feat: add Rate Limits on trello API and own parallisation

- Limits the rate of API calls to Trello
- Limits the number of stacks, cards, labels and attachments that are running in parallel

src/import/attachment.rs

@@ -1,12 +1,15 @@
+use color_eyre::eyre::Context;
 //
-use kameo::{mailbox::unbounded::UnboundedMailbox, Actor};
+use crate::import::rate_limit::{RateLimitActor, RequestToken};
-
 use crate::{
+    ask, e, f,
     nextcloud::model::{NextcloudBoardId, NextcloudCardId, NextcloudStackId},
     on_actor_start, p,
     trello::model::{attachment::TrelloAttachment, card::TrelloCardId},
     FullCtx,
 };
+use kameo::actor::ActorRef;
+use kameo::{mailbox::unbounded::UnboundedMailbox, Actor};
 
 pub(crate) struct ImportAttachmentActor {
     ctx: FullCtx,
@@ -15,6 +18,7 @@ pub(crate) struct ImportAttachmentActor {
     nextcloud_card_id: NextcloudCardId,
     trello_card_id: TrelloCardId,
     trello_attachment: TrelloAttachment,
+    rate_limit_actor: ActorRef<RateLimitActor>,
 }
 impl ImportAttachmentActor {
     pub fn new(
@@ -24,6 +28,7 @@ impl ImportAttachmentActor {
         nextcloud_card_id: NextcloudCardId,
         trello_card_id: TrelloCardId,
         trello_attachment: TrelloAttachment,
+        rate_limit_actor: ActorRef<RateLimitActor>,
     ) -> Self {
         Self {
             ctx,
@@ -32,6 +37,7 @@ impl ImportAttachmentActor {
             nextcloud_card_id,
             trello_card_id,
             trello_attachment,
+            rate_limit_actor,
         }
     }
 }
@@ -49,6 +55,7 @@ impl Actor for ImportAttachmentActor {
         let dir = this.ctx.temp_fs.dir(&dir_path);
         dir.create()?;
         let file_path = dir_path.join(&*this.trello_attachment.name);
+        ask!(this.rate_limit_actor, RequestToken)?;
         let attachment_path = this
             .ctx
             .trello_client()
@@ -58,7 +65,8 @@ impl Actor for ImportAttachmentActor {
                 Some(&file_path),
                 &this.ctx.temp_fs,
             )
-            .await?;
+            .await
+            .with_context(|| f!("saving attachment {}", file_path.display()))?;
         let attachment_file = this.ctx.temp_fs.file(&attachment_path);
         // - - - upload the attachment to nextcloud card
         let attachment = this
@@ -71,7 +79,16 @@ impl Actor for ImportAttachmentActor {
                 &attachment_file,
             )
             .await
-            .result?;
+            .result
+            .inspect_err(|e| {
+                e!(
+                    this.ctx.prt,
+                    ">> Error adding attachment {} to card: {}",
+                    file_path.display(),
+                    e.to_string()
+                );
+            })
+            .with_context(|| f!("adding attachment to card {}", file_path.display()))?;
         p!(
             this.ctx.prt,
             ">> Attachment added: {}:{}",
@@ -80,8 +97,11 @@ impl Actor for ImportAttachmentActor {
             // attachment.extended_data.mimetype
         );
         // delete local copy of attachment
-        attachment_file.remove()?;
+        attachment_file
-        dir.remove()?;
+            .remove()
+            .with_context(|| f!("deleting temp file {attachment_file}"))?;
+        dir.remove()
+            .with_context(|| f!("deleting temp dir {dir}"))?;
 
         Ok(actor_ref.stop_gracefully().await?)
     });

src/import/card.rs

@@ -8,7 +8,10 @@ use kameo::{
     Actor,
 };
 
+use crate::import::rate_limit::{RateLimitActor, RequestToken};
+use crate::rate_limit::RateLimit;
 use crate::{
+    ask,
     import::{attachment::ImportAttachmentActor, label::ImportLabelActor, labels::LabelsActor},
     nextcloud::{
         client::DeckClient,
@@ -28,6 +31,7 @@ pub(crate) struct ImportCardActor {
     nextcloud_board_id: NextcloudBoardId,
     nextcloud_stack_id: NextcloudStackId,
     labels_actor_ref: ActorRef<LabelsActor>,
+    rate_limit_actor: ActorRef<RateLimitActor>,
 
     labels_children: HashMap<ActorID, (TrelloLabelName, ActorRef<ImportLabelActor>)>,
     attachments_children: HashMap<ActorID, (TrelloAttachmentName, ActorRef<ImportAttachmentActor>)>,
@@ -39,6 +43,7 @@ impl ImportCardActor {
         nextcloud_board_id: NextcloudBoardId,
         nextcloud_stack_id: NextcloudStackId,
         labels_actor_ref: ActorRef<LabelsActor>,
+        rate_limit_actor: ActorRef<RateLimitActor>,
     ) -> Self {
         Self {
             ctx,
@@ -46,6 +51,7 @@ impl ImportCardActor {
             nextcloud_board_id,
             nextcloud_stack_id,
             labels_actor_ref,
+            rate_limit_actor,
 
             labels_children: Default::default(),
             attachments_children: Default::default(),
@@ -77,10 +83,13 @@ impl Actor for ImportCardActor {
             .await
             .result?;
 
+        let mut limit = RateLimit::new("labels & attachments", 10, 10.0, this.ctx.now());
+
         // - - for each label on the trello card
         let mut labels = vec![];
         std::mem::swap(&mut this.trello_card.labels, &mut labels);
         for trello_label in labels.into_iter() {
+            limit.block_until_token_available(this.ctx.now()).await;
             let trello_label_name = trello_label.name.clone();
             let child = spawn_in_thread!(
                 actor_ref,
@@ -97,12 +106,14 @@ impl Actor for ImportCardActor {
                 .insert(child.id(), (trello_label_name, child.clone()));
         }
 
+        ask!(this.rate_limit_actor, RequestToken)?;
         let attachments = trello_client
             .card(&this.trello_card.id)
             .await
             .result?
             .attachments;
         for trello_attachment in attachments.into_iter() {
+            limit.block_until_token_available(this.ctx.now()).await;
             if trello_attachment.is_upload {
                 let trello_attachment_name = trello_attachment.name.clone();
                 let child = spawn_in_thread!(
@@ -114,6 +125,7 @@ impl Actor for ImportCardActor {
                         nextcloud_card.id,
                         this.trello_card.id.clone(),
                         trello_attachment,
+                        this.rate_limit_actor.clone(),
                     )
                 );
                 this.attachments_children

src/import/mod.rs

@@ -4,8 +4,12 @@ use kameo::actor::spawn_in_thread;
 
 use crate::{
     f,
-    import::{labels::LabelsActor, stacks::ImportStacksActor, supervisor::Supervisor},
+    import::{
+        labels::LabelsActor, rate_limit::RateLimitActor, stacks::ImportStacksActor,
+        supervisor::Supervisor,
+    },
     nextcloud::model::NextcloudBoardTitle,
+    rate_limit::RateLimit,
     spawn_in_thread,
     trello::model::board::{TrelloBoardId, TrelloBoardName, TrelloBoards},
     FullCtx,
@@ -15,6 +19,7 @@ mod attachment;
 mod card;
 mod label;
 mod labels;
+mod rate_limit;
 mod stack;
 mod stacks;
 mod supervisor;
@@ -72,13 +77,16 @@ pub(crate) async fn run(ctx: &FullCtx) -> color_eyre::Result<()> {
 
     // Start importing
 
+    println!("> Loading Nextcloud board: {nextcloud_board_name}");
     // get the id of the selected board
     let nextcloud_board_id = nextcloud_boards
         .iter()
         .find(|b| b.title == nextcloud_board_name)
         .map(|b| b.id)
         .expect("find selected board");
+    println!("> Loaded Nextcloud board: {nextcloud_board_name}");
 
+    println!("> Spawning actors");
     let supervisor = spawn_in_thread(Supervisor);
 
     let labels_actor_ref = spawn_in_thread!(
@@ -86,6 +94,12 @@ pub(crate) async fn run(ctx: &FullCtx) -> color_eyre::Result<()> {
         LabelsActor::new(ctx.clone(), nextcloud_board_id)
     );
 
