feat(node): accept an AbortSignal to cancel blob reads (#4350)

Closes #4288.

Blob reads in Node can now be cancelled. `BlobFile.read`, `readRange`
and `readRanges`, and `Table.fetchBlobs` and `fetchBlobFiles` take an
optional `{ signal?: AbortSignal }` (new `BlobReadOptions` type). When
the signal aborts, the call rejects with `signal.reason` (an
`AbortError` unless the caller gave a reason) and the native future is
dropped, so the read stops instead of running to completion. A signal
that is already aborted rejects before any read starts.

How it works:

- On the Rust side, these five methods are now sync `#[napi]` methods
that return `env.spawn_future(...)`. The future is wrapped in
`futures::future::abortable`, and the abort handle is registered with
napi's `AbortSignal::on_abort`. napi's `AbortSignal` is `Rc`-based, so
it can't be an argument of an `async fn`, which is why the methods
changed shape. The TypeScript signatures are unchanged apart from the
new optional argument.
- napi-rs installs its own `onabort` on the signal it receives, which
would overwrite a handler the caller set. So the TypeScript wrapper
passes a private `AbortController` signal to the native call for each
request, forwards the caller's abort to it with `addEventListener(..., {
once: true })`, and removes the listener when the call settles.

Tests in `table.test.ts`:

- an already aborted signal rejects on all five methods, and the handle
still reads afterwards
- a custom abort reason is what the call rejects with
- a read with a signal that never aborts returns normally and leaves no
`abort` listener on the signal
- aborting a 32 MiB `readRange` right after it starts rejects with
`AbortError`, and the same handle reads again afterwards

The Node docs under `docs/src/js` are regenerated.

