Class LayeredBlobStore

Read-through / write-through layering: writes go to local and remote in parallel; reads check local first and fall back to remote, caching the result locally on hit.

This is what we recommend for browser apps: instant local reads, durable remote storage, and offline-first semantics for already-fetched blobs.

Implements

BlobStore

Index

Constructors

Methods

delete get has hashBytes list put

Constructors

constructor

new LayeredBlobStore(local: BlobStore, remote: BlobStore): LayeredBlobStore
Parameters
- local: BlobStore
- remote: BlobStore
Returns LayeredBlobStore

Methods

delete

delete(hash: string): Promise<boolean>
Parameters
- hash: string
Returns Promise<boolean>

Implementation of BlobStore.delete

get

get(hash: string): Promise<Uint8Array<ArrayBufferLike> | null>
Parameters
- hash: string
Returns Promise<Uint8Array<ArrayBufferLike> | null>

Implementation of BlobStore.get

has

has(hash: string): Promise<boolean>
Parameters
- hash: string
Returns Promise<boolean>

Implementation of BlobStore.has

hashBytes

hashBytes(bytes: Uint8Array): string
Recompute this backend's content identity without writing the blob.
Parameters
- bytes: Uint8Array
Returns string

Implementation of BlobStore.hashBytes

list

list(): Promise<string[]>

List currently-known blob hashes; ordering is implementation-defined.

Returns Promise<string[]>

Implementation of BlobStore.list

put

put(
    bytes: Uint8Array,
    contentType?: string,
    options?: BlobPutOptions,
): Promise<BlobMeta>
Parameters
- bytes: Uint8Array
- OptionalcontentType: string
- Optionaloptions: BlobPutOptions
Returns Promise<BlobMeta>

Implementation of BlobStore.put