locale entries In the example below, we are changing the entry for the "Not Blank" error. Since we are not specifying the
localeCode
, we are using and replacing the default English locale. However, you can also specify another localeCode if necessary.// Creating a new validation session and changing a locale entry val := v.New(v.Options{ Locale: &Locale{ ErrorKeyNotBlank: "{{title}} should not be blank", } }) // Testing the output val := val.Check(v.String(" ", "name").Not().Blank()) out, _ := json.MarshalIndent(val.Error(), "", " ") fmt.Println(string(out))
output:
{ "name": [ "Name should not be blank" ] }
-
Adding a new locale As mentioned previously, Valgo currently only has the English, Spanish and German locales, but we hope to have more soon. However, you can add your own custom locale. Below is an example using the Estonian language:
// Creating a new validation session and adding a new locale with two entries val := v.New(v.Options{ LocaleCode: "ee", Locale: &Locale{ ErrorKeyNotBlank: "{{title}} ei tohi olla tühi", ErrorKeyNotFalse: "{{title}} ei tohi olla vale", } }) // Testing the output val := val.Is( v.String(" ", "name").Not().Blank(), v.Bool(false, "active").Not().False(), ) out, _ := json.MarshalIndent(val.Error(), "", " ") fmt.Println(string(out))
output:
{ "name": [ "Name ei tohi olla tühi" ], "active": [ "Active ei tohi olla vale" ] }
For entries not specified in the custom locale, the default Valgo locale (English) will be used. If you wish to change the default locale, you can use the
Factory
function, which is further explained in the Factory section.
We welcome pull requests for adding new locale messages, but please ensure that the translations are of high quality.
Valgo provides the Factory()
function which allows you to create a valgo factory. With a valgo factory, you can create Validation
sessions with preset options, avoiding having to pass options each time when a Validation is created. This allows more flexibility and easier management of options when creating Validation
sessions.
The Factory
function takes a parameter of type FactoryOptions
struct, which allows you to modify the default locale code, add new locales, and set a custom JSON marshaler for errors. The ValidationFactory
instance created by this function has all the functions to create Validations available in the package level (Is()
, In()
, Check()
, New()
) which creates a new Validation session with the preset options in the factory.
In the following example, we create a Factory
with the default locale code set to Spanish, a new locale added for Estonian. This factory instance enables us to create validation sessions.
factory := v.Factory(v.FactoryOptions{
LocaleCodeDefault: "es",
Locales: map[string]*Locale{
"ee": {
v.ErrorKeyNotBlank: "{{title}} ei tohi olla tühi",
v.ErrorKeyNotFalse: "{{title}} ei tohi olla vale",
},
}
})
// Error will contain the spanish error "Nombre no puede estar en blanco"
v1 := factory.Is(String(" ", "nombre").NotBlank())
// Error will contain the spanish error "Nime ei tohi olla tühi"
v2 := factory.New(Options{LocaleCode: "ee"}).Is(String(" ", "nime").Not().Blank())
It is possible to use the MarshalJsonFunc
parameter of Factory
for customizing the JSON output for errors.
customMarshalJson := func(e *Error) ([]byte, error) {
errors := map[string]interface{}{}
for k, v := range e.errors {
errors[k] = v.Messages()
}
// Add a root key level called errors, which is not set by default in the Valgo implementation.
return json.Marshal(map[string]map[string]interface{}{"errors": errors})
}
// Set the custom Marshal JSON function
factory := v.Factory(v.FactoryOptions{
MarshalJsonFunc: customMarshalJson
})
// Now validate something to check if the output JSON contains the errors root key
val := factory.Is(v.String("", "name").Not().Empty())
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
output:
{
"errors": {
"name": [
"Name can't be empty"
]
}
}
Validators establish the rules for validating a value. The validators are passed to a Validation
session.
For each primitive Golang value type, Valgo provides a Validator
. A Validator
has different functions that set its value's validation rules.
