npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

lmdb-indexeddb

v0.0.9

Published

LMDB API wrapped around IndexedDB to make LMDB available in the browser

Downloads

11

Readme

lmdb-indexeddb

LMDB API wrapped around IndexedDB to make LMDB available in the browser.

This is ALPHA software. Most unit tests are in place, but edge and error generating cases remain to be written. The software has not been heavily used. Replication in the context of transactions has not been tested. Load and performance testing has not been done. Encryption and compression compatibility with the NodeJS version has not been tested.

See also:

LMDB Cluster - A clustered version of LMDB that supports a REST API with sockets planned.

LMDB Query - A higher level query mechanism for LMDB supporting functional, declarative and RegExp filters without the overhead of an entire database wrapper.

Installation

npm install lmdb-indexeddb

Usage

See lmdb for documentation on the LMDB API.


```javascript
import {open} from "lmdb-indexeddb"

Examples

See the directory examples for examples.

Kitchen Sink

<script type="module">
    import {open} from "../lmdb-indexeddb.js";
    const db1 = await open("dummy",{useVersions:true,compression:true,cache:true}), //encryptionKey:"my password"
            db2 = await db1.openDB("child",{useVersions:true,encryptionKey:"my password",cache:true}),
            db3 = await db1.openDB("child2",{useVersions:true,encryptionKey:"my password",cache:true});
    console.log(db1.get("test1"));
    console.log(db2.get("test2"));
    await db1.put("test1","test1test1");
    await db2.put("test2","test2");
    console.log(db1.get("test1"));
    console.log(db2.get("test2"));
    console.log(await db1.getAsync("test1",true));
    console.log(await db2.getAsync("test2",true));
    // await db1.remove("test1");
    // await db2.remove("test2");
    console.log(db1.get("test1"));
    console.log(db2.get("test2"));
    await db2.transaction(() => {
        db