+    let now = ctx.now();
+    let trello_rate_limiter = spawn_in_thread!(
+        supervisor,
+        RateLimitActor::new(RateLimit::new("trello", 100, 10.0, now))
+    );
+
     let _main = spawn_in_thread!(
         supervisor,
         ImportStacksActor::new(
@@ -94,8 +108,10 @@ pub(crate) async fn run(ctx: &FullCtx) -> color_eyre::Result<()> {
             selected_trello_stack_names.into_iter().cloned().collect(),
             nextcloud_board_id,
             labels_actor_ref.clone(),
+            trello_rate_limiter.clone(),
         )
     );
+    println!("> Spawned main actors");
 
     supervisor.wait_for_stop().await;
 

src/import/rate_limit.rs

@@ -0,0 +1,34 @@
+use crate::{message, rate_limit::RateLimit};
+use kameo::{
+    mailbox::unbounded::UnboundedMailbox,
+    message::{Context, Message},
+    Actor,
+};
+use tokio::time::Instant;
+
+message!(RequestToken, "Request a rate limited token");
+
+pub(crate) struct RateLimitActor {
+    limiter: RateLimit,
+}
+impl RateLimitActor {
+    pub fn new(limiter: RateLimit) -> Self {
+        Self { limiter }
+    }
+}
+impl Actor for RateLimitActor {
+    type Mailbox = UnboundedMailbox<Self>;
+}
+impl Message<RequestToken> for RateLimitActor {
+    type Reply = ();
+
+    async fn handle(
+        &mut self,
+        _msg: RequestToken,
+        _ctx: Context<'_, Self, Self::Reply>,
+    ) -> Self::Reply {
+        self.limiter
+            .block_until_token_available(Instant::now())
+            .await
+    }
+}

src/import/stack.rs

@@ -8,7 +8,10 @@ use kameo::{
     Actor,
 };
 
+use crate::import::rate_limit::{RateLimitActor, RequestToken};
+use crate::rate_limit::RateLimit;
 use crate::{
+    ask,
     import::{card::ImportCardActor, labels::LabelsActor},
     nextcloud::model::{NextcloudBoardId, NextcloudStackId},
     on_actor_link_died, on_actor_start, spawn_in_thread,
@@ -28,6 +31,7 @@ pub(super) struct ImportStackActor {
     nextcloud_board_id: NextcloudBoardId,
     nextcloud_stack_id: NextcloudStackId,
     labels_actor_ref: ActorRef<LabelsActor>,
+    rate_limit_actor: ActorRef<RateLimitActor>,
     children: HashMap<ActorID, (TrelloCardName, ActorRef<ImportCardActor>)>,
 }
 impl ImportStackActor {
@@ -37,6 +41,7 @@ impl ImportStackActor {
         nextcloud_board_id: NextcloudBoardId,
         nextcloud_stack_id: NextcloudStackId,
         labels_actor_ref: ActorRef<LabelsActor>,
+        rate_limit_actor: ActorRef<RateLimitActor>,
     ) -> Self {
         Self {
             ctx,
@@ -44,6 +49,7 @@ impl ImportStackActor {
             nextcloud_board_id,
             nextcloud_stack_id,
             labels_actor_ref,
+            rate_limit_actor,
             children: Default::default(),
         }
     }
@@ -56,6 +62,7 @@ impl Actor for ImportStackActor {
 
         crate::p!(this.ctx.prt, "Importing stack: {}", this.trello_stack.name);
         // - get the list of trello cards in the stack
+        ask!(this.rate_limit_actor, RequestToken)?;
         let mut trello_cards = trello_client
             .list_cards(&TrelloListId::new(this.trello_stack.id.as_ref()))
             .await
@@ -63,8 +70,12 @@ impl Actor for ImportStackActor {
         // sort cards by their position
         trello_cards.sort_by_key(|card| card.pos);
 
+        let mut limit = RateLimit::new("cards", 10, 10.0, this.ctx.now());
+
         // - for each card in the trello stack
         for trello_card in trello_cards.into_iter() {
+            limit.block_until_token_available(this.ctx.now()).await;
+            ask!(this.rate_limit_actor, RequestToken)?;
             let trello_card_name = trello_card.name.clone();
             let child: ActorRef<ImportCardActor> = spawn_in_thread!(
                 actor_ref.clone(),
@@ -74,6 +85,7 @@ impl Actor for ImportStackActor {
                     this.nextcloud_board_id,
                     this.nextcloud_stack_id,
                     this.labels_actor_ref.clone(),
+                    this.rate_limit_actor.clone(),
                 )
             );
             this.children

src/import/stacks.rs

@@ -8,6 +8,8 @@ use kameo::{
     Actor,
 };
 
+use crate::import::rate_limit::RateLimitActor;
+use crate::rate_limit::RateLimit;
 use crate::{
     import::{labels::LabelsActor, stack::ImportStackActor},
     nextcloud::{
@@ -72,6 +74,7 @@ pub(super) struct ImportStacksActor {
     selected_trello_stack_names: Vec<String>,
     nextcloud_board_id: NextcloudBoardId,
     labels_actor_ref: ActorRef<LabelsActor>,
+    rate_limit_actor: ActorRef<RateLimitActor>,
 
     children: HashMap<ActorID, (TrelloListName, ActorRef<ImportStackActor>)>,
 }
@@ -82,6 +85,7 @@ impl ImportStacksActor {
         selected_trello_stack_names: Vec<String>,
         nextcloud_board_id: NextcloudBoardId,
         labels_actor_ref: ActorRef<LabelsActor>,
+        rate_limit_actor: ActorRef<RateLimitActor>,
     ) -> Self {
         Self {
             ctx,
@@ -89,6 +93,7 @@ impl ImportStacksActor {
             selected_trello_stack_names,
             nextcloud_board_id,
             labels_actor_ref,
+            rate_limit_actor,
             children: Default::default(),
         }
     }
@@ -126,8 +131,11 @@ impl Actor for ImportStacksActor {
             .await
             .result?;
 
+        let mut limit = RateLimit::new("stacks", 1, 0.1, this.ctx.now());
+
         // for each selected trello stack
         for selected_trello_stack in selected_trello_stacks {
+            limit.block_until_token_available(this.ctx.now()).await;
             let nextcloud_stack_id = //this.
                 nextcloud_stacks
                 .iter()
@@ -142,6 +150,7 @@ impl Actor for ImportStacksActor {
                     this.nextcloud_board_id,
                     nextcloud_stack_id,
                     this.labels_actor_ref.clone(),
+                    this.rate_limit_actor.clone()
                 )
             );
             this.children

src/lib.rs

@@ -5,10 +5,10 @@ use crate::{nextcloud::client::DeckClient, trello::client::TrelloClient};
 use clap::Parser;
 use color_eyre::eyre::eyre;
 use config::AppConfig;
+use execute::Execute;
 use kxio::fs::TempFileSystem;
 use kxio::{fs::FileSystem, kxeprintln as e, kxprintln as p, net::Net, print::Printer};
-
+use tokio::time::Instant;
-use execute::Execute;
 
 mod api_result;
 mod check;
@@ -19,6 +19,7 @@ mod import;
 mod init;
 pub mod macros;
 mod nextcloud;
+mod rate_limit;
 mod template;
 mod trello;
 
@@ -86,6 +87,13 @@ pub(crate) struct FullCtx {
     pub prt: Printer,
     pub cfg: AppConfig,
 }
+
+impl FullCtx {
+    pub(crate) fn now(&self) -> Instant {
+        Instant::now()
+    }
+}
+
 impl FullCtx {
     pub(crate) const fn deck_client(&self) -> DeckClient {
         DeckClient::new(self)

src/macros/message.rs

@@ -0,0 +1,15 @@
+#[macro_export]
+macro_rules! message {
+    ($name:ident, $value:ty, $docs:literal) => {
+        $crate::newtype!($name, $value, $docs);
+    };
+    ($name:ident, $docs:literal) => {
+        $crate::newtype!($name, $docs);
+    };
+    ($name:ident, $value:ty => $result:ty, $docs:literal) => {
+        $crate::newtype!($name, $value, $docs);
+    };
+    ($name:ident => $result:ty, $docs:literal) => {
+        $crate::newtype!($name, $docs);
+    };
+}

src/macros/mod.rs

@@ -1,6 +1,7 @@
 //
 mod actor;
 mod backoff;
+mod message;
 mod newtype;
 mod print;
 mod send;