Although Valgo has multiple types of validators, it can be extended with custom validators. Check the section Extending Valgo with Custom Validators for more information.
Validators only require the value to be validated, so, for example, the following code validates a string value by checking if it is empty.
val := v.New(v.String("").Empty())
val.Error()
output:
{
"value_0": [
"Value 0 can't be empty",
]
}
In the example above, since we didn't specify a name for the value, Valgo generates a value_0
name and consequently the Value 0
title in the error message.
However, Validators allow you, optionally, to specify the value's name and title, as shown below:
Validator with value's name:
val := v.New(v.String("", "company_name").Not().Empty())
val.Error()
output:
{
"company_name": [
"Company name can't be empty",
]
}
Validator with value's name and title:
val := v.New(v.String("", "company_name", "Customer").Not().Empty())
val.Error()
output:
{
"company_name": [
"Customer can't be empty",
]
}
Valgo validators have a Not()
function to invert the boolean value associated with the next validator rule function.
In the following example, the call to Valid()
will return false because Not()
inverts the boolean value associated with the Zero()
function.
valid := Is(v.Number(0).Not().Zero()).Valid()
fmt.Println(valid)
output:
false
The ValidatorString
provides functions for setting validation rules for a string
type value, or a custom type based on a string
.
Below is a valid example for every String validator rule.
v.Is(v.String("Dennis Ritchie").EqualTo("Dennis Ritchie"))
v.Is(v.String("Steve Jobs").GreaterThan("Bill Gates"))
v.Is(v.String("Steve Jobs").GreaterOrEqualTo("Elon Musk"))
v.Is(v.String("C#").LessThan("Go"))
v.Is(v.String("Go").LessOrEqualTo("Golang"))
v.Is(v.String("Rust").Between("Go", "Typescript")) // Inclusive
v.Is(v.String("").Empty())
v.Is(v.String(" ").Blank())
v.Is(v.String("Dart").Passing(func(val string) bool { return val == "Dart" }))
v.Is(v.String("processing").InSlice([]string{"idle", "processing", "ready"})
v.Is(v.String("123456").MaxLength(6))
v.Is(v.String("123").MinLength(3))
v.Is(v.String("1234").MinLength(4))
v.Is(v.String("12345").LengthBetween(4,6)) // Inclusive
regex, _ := regexp.Compile("pre-.+"); v.Is(String("pre-approved").MatchingTo(regex))
The ValidatorStringP
provides functions for setting validation rules for a string
type pointer, or a custom type based on a string
pointer.
Below is a valid example for every String pointer validator rule.
x := "Dennis Ritchie"; v.Is(v.StringP(&x).EqualTo("Dennis Ritchie"))
x := "Steve Jobs"; v.Is(v.StringP(&x).GreaterThan("Bill Gates"))
x := "Steve Jobs"; v.Is(v.StringP(&x).GreaterOrEqualTo("Elon Musk"))
x := "C#"; v.Is(v.StringP(&x).LessThan("Go"))
x := "Go"; v.Is(v.StringP(&x).LessOrEqualTo("Golang"))
x := "Rust"; v.Is(v.StringP(&x).Between("Go", "Typescript")) // Inclusive
x := ""; v.Is(v.StringP(&x).Empty())
x := " "; v.Is(v.StringP(&x).Blank())
x := "Dart"; v.Is(v.StringP(&x).Passing(func(val *string) bool { return *val == "Dart" }))
x := "processing"; v.Is(v.StringP(&x).InSlice([]string{"idle", "processing", "ready"}))
x := "123456"; v.Is(v.StringP(&x).MaxLength(6))
x := "123"; v.Is(v.StringP(&x).MinLength(3))
x := "1234"; v.Is(v.StringP(&x).MinLength(4))
x := "12345"; v.Is(v.StringP(&x).LengthBetween(4,6)) // Inclusive
x := "pre-approved"; regex, _ := regexp.Compile("pre-.+"); v.Is(StringP(&x).MatchingTo(regex))
x := ""; v.Is(v.StringP(&x).EmptyOrNil())
x := " "; v.Is(v.StringP(&x).BlankOrNil())
var x *string; v.Is(v.StringP(x).Nil())
The Number validator provides functions for setting validation rules for a TypeNumber
value, or a custom type based on a TypeNumber
.