I didn't add a test for where the `read()` cursor ends up after a
mid-flight abort on a local table. The remote `BlobFile` documents that
a cancelled read leaves the cursor unchanged, but I haven't confirmed
the same for the native Lance handle, so the docs don't claim it.
This commit is contained in:
Bilal Atique
2026-09-29 14:01:17 -05:00
committed by GitHub
parent ad79112b8e
commit 0a1c1a04c7
10 changed files with 293 additions and 64 deletions
+75 -27
View File
@@ -1,12 +1,15 @@
// SPDX-License-Identifier: Apache-2.0
// SPDX-FileCopyrightText: Copyright The LanceDB Authors
use std::future::Future;
use std::ops::Range;
use std::sync::Arc;
use std::sync::atomic::{AtomicBool, Ordering};
use arrow_array::{Array, LargeBinaryArray};
use futures::future::abortable;
use lancedb::blob::BlobFile as LanceBlobFile;
use napi::Env;
use napi::bindgen_prelude::*;
use napi_derive::napi;
@@ -49,48 +52,73 @@ impl BlobFile {
}
#[napi]
pub async fn read(&self, max_bytes: Option<BigInt>) -> napi::Result<Buffer> {
pub fn read<'env>(
&self,
env: &'env Env,
max_bytes: Option<BigInt>,
signal: Option<AbortSignal>,
) -> napi::Result<PromiseRaw<'env, Buffer>> {
self.ensure_open()?;
let bytes = match max_bytes {
None => self.inner.read().await,
Some(max_bytes) => {
let max_bytes = usize::try_from(parse_u64(max_bytes, "maxBytes")?)
.map_err(|_| napi::Error::from_reason("maxBytes is too large"))?;
self.inner.read_up_to(max_bytes).await
let max_bytes = max_bytes
.map(|max_bytes| {
usize::try_from(parse_u64(max_bytes, "maxBytes")?)
.map_err(|_| napi::Error::from_reason("maxBytes is too large"))
})
.transpose()?;
let inner = self.inner.clone();
spawn_abortable(env, signal, async move {
let bytes = match max_bytes {
None => inner.read().await,
Some(max_bytes) => inner.read_up_to(max_bytes).await,
}
}
.map_err(|err| convert_error(&err))?;
Ok(Buffer::from(bytes.as_ref()))
.map_err(|err| convert_error(&err))?;
Ok(Buffer::from(bytes.as_ref()))
})
}
#[napi]
pub async fn read_range(&self, start: BigInt, end: BigInt) -> napi::Result<Buffer> {
pub fn read_range<'env>(
&self,
env: &'env Env,
start: BigInt,
end: BigInt,
signal: Option<AbortSignal>,
) -> napi::Result<PromiseRaw<'env, Buffer>> {
self.ensure_open()?;
let range = bigint_range(start, end)?;
let bytes = self
.inner
.read_range(range)
.await
.map_err(|err| convert_error(&err))?;
Ok(Buffer::from(bytes.as_ref()))
let inner = self.inner.clone();
spawn_abortable(env, signal, async move {
let bytes = inner
.read_range(range)
.await
.map_err(|err| convert_error(&err))?;
Ok(Buffer::from(bytes.as_ref()))
})
}
#[napi]
pub async fn read_ranges(&self, ranges: Vec<BlobRange>) -> napi::Result<Vec<Buffer>> {
pub fn read_ranges<'env>(
&self,
env: &'env Env,
ranges: Vec<BlobRange>,
signal: Option<AbortSignal>,
) -> napi::Result<PromiseRaw<'env, Vec<Buffer>>> {
self.ensure_open()?;
let ranges = ranges
.into_iter()
.map(|range| bigint_range(range.start, range.end))
.collect::<napi::Result<Vec<_>>>()?;
let buffers = self
.inner
.read_ranges(&ranges)
.await
.map_err(|err| convert_error(&err))?;
Ok(buffers
.iter()
.map(|bytes| Buffer::from(bytes.as_ref()))
.collect())
let inner = self.inner.clone();
spawn_abortable(env, signal, async move {
let buffers = inner
.read_ranges(&ranges)
.await
.map_err(|err| convert_error(&err))?;
Ok(buffers
.iter()
.map(|bytes| Buffer::from(bytes.as_ref()))
.collect())
})
}
#[napi]
@@ -124,6 +152,26 @@ impl BlobFile {
}
}
/// Runs `fut` on the napi runtime. Aborting `signal` drops the future, which
/// cancels in-flight reads.
pub(crate) fn spawn_abortable<'env, T>(
env: &'env Env,
signal: Option<AbortSignal>,
fut: impl Future<Output = napi::Result<T>> + Send + 'static,
) -> napi::Result<PromiseRaw<'env, T>>
where
T: ToNapiValue + Send + 'static,
{
let (fut, handle) = abortable(fut);
if let Some(signal) = signal {
signal.on_abort(move || handle.abort());
}
env.spawn_future(async move {
fut.await
.unwrap_or_else(|_| Err(napi::Error::from_reason("blob read aborted")))
})
}
fn bigint_range(start: BigInt, end: BigInt) -> napi::Result<Range<u64>> {
let start = parse_u64(start, "start")?;
let end = parse_u64(end, "end")?;
+29 -20
View File
@@ -11,11 +11,12 @@ use lancedb::table::{
FieldMetadataUpdate as LanceFieldMetadataUpdate, FtsToken as LanceDbFtsToken,
NewColumnTransform, OptimizeAction, OptimizeOptions, Ref, Table as LanceDbTable,
};
use napi::Env;
use napi::bindgen_prelude::*;
use napi::threadsafe_function::{ThreadsafeFunction, ThreadsafeFunctionCallMode};
use napi_derive::napi;
use crate::blob::{BlobFile, copy_blob_buffers, parse_row_ids};
use crate::blob::{BlobFile, copy_blob_buffers, parse_row_ids, spawn_abortable};
use crate::error::NapiErrorExt;
use crate::index::Index;
use crate::merge::NativeMergeInsertBuilder;
@@ -340,36 +341,44 @@ impl Table {
}
#[napi(catch_unwind)]
pub async fn fetch_blobs(
pub fn fetch_blobs<'env>(
&self,
env: &'env Env,
column: String,
row_ids: Vec<BigInt>,
) -> napi::Result<Vec<Option<Buffer>>> {
signal: Option<AbortSignal>,
) -> napi::Result<PromiseRaw<'env, Vec<Option<Buffer>>>> {
let row_ids = parse_row_ids(row_ids)?;
let array = self
.inner_ref()?
.fetch_blobs(column.as_str(), &row_ids)
.await
.default_error()?;
Ok(copy_blob_buffers(array))
let table = self.inner_ref()?.clone();
spawn_abortable(env, signal, async move {
let array = table
.fetch_blobs(column.as_str(), &row_ids)
.await
.default_error()?;
Ok(copy_blob_buffers(array))
})
}
#[napi(catch_unwind)]
pub async fn fetch_blob_files(
pub fn fetch_blob_files<'env>(
&self,
env: &'env Env,
column: String,
row_ids: Vec<BigInt>,
) -> napi::Result<Vec<Option<BlobFile>>> {
signal: Option<AbortSignal>,
) -> napi::Result<PromiseRaw<'env, Vec<Option<BlobFile>>>> {
let row_ids = parse_row_ids(row_ids)?;
let files = self
.inner_ref()?
.fetch_blob_files(column.as_str(), &row_ids)
.await
.default_error()?;
Ok(files
.into_iter()
.map(|file| file.map(BlobFile::new))
.collect())
let table = self.inner_ref()?.clone();
spawn_abortable(env, signal, async move {
let files = table
.fetch_blob_files(column.as_str(), &row_ids)
.await
.default_error()?;
Ok(files
.into_iter()
.map(|file| file.map(BlobFile::new))
.collect())
})
}
#[napi(catch_unwind)]