src/rate_limit.rs

@@ -0,0 +1,96 @@
+use std::time::Duration;
+use tokio::time::Instant;
+use tracing::instrument;
+
+#[derive(Debug)]
+pub(crate) struct RateLimit {
+    /// The name of the limit
+    pub(crate) name: String,
+
+    /// Last token was consumed at
+    pub(crate) last_consumed_at: Instant,
+
+    /// The number of tokens available after the last was consumed
+    pub(crate) last_token_count: u32,
+
+    /// Maximum tokens we can have at any one time
+    pub(crate) max: u32,
+
+    /// Tokens replenish rate
+    pub(crate) tokens_per_second: f32,
+}
+
+#[derive(Debug, PartialEq, Eq)]
+pub(crate) enum Token {
+    Available,
+    NotAvailable { sleep_nanos: u64 },
+}
+impl RateLimit {
+    // 300 tokens / 10 seconds => max 300, tokens_per_second 30
+    pub(crate) fn new(
+        name: impl Into<String>,
+        max: u32,
+        tokens_per_second: f32,
+        now: Instant,
+    ) -> Self {
+        Self {
+            name: name.into(),
+            last_consumed_at: now,
+            last_token_count: max,
+            max,
+            tokens_per_second,
+        }
+    }
+
+    pub(crate) async fn block_until_token_available(&mut self, now: Instant) {
+        let mut now = now;
+        loop {
+            match self.acquire_token(now) {
+                Token::Available => {
+                    return;
+                }
+                Token::NotAvailable { sleep_nanos } => {
+                    tracing::debug!("sleeping for {}ms", sleep_nanos / 1_000_000);
+                    tokio::time::sleep(Duration::from_nanos(sleep_nanos)).await;
+                    now += Duration::from_nanos(sleep_nanos);
+                }
+            }
+        }
+    }
+
+    #[instrument]
+    pub(crate) fn acquire_token(&mut self, now: Instant) -> Token {
+        debug_assert!(now > self.last_consumed_at, "time must pass");
+        let elapsed_nanos = now.duration_since(self.last_consumed_at).as_nanos();
+        tracing::trace!(?elapsed_nanos, "ELAPSED");
+
+        // Calculate new tokens generated since last consumption
+        let nanos_per_token = (Duration::from_secs(1).as_nanos() as f32) / self.tokens_per_second;
+        let new_tokens = (elapsed_nanos as f32 / nanos_per_token) as u32;
+        let current_tokens = (self.last_token_count + new_tokens).min(self.max);
+        tracing::trace!(%elapsed_nanos, %nanos_per_token, %new_tokens, %current_tokens, "tokens?");
+
+        if current_tokens > 0 {
+            self.last_token_count = current_tokens - 1;
+            self.last_consumed_at = now;
+            tracing::trace!("token available");
+            return Token::Available;
+        }
+
+        // Calculate time until next token is available
+        let time_to_next = if current_tokens == 0 {
+            let tokens_needed = 1;
+            let seconds = (tokens_needed as f64) / (self.tokens_per_second as f64);
+            tracing::trace!(?seconds, "time until next token");
+            Duration::from_secs_f64(seconds)
+        } else {
+            Duration::from_secs(0)
+        };
+
+        tracing::trace!(?time_to_next, "token not available");
+        println!("[{}] sleep {}ms", self.name, time_to_next.as_millis());
+        Token::NotAvailable {
+            sleep_nanos: time_to_next.as_nanos() as u64,
+        }
+    }
+}

src/tests/mod.rs

@@ -13,6 +13,7 @@ mod check;
 mod config;
 pub(crate) mod given;
 mod init;
+mod rate_limit;
 mod template;
 
 #[test]

