# Async usage

**URL:** https://discourse.julialang.org/t/async-usage/59218
**Category:** New to Julia
**Created:** [April 13, 2021, 5:48pm UTC](https://discourse.julialang.org/t/async-usage/59218 "2021-04-13T17:48:02Z")
**Posts on this page:** 3
**Page:** 1

<div class="post-metadata">

### Author: ![LinasBa](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/linasba/32/20957_2.png) [@LinasBa](https://discourse.julialang.org/u/LinasBa)
#### Post date: [April 13, 2021, 5:48pm UTC](https://discourse.julialang.org/t/async-usage/59218/1 "2021-04-13T17:48:03Z")

</div>

Hello,

I’m trying to use Telegram API with asyncronous programming. My program basically validates logs for specifeid arguments.

My first attempt was:

```julia
using Telegram, Random

# validation is IO bound log checking
function validate1(name) 
    sleep(1) 
    rand() < 0.9
end

function validate2(name) 
    sleep(1) 
    rand() < 0.8
end

action_telegram(name) = Telegram.API.sendMessage(text="$name")
action2(name) = sleep(10) # long IO bound action

function validate(name)
    # this list will have many pairs
    fs = [validate1 => action_telegram, validate2 => action2]
    for (f, action) ∈ fs
        if f(name)
            println("$(nameof(f)) -> $(nameof(action))")
            action(name)
            return
        end
    end
end

function test1()
    tg = TelegramClient("", chat_id = "")
    names = repeat(["test"], 20)
    @sync for name ∈ names
        @async validate(name)
    end
end

```

This would be perfect, but I get “Too many request” from Telegram servers, so ideally I would like to group all messages in one.

I did this workaround:

```julia
function validate_alt(name)
    fs = [validate1 => action_telegram, validate2 => action2]
    for (f, action) ∈ fs
        if f(name)
            println("$(nameof(f)) -> $(nameof(action))")
            if nameof(action) == :action_telegram
                return true
            else
                action(name)
            end
        end
    end
    return false
end

function test2()
    tg = TelegramClient("", chat_id = "")
    names = repeat(["test"], 20)
    mask = zeros(Bool, length(names))
    @sync for (i, name) ∈ enumerate(names)
        @async mask[i] = validate_alt(name)
    end
    action_telegram(join(names[mask], "\n"))
    return
end

```

But this is rather inconvenient since in my real code I have many “action” and “validation” functions, so this workaround looks out of place.

I would be greateful for any suggestions for more elegant solution. I don’t mind restructuring my code.

---

<div class="post-metadata">

### Author: ![Skoffer](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/skoffer/32/378_2.png) [@Skoffer](https://discourse.julialang.org/u/Skoffer)
#### Post date: [April 13, 2021, 7:14pm UTC](https://discourse.julialang.org/t/async-usage/59218/2 "2021-04-13T19:14:39Z")

</div>

You can use producer/consumers pattern. Just make a `Channel` which will hold commands, and fill it with producer. On the other side of this channel run as many consumers as telegram allows, `take!` messages and send them. This is pretty standard tactic.

Some relevant information:  
[https://docs.julialang.org/en/v1/manual/asynchronous-programming/#More-on-Channels](https://docs.julialang.org/en/v1/manual/asynchronous-programming/#More-on-Channels)

---

<div class="post-metadata">

### Author: ![LinasBa](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/linasba/32/20957_2.png) [@LinasBa](https://discourse.julialang.org/u/LinasBa)
#### Post date: [April 14, 2021, 11:43am UTC](https://discourse.julialang.org/t/async-usage/59218/3 "2021-04-14T11:43:48Z")

</div>

Thanks, Channel was a perfect fit, got it working nicely.