telega/chat_action

Keeps a chat action indicator (“typing…”, “sending photo…”) alive while a long-running handler executes.

Telegram clears a chat action after ~5 seconds, so a single api.send_chat_action call is not enough for slow handlers. with_action re-sends the action every ~4 seconds until the wrapped function returns:

import telega/chat_action

fn handler(ctx, _) {
  use <- chat_action.with_action(ctx, chat_action.Typing)
  // ... long-running work: LLM call, file processing, etc.
  reply.with_text(ctx, "Done!")
}

The repeating sender runs in an unlinked worker process that monitors the caller: it stops when the wrapped function returns or when the calling process dies, so no processes are leaked even if the handler crashes.

Types

Action

</>

Chat action to broadcast, mirrors ChatAction.

pub type Action {
  Typing
  UploadPhoto
  RecordVideo
  UploadVideo
  RecordVoice
  UploadVoice
  UploadDocument
  ChooseSticker
  FindLocation
  RecordVideoNote
  UploadVideoNote
}

Constructors

```
Typing
```
```
UploadPhoto
```
```
RecordVideo
```
```
UploadVideo
```
```
RecordVoice
```
```
UploadVoice
```
```
UploadDocument
```
```
ChooseSticker
```
```
FindLocation
```
```
RecordVideoNote
```
```
UploadVideoNote
```

Values

with_action

</>

pub fn with_action(
  ctx ctx: bot.Context(session, error, dependencies),
  action action: Action,
  run run: fn() -> a,
) -> a

Sends action immediately and then re-sends it every ~4 seconds while run executes. Returns the result of run.

use <- chat_action.with_action(ctx, chat_action.Typing)
slow_work(ctx)

with_action_every

</>

pub fn with_action_every(
  ctx ctx: bot.Context(session, error, dependencies),
  action action: Action,
  interval interval: Int,
  run run: fn() -> a,
) -> a

Same as with_action but with a custom re-send interval in milliseconds. Useful for tests and for long uploads where a different cadence is needed.