src/tests/rate_limit.rs

@@ -0,0 +1,141 @@
+//
+use std::ops::Add;
+
+use tokio::time::{Duration, Instant};
+
+use crate::rate_limit::{RateLimit, Token};
+
+#[tokio::test]
+async fn test_initial_token_available() {
+    //given
+    let mut rate_limit = RateLimit::new("Test", 10, 10.0, Instant::now()); // 10 tokens per second, max 10 tokens
+
+    //when
+    let token = rate_limit.acquire_token(Instant::now());
+
+    //then
+    assert_eq!(token, Token::Available);
+}
+
+#[tokio::test]
+async fn test_max_tokens_respected() {
+    //given
+    let start = Instant::now();
+    let mut rate_limit = RateLimit::new("Test", 5, 10.0, start); // 10 tokens per second, max 5 tokens
+
+    //when
+    let token = rate_limit.acquire_token(start.add(Duration::from_secs(1)));
+
+    //then
+    assert_eq!(token, Token::Available);
+    assert_eq!(rate_limit.last_token_count, 4); // Should be max-1 after consuming one token
+}
+
+#[tokio::test]
+async fn test_token_exhaustion() {
+    //given
+    let mut rate_limit = RateLimit::new("Test", 1, 1.0, Instant::now()); // 1 token per second, max 1 token
+
+    // Consume the only available token
+    assert_eq!(rate_limit.acquire_token(Instant::now()), Token::Available);
+
+    //when
+    let token = rate_limit.acquire_token(Instant::now());
+
+    //then
+    // Next request should not be available
+    match token {
+        Token::NotAvailable { sleep_nanos } => {
+            assert!(sleep_nanos > 0);
+        }
+        Token::Available => panic!("Should not have token available"),
+    }
+}
+
+#[test]
+fn add_duration_to_instant() {
+    //given
+    let start = Instant::now();
+    //when
+    let later = start.add(Duration::from_secs(1));
+    //then
+    assert!(later > start);
+    assert!(later.duration_since(start) == Duration::from_secs(1));
+}
+
+#[test_log::test(tokio::test)]
+async fn test_token_regeneration() {
+    //given
+    let mut rate_limit = RateLimit::new("Test", 1, 1.0, Instant::now()); // 1 token per second, max 1 token
+
+    tracing::info!("Consume initial token");
+    assert_eq!(rate_limit.acquire_token(Instant::now()), Token::Available);
+
+    //when
+    let later = Instant::now().add(Duration::from_millis(1100));
+    tracing::info!(later=?later.elapsed().as_nanos(), "Wait for more than 1 second");
+    let token = rate_limit.acquire_token(later); // 1.1 seconds
+
+    //then
+    // Should have generated a new token
+    assert_eq!(token, Token::Available);
+}
+
+#[test_log::test(tokio::test)]
+async fn test_block_until_token_available() {
+    //given
+    let mut rate_limit = RateLimit::new("Test", 1, 2.0, Instant::now()); // 2 tokens per second, max 1 token
+    tracing::info!("consume initial token");
+    rate_limit.block_until_token_available(Instant::now()).await; // Consume initial tokens
+
+    //then
+    tracing::info!("blocking until next token available");
+    let start = Instant::now();
+    tokio::select! {
+        _ = rate_limit.block_until_token_available(Instant::now()) => { // This should block for approximately 0.5 seconds
+            tracing::info!("received token");
+        }
+        _ = tokio::time::sleep(Duration::from_secs(1)) => {
+            panic!("Timeout waiting for token");
+        }
+    }
+    let elapsed = start.elapsed();
+
+    //then
+    assert!(elapsed >= Duration::from_millis(450), "{elapsed:?}")
+}
+
+#[tokio::test]
+async fn test_multiple_concurrent_requests() {
+    let rate_limit = std::sync::Arc::new(tokio::sync::Mutex::new(
+        RateLimit::new("Test", 2, 2.0, Instant::now()), // 2 tokens per second, max 1 token
+    ));
+
+    let mut handles = vec![];
+
+    // Launch 3 concurrent requests
+
+    tokio::select! {
+        _ = {
+           // should take ~0.5 second for third token to become available
+           async{
+               for _ in 0..3 {
+                   let rate_limit_clone = rate_limit.clone();
+                   handles.push(tokio::spawn(async move {
+                       let mut rate_limit = rate_limit_clone.lock().await;
+                       rate_limit.block_until_token_available(Instant::now()).await;
+                   }));
+               }
+               // Wait for all requests to complete
+               for handle in handles {
+                   handle.await.unwrap();
+               }
+           }
+        } => {
+            tracing::info!("concurrent requests okay");
+        }
+        _ = tokio::time::sleep(Duration::from_secs(1)) => {
+            panic!("Timeout waiting for token");
+        }
+    }
+}

src/trello/client.rs

@@ -97,6 +97,18 @@ impl<'ctx> TrelloClient<'ctx> {
         .bytes()
         .await
         .context("downloading attachment")?;
+        let Some(expected_file_size) = attachment.bytes else {
+            crate::e!(self.ctx.prt, "==> no file size in attachment");
+            return Err(color_eyre::eyre::eyre!("no file size in attachment"));
+        };
+        if resp.len() != expected_file_size {
+            crate::e!(
+                self.ctx.prt,
+                "==> expected file size: {} != actual file size: {}",
+                expected_file_size,
+                resp.len()
+            );
+        }
         let file = fs.file(&file_name);
         file.write(resp)
             .map_err(|e| {

src/trello/model/attachment.rs

@@ -23,6 +23,7 @@ newtype!(
 pub(crate) struct TrelloAttachment {
     pub(crate) id: TrelloAttachmentId, // "5abbe4b7ddc1b351ef961414",
     pub(crate) name: TrelloAttachmentName, //"Deprecation Extension Notice",
+    pub(crate) bytes: Option<usize>,
     pub(crate) url: TrelloAttachmentUrl, //"https://admin.typeform.com/form/RzExEM/share#/link",
     #[serde(rename = "fileName")]
     pub(crate) file_name: TrelloAttachmentFilename,