TypeNumber
is a generic interface defined by Valgo that generalizes any standard Golang type. Below is Valgo's definition of TypeNumber
:
type TypeNumber interface {
~int |
~int8 |
~int16 |
~int32 |
~int64 |
~uint |
~uint8 |
~uint16 |
~uint32 |
~uint64 |
~float32 |
~float64
}
Below is a valid example for every Number validator rule.
v.Is(v.Number(10).EqualTo(10))
v.Is(v.Number(11).GreaterThan(10))
v.Is(v.Number(10).GreaterOrEqualTo(10))
v.Is(v.Number(10).LessThan(11))
v.Is(v.Number(10).LessOrEqualTo(10))
v.Is(v.Number(11).Between(10, 12)) // Inclusive
v.Is(v.Number(0).Zero())
v.Is(v.Number(10).Passing(func(val int) bool { return val == 10 }))
v.Is(v.Number(20).InSlice([]int{10, 20, 30}))
The Number pointer validator provides functions for setting validation rules for a TypeNumber
pointer, or a custom type based on a TypeNumber
pointer.
As it's explained in Number validator, the TypeNumber
is a generic interface defined by Valgo that generalizes any standard Golang type.
Below is a valid example for every Number pointer validator rule.
x := 10; v.Is(v.NumberP(&x).EqualTo(10))
x := 11; v.Is(v.NumberP(&x).GreaterThan(10))
x := 10; v.Is(v.NumberP(&x).GreaterOrEqualTo(10))
x := 10; v.Is(v.NumberP(&x).LessThan(11))
x := 10; v.Is(v.NumberP(&x).LessOrEqualTo(10))
x := 11; v.Is(v.NumberP(&x).Between(10, 12)) // Inclusive
x := 0; v.Is(v.NumberP(&x).Zero())
x := 10; v.Is(v.NumberP(&x).Passing(func(val *int) bool { return *val == 10 }))
x := 20; v.Is(v.NumberP(&x).InSlice([]int{10, 20, 30}))
x := 0; v.Is(v.NumberP(&x).ZeroOrNil())
var x *int; v.Is(v.NumberP(x).Nil())
While the validator Number
works with all golang number types, Valgo also has a validator for each type. You can use them if you prefer or need a stronger safe type code.
Following is a list of functions for every specific number type validator, along with their equivalent pointer validators.
Int(v int) IntP(v *int)
Int8(v int8) Int8P(v *int8)
Int16(v int16) Int16P(v *int16)
Int32(v int32) Int32P(v *int32)
Int64(v int64) Int64P(v *int64)
Uint(v uint) UintP(v *uint)
Uint8(v uint8) Uint8P(v *uint8)
Uint16(v uint16) Uint16P(v *uint16)
Uint32(v uint32) Uint32P(v *uint32)
Uint64(v uint64) Uint64P(v *uint64)
Float32(v float32) Float32P(v *float32)
Float64(v float64) Float64P(v *float64)
Byte(v byte) ByteP(v *byte)
Rune(v byte) RuneP(v *byte)
These validators have the same rule functions as the Number
validator.
Similar to the Number
validator, custom types can be passed based on the specific number type.
The Bool validator provides functions for setting validation rules for a bool
type value, or a custom type based on a bool
.
