This library makes it easy to build and change business logic.
It makes code:
- Match flow diagrams.
- Easy to read.
- Easy to test.
- Modular.
go get github.com/ofabricio/adt
package main
import "fmt"
import . "github.com/ofabricio/adt/a"
func main() {
fizzbuzz := Branch(
Case(And(IsMod(3), IsMod(5)), Print("FizzBuzz")),
Case(IsMod(3), Print("Fizz")),
Case(IsMod(5), Print("Buzz")),
Else(PrintIndex),
)
for i := 1; i <= 15; i++ {
fizzbuzz.Run(i)
}
// Output:
// 1 2 Fizz 4 Buzz Fizz 7 8 Fizz Buzz 11 Fizz 13 14 FizzBuzz
}
func IsMod(v int) Cond[int] {
return func(i int) bool {
return i%v == 0
}
}
func Print(s string) Func[int] {
return func(int) error {
fmt.Print(s, " ")
return nil
}
}
func PrintIndex(i int) error {
fmt.Print(i, " ")
return nil
}Don't let this simple example fool you. This is a very simple, yet powerful lib.
This lib makes code maintainable. You can plug and unplug functions and conditions like Lego; and run parallel code with a single change. Also your code will look like your flow diagram and this allows you to change your code as easy as changing your diagram.
Think of a function (Func[T] type) as the rectangular boxes in a flow diagram;
and think of a branch (Branch operator) as the diamonds in a flow diagram.
Do that and you're good to go.
graph LR
A[Func] --> B{Branch}
B -->|Case| C[Func]
B -->|Else| D[Func]
Run runs the ADT. It is the entry point where you feed the tree with the input for processing.
import . "github.com/ofabricio/adt/a"
func Example() {
a := All(Add(2), Add(3), Add(5))
job := 2
err := a.Run(&job)
fmt.Println(job, err)
// Output:
// 12 <nil>
}
func Add(n int) Func[*int] {
return func(v *int) error {
*v += n
return nil
}
}The code above matches this diagram:
graph LR
S((2)) --> A
subgraph All
A[Add 2] --> B[Add 3] --> C[Add 5]
end
C --> E((12))
Tip! You can use complex structs as a job input so that you can have context, etc. Example:
type Job struct {
Input map[string]int // Just an input example.
Result map[string]int // Maybe you want the result separately.
mut sync.Mutex // Maybe you want concurrency control.
ctx context.Context // Maybe you want context.
log log.Logger // Maybe you want log.
}Branch operator works like a switch statement, it runs the first case that evaluates to true.
Inside a branch you can use Case and Else operators.
Caseruns its function if its condition evaluates to true.Elsealways runs its function because it always evaluates to true.
Note that you can use Branch inside a Case or Else to create a decision tree.
func Example() {
a := Branch(
Case(Is(1), Say("One")),
Case(Is(2), Say("Two")),
Else(Say("Ten")),
)
a.Run(2)
// Output:
// Two
}
func Is(n int) Cond[int] {
return func(v int) bool {
return v == n
}
}
func Say(msg int) Func[int] {
return func(v int) error {
fmt.Println(msg)
return nil
}
}The code above matches this diagram:
graph LR
S((2)) --> A
subgraph Branch
A{Branch} -->|Case 1| B[Say One]
A -->|Case 2| C[Say Two]
A -->|Else| D[Say Ten]
end
C --> E((Two))
Case is part of the Branch operator.
It runs its function if its condition evaluates to true. See Branch operator example.
Else is part of the Branch operator.
It always runs its function because it always evaluates to true.
See Branch operator example.
All runs functions sequentially. It stops if one of them returns an error.
func Example() {
a := All(Add(2), Add(3), Add(5))
job := 2
a.Run(&job)
fmt.Println(job)
// Output:
// 12
}The code above matches this diagram:
graph LR
S((2)) --> A
subgraph All
A[Add 2] --> B[Add 3] --> C[Add 5]
end
C --> E((12))
Now supposing that Add(3) failed, it would behave like this:
graph LR
S((2)) --> A
subgraph All
A[Add 2] --> B[Error] --> C[Add 5]
end
B ---> E((4))
style B fill:#eb6c6c
Retry retries a function if it returns an error.
In the following example, suppose that Say("Two") returns and prints an error Oops the first time it runs.
The output would be like below.
func Example() {
a := All(Say("One"), Say("Two").Retry(2), Say("Ten"))
a.Run(0)
// Output:
// One
// Oops
// Two
// Ten
}Catch is called when an error happens.
Note that whenever any function in the ADT returns an error the ADT stops and Run() returns the error.
In order to not stop it you need to catch and handle the error (by returning nil).
In the following example, suppose that Say("Two") returns and prints an error Oops the first time it runs.
By handling it the output would be like below.
func Example() {
a := All(Say("One"), Say("Two").Catch(Handle), Say("Ten"))
a.Run(0)
// Output:
// One
// Oops
// Got the error oops, but it's fine now. Keep going.
// Ten
}
func Handle(v int, err error) error {
fmt.Printf("Got the error %v, but it's fine now. Keep going.\n", err)
return nil
}On is called on either success or error.
func Example() {
a := Say("One").On(Handle)
a.Run(0)
// Output:
// One
// 0 nil
}
func Handle(v int, err error) error {
fmt.Println(v, err)
return err
}Parallel runs functions concurrently. It awaits until all functions are complete. It returns the first error found.
func Example() {
a := Parallel(Say("One"), Say("Two"), Say("Ten"))
a.Run(0)
// Output:
// Different output on each run.
}The code above matches this diagram:
flowchart LR
subgraph Parallel
direction LR
A[Say One]
B[Say Two]
C[Say Ten]
end
S(( )) --> Parallel --> E
E{{"Ten<br>One<br>Two"}}
Semaphore is like Parallel, but you control the number of go routines it uses.
func Example() {
a := Semaphore(2, Say("One"), Say("Two"), Say("Ten"))
a.Run(0)
// Output:
// Different output on each run.
}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent