mirror of
https://github.com/lancedb/lancedb.git
synced 2026-10-26 13:36:13 +00:00
fix(node): read each distinct Blob once in resolveBlobInputs
Key the read cache by the Blob itself instead of the column value, so
rows sharing a Blob through separate { data } structs read it once and
share the bytes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012U4KdHofT5edv3LcXHXbLv
This commit is contained in:
@@ -146,6 +146,31 @@ describe("resolveBlobInputs", () => {
|
||||
expect(data[0].images?.[0]).toBe(shared);
|
||||
});
|
||||
|
||||
it("reads each distinct Blob once, bare or inside data structs", async () => {
|
||||
class CountingBlob extends Blob {
|
||||
calls = 0;
|
||||
async arrayBuffer(): Promise<ArrayBuffer> {
|
||||
this.calls++;
|
||||
return super.arrayBuffer();
|
||||
}
|
||||
}
|
||||
const shared = new CountingBlob(["same"]);
|
||||
const schema = new Schema([blob("image")]);
|
||||
const resolved = await resolveBlobInputs(
|
||||
[
|
||||
{ image: { data: shared } },
|
||||
{ image: { data: shared } },
|
||||
{ image: shared },
|
||||
],
|
||||
schema,
|
||||
);
|
||||
expect(shared.calls).toBe(1);
|
||||
const first = (resolved[0].image as { data: Uint8Array }).data;
|
||||
expect(Buffer.from(first).toString()).toBe("same");
|
||||
expect((resolved[1].image as { data: Uint8Array }).data).toBe(first);
|
||||
expect(resolved[2].image).toBe(first);
|
||||
});
|
||||
|
||||
it("returns the input when there is nothing to read", async () => {
|
||||
const data = [{ image: Buffer.from("x") }];
|
||||
const schema = new Schema([blob("image")]);
|
||||
|
||||
Reference in New Issue
Block a user