Below is a valid example for every Bool validator rule.
v.Is(v.Bool(true).EqualTo(true))
v.Is(v.Bool(true).True())
v.Is(v.Bool(false).False())
v.Is(v.Bool(true).Passing(func(val bool) bool { return val == true }))
v.Is(v.Bool(true).InSlice([]string{true, false}))
The Bool pointer validator provides functions for setting validation rules for a bool
pointer, or a custom type based on a bool
pointer.
Below is a valid example for every Bool pointer validator rule.
x := true; v.Is(v.BoolP(&x).EqualTo(true))
x := true; v.Is(v.BoolP(&x).True())
x := false; v.Is(v.BoolP(&x).False())
x := true; v.Is(v.BoolP(&x).Passing(func(val *bool) bool { return val == true }))
x := true; v.Is(v.BoolP(&x).InSlice([]string{true, false}))
x := false; v.Is(v.BoolP(&x).FalseOrNil())
var x *bool; v.Is(v.BoolP(x).Nil())
The ValidatorTime
provides functions for setting validation rules for a time.Time
type value, or a custom type based on a time.Time
.
Below is a valid example for every Time validator rule.
import "time"
v.Is(v.Time(time.Now()).EqualTo(time.Now()))
v.Is(v.Time(time.Now()).After(time.Now().Add(-time.Hour)))
v.Is(v.Time(time.Now()).AfterOrEqualTo(time.Now().Add(-time.Hour)))
v.Is(v.Time(time.Now()).Before(time.Now().Add(time.Hour)))
v.Is(v.Time(time.Now()).BeforeOrEqualTo(time.Now().Add(time.Hour)))
v.Is(v.Time(time.Now()).Between(time.Now().Add(-time.Hour), time.Now().Add(2*time.Hour))) // Inclusive
v.Is(v.Time(time.Time{}).Zero())
v.Is(v.Time(time.Now()).Passing(func(val time.Time) bool { return val.Before(time.Now().Add(2*time.Hour)) }))
v.Is(v.Time(time.Now()).InSlice([]time.Time{time.Now(), time.Now().Add(time.Hour)}))
The ValidatorTimeP
provides functions for setting validation rules for a time.Time
type pointer, or a custom type based on a time.Time
pointer.
Below is a valid example for every Time pointer validator rule.
import "time"
x := time.Now(); v.Is(v.TimeP(&x).EqualTo(time.Now()))
x = time.Now(); v.Is(v.TimeP(&x).After(time.Now().Add(-time.Hour)))
x = time.Now(); v.Is(v.TimeP(&x).AfterOrEqualTo(time.Now().Add(-time.Hour)))
x = time.Now(); v.Is(v.TimeP(&x).Before(time.Now().Add(time.Hour)))
x = time.Now(); v.Is(v.TimeP(&x).BeforeOrEqualTo(time.Now().Add(time.Hour)))
x = time.Now(); v.Is(v.TimeP(&x).Between(time.Now().Add(-time.Hour), time.Now().Add(2*time.Hour))) // Inclusive
x = time.Time{}; v.Is(v.TimeP(&x).Zero())
x = time.Now(); v.Is(v.TimeP(&x).Passing(func(val *time.Time) bool { return val.Before(time.Now().Add(2*time.Hour)) }))
x = time.Now(); v.Is(v.TimeP(&x).InSlice([]time.Time{time.Now(), time.Now().Add(time.Hour)}))
var x *time.Time; v.Is(v.TimeP(x).Nil())
x = new(time.Time); v.Is(v.TimeP(x).NilOrZero())
With the Any validator, you can set validation rules for any value or pointer.
Below is a valid example of every Any validator rule.
v.Is(v.Any("react").EqualTo("react"))
v.Is(v.Any("svelte").Passing(func(val *bool) bool { return val == "svelte" }))
var x *bool; v.Is(v.Any(x).Nil())
For the EqualTo(v any)
rule function, the parameter type must match the type used by the Any()
function, otherwise it will be invalid. In the following example, since the value passed to the Any(...)
function is int
, and EqualTo(...)
compares it with int64, the validation is invalid.
valid := v.Is(v.Any(10).EqualTo(int64(10))).Valid()
fmt.Println(valid)
output
false
If a pointer is used, the same pointer must be passed to EqualTo(v any)
as it is passed to Any(v any)
, in order to get a valid validation. The following example illustrates it.
