@zbsearch/plugin-data-persistence
v4.0.0
Published
Data persistence plugin for ZBSearch
Maintainers
michelerivaReadme
Data Persistence Plugin
This plugin aims to provide data persistence capabilities to ZBSearch.
Usage
For the complete usage guide, please refer to the official plugin documentation.
Storage adapters
Besides serializing to a string/buffer (persist/restore) or to disk
(persistToFile/restoreFromFile, via /server), you can persist a snapshot
directly into a pluggable storage backend with persistToStorage /
restoreFromStorage. The whole index is written as one compact byte payload;
the in-memory database is untouched and stays fully queryable, and restoring it
loads the snapshot back into memory.
Any backend implementing the PersistenceStorage contract works. It is
structurally compatible with ObjectStorage from @zbsearch/edge-core, so an
S3ObjectStorage can be passed as-is.
import { persistToStorage, restoreFromStorage } from '@zbsearch/plugin-data-persistence'
import { IndexedDBStorage } from '@zbsearch/plugin-data-persistence/indexeddb'
const storage = new IndexedDBStorage() // durable, in-browser
await persistToStorage(db, storage, 'my-index') // defaults to compact `binary`
// later — e.g. after a page reload:
const db = await restoreFromStorage(storage, 'my-index')persistToStorage(db, storage, key, { format }) and
restoreFromStorage(storage, key, { format }) accept the same formats as
persist (json | dpack | binary | seqproto); the format used to
restore must match the one used to persist. For IndexedDB, binary (raw
msgpack, no hex doubling) or json are recommended.