Macro is a unified workspace for teams: email, chat, docs, tasks, agents, calls, and CRM — @-linked together with shared AI memory.
TypeScript agentaiai-agentsall-in-onecrmcrm-systememaillinearmcpmessagingnotesnotionofficerustslackslack-alternativestartupstartup-toolstypescriptworkspace
use aws_sdk_sqs::types::SendMessageBatchRequestEntry;
use std::collections::HashMap;
use crate::{MAX_BATCH_SIZE, message_attribute::build_string_message_attribute};
fn construct_message_attributes(
user_id: Option<&str>,
document_id: &str,
) -> anyhow::Result<HashMap<String, aws_sdk_sqs::types::MessageAttributeValue>> {
let mut message_attributes = HashMap::new();
if let Some(user_id) = user_id {
message_attributes.insert(
"user_id".to_string(),
build_string_message_attribute(user_id)?,
);
}
message_attributes.insert(
"document_id".to_string(),
build_string_message_attribute(document_id)?,
);
Ok(message_attributes)
}
/// Bulk enqueues document delete messages to the document delete queue
#[tracing::instrument(skip(sqs_client, documents))]
pub(crate) async fn bulk_enqueue_document_delete(
sqs_client: &aws_sdk_sqs::Client,
queue_url: &str,
documents: Vec<String>,
) -> anyhow::Result<()> {
let mut entries: Vec<SendMessageBatchRequestEntry> = vec![];
for document_id in documents {
tracing::trace!(document_id=?document_id, "enqueueing document delete");
let message_attributes = construct_message_attributes(None, document_id.as_str())?;
let batch_requesst = SendMessageBatchRequestEntry::builder()
.id(&document_id)
.message_body(&document_id)
.set_message_attributes(Some(message_attributes))
.build()?;
entries.push(batch_requesst);
}
if entries.is_empty() {
tracing::warn!("no entries to enqueue");
return Ok(());
}
// Batch the entries in chunks of 10 and send each batch separately
for chunk in entries.chunks(MAX_BATCH_SIZE) {
let chunk_to_send = chunk.to_vec();
// Send the batch
sqs_client
.send_message_batch()
.set_entries(Some(chunk_to_send))
.queue_url(queue_url)
.send()
.await?;
}
Ok(())
}
#[tracing::instrument(skip(sqs_client, documents))]
pub(crate) async fn bulk_enqueue_document_delete_with_owner(
sqs_client: &aws_sdk_sqs::Client,
queue_url: &str,
documents: Vec<(String, String)>,
) -> anyhow::Result<()> {
let mut entries: Vec<SendMessageBatchRequestEntry> = vec![];
for (document_id, user_id) in documents {
tracing::trace!(document_id=?document_id, "enqueueing document delete");
let message_attributes =
construct_message_attributes(Some(user_id.as_str()), document_id.as_str())?;
let batch_requesst = SendMessageBatchRequestEntry::builder()
.id(&document_id)
.message_body(&document_id)
.set_message_attributes(Some(message_attributes))
.build()?;
entries.push(batch_requesst);
}
if entries.is_empty() {
tracing::warn!("no entries to enqueue");
return Ok(());
}
// Batch the entries in chunks of 10 and send each batch separately
for chunk in entries.chunks(MAX_BATCH_SIZE) {
let chunk_to_send = chunk.to_vec();
// Send the batch
sqs_client
.send_message_batch()
.set_entries(Some(chunk_to_send))
.queue_url(queue_url)
.send()
.await?;
}
Ok(())
}
/// Enqueues document delete message to the document delete queue
#[tracing::instrument(skip(sqs_client))]
pub(crate) async fn enqueue_document_delete(
sqs_client: &aws_sdk_sqs::Client,
queue_url: &str,
user_id: &str,
document_id: &str,
) -> anyhow::Result<()> {
let message_attributes = construct_message_attributes(Some(user_id), document_id)?;
sqs_client
.send_message()
.queue_url(queue_url)
.set_message_attributes(Some(message_attributes))
.message_body(document_id)
.send()
.await?;
Ok(())
}