// Valid since the same pointers are compared
numberA := 10
v.Is(v.Any(&numberA).EqualTo(&numberA)).Valid()
// Invalid since different pointers are compared
numberB := 10
v.Is(v.Any(&numberA).EqualTo(&numberB)).Valid()
All golang validators allow to pass a custom type based on its value type. Bellow some valid examples.
type Status string
var status Status = "up"
val := v.Is(v.String(status).InSlice([]Status{"down", "up", "paused"}))
type Level int
var level Level = 1
val = v.Is(v.Int(level).LessThan(Level(2)))
type Stage int64
var stage Stage = 2
val := v.Is(v.NumberP(&stage).GreaterThan(Stage(1)))
The Or
operator function enables developers to combine validator rules using a logical OR chain. This addition allows for more nuanced validator scenarios, where a value may satisfy one of multiple conditions to be considered valid.
In Valgo, validator rules are typically chained together using an implicit AND logic. This means that for a value to be deemed valid, it must satisfy all specified conditions. The Or
operator provides an alternative by allowing conditions to be linked with OR logic. In such cases, a value is considered valid if it meets at least one of the chained conditions.
The Or
operator follows a simple left-to-right boolean priority, akin to the Go language's approach to evaluating boolean expressions. Valgo does not have an equivalent to parentheses in API functions, in order to keep the syntax simple and readable. We believe that complex boolean logic becomes harder to read with a Fluent API interface, so for those cases, it is preferred to use imperative Go programming language constructs.
To utilize the Or
operator, simply insert .Or().
between two conditions within your validator chain. Here's a basic example:
v := Is(Bool(true).True().Or().False())
In this case, the validator passes because the boolean value true
satisfies the first condition before the Or()
operator.
- Implicit AND Logic: By default, when validators are chained without specifying the
Or()
operator, they are combined using an AND logic. Each condition must be met for the validation to pass. - No Short-circuiting for
Check
: Unlike theIs
function, which evaluates conditions lazily and may short-circuit (stop evaluating once the overall outcome is determined), theCheck
function ensures that all conditions are evaluated, regardless of their order and the use ofOr
.
Below are examples demonstrating different scenarios using the Or
operator, including combinations with the Not
operator and multiple Or
conditions in sequence. These examples illustrate how you can tailor complex validation logic to suit your needs.
// Validation with two valid OR conditions
v = Is(Bool(true).True().Or().True())
assert.True(t, v.Valid())
// Validation with a valid OR condition followed by an invalid AND condition
v = Is(Bool(true).False().Or().True().False())
assert.False(t, v.Valid())
// Validation combining NOT and OR operators
v = Is(Bool(true).Not().False().Or().False())
assert.True(t, v.Valid())
These examples are intended to provide a clear understanding of how to effectively use the Or
operator in your validations. By leveraging this functionality, you can create more flexible and powerful validation rules, enhancing the robustness and usability of your applications.
While all validators in Golang provide a Passing(...)
function, which allows you to use a custom validator function, Valgo also allows you to create your own validator.
With this functionality Valgo can be extended with Validator libraries, which we encourage the community to do.
For example, let's say we want to create a validation for the following ID
struct, where a user must provide at least one property.
The struct to validate:
// Type to validate
type ID struct {
Phone string
Email string
}
the custom validator code:
// The custom validator type
type ValidatorID struct {
context *valgo.ValidatorContext
}
// The custom validator implementation of `valgo.Validator`
func (validator *ValidatorID) Context() *valgo.ValidatorContext {
return validator.context
}
// Here is the function that passes the value to the custom validator
func IDValue(value ID, nameAndTitle ...string) *ValidatorID {
return &ValidatorID{context: valgo.NewContext(value, nameAndTitle...)}
}
// The Empty rule implementation
func (validator *ValidatorID) Empty(template ...string) *ValidatorID {
validator.context.Add(
func() bool {
return len(strings.Trim(validator.context.Value().(ID).Phone, " ")) == 0 &&
len(strings.Trim(validator.context.Value().(ID).Email, " ")) == 0
},
v.ErrorKeyEmpty, template...)
return validator
}
// It would be possible to create a rule NotEmpty() instead of Empty(), but if you add a Not() function then your validator will be more flexible.
func (validator *ValidatorID) Not() *ValidatorID {
validator.context.Not()
return validator
}
using our validator:
val := v.Is(IDValue(ID{}, "id").Not().Empty())
out, _ := json.MarshalIndent(val.Error(), "", " ")
fmt.Println(string(out))
output:
{
"identification": [
"Id can't be empty"
]
}
-
String
validatorEqualTo
GreaterThan
GreaterOrEqualTo
LessThan
LessOrEqualTo
Between
Empty
Blank
Passing
InSlice
MatchingTo
MaxLength
MinLength
Length
LengthBetween
-
StringP
validator - for string pointerEqualTo
GreaterThan
GreaterOrEqualTo
LessThan
LessOrEqualTo
Between
Empty
Blank
Passing
InSlice
MatchingTo
MaxLength
MinLength
Length
LengthBetween
BlankOrNil
EmptyOrNil
Nil
-
Bool
validatorEqualTo
Passing
True
False
InSlice
-
BoolP
validator - for boolean pointerEqualTo
Passing
True
False
InSlice
FalseOrNil
Nil
-
Number
andInt
,Int8
,Int16
,Int32
,Int64
,Uint
,Uint8
,Uint16
,Uint32
,Uint64
,Float32
,Float64
,Byte
,Rune
- for number pointerEqualTo
GreaterThan
GreaterOrEqualTo
LessThan
LessOrEqualTo
Between
Zero
InSlice
Passing
-
NumberP
andIntP
,Int8P
,Int16P
,Int32P
,Int64P
,UintP
,Uint8P
,Uint16P
,Uint32P
,Uint64P
,Float32P
,Float64P
,ByteP
,RuneP
- for number pointerEqualTo
GreaterThan
GreaterOrEqualTo
LessThan
LessOrEqualTo
Between
Zero
InSlice
Passing
ZeroOrNil
Nil
-
Time
validatorEqualTo
After
AfterOrEqualTo
Before
BeforeOrEqualTo
Between
Zero
Passing
InSlice
-
TimeP
validator - fortime.Time
pointerEqualTo
After
AfterOrEqualTo
Before
BeforeOrEqualTo
Between
Zero
Passing
InSlice
Nil
NilOrZero
-
Any
validatorEqualTo
Passing
Nil
We welcome contributions to our project! To make the process smooth and efficient, please follow these guidelines when submitting code:
-
Discuss changes with the community: We encourage contributors to discuss their proposed changes or improvements with the community before starting to code. This ensures that the changes align with the focus and purpose of the project, and that other contributors are aware of the work being done.
-
Make commits small and cohesive: It is important to keep your commits focused on a single task or change. This makes it easier to review and understand your changes.
-
Check code formatting with go fmt: Before submitting your code, please ensure that it is properly formatted using the go fmt command.
-
Make tests to cover your changes: Please include tests that cover the changes you have made. This ensures that your code is functional and reduces the likelihood of bugs.
-
Update golang docs and README to cover your changes: If you have made changes that affect documentation or the README file, please update them accordingly.
-
Keep a respectful language with a collaborative tune: We value a positive and collaborative community. Please use respectful language when communicating with other contributors or maintainers.
Copyright © 2023 Carlos Forero
Valgo is released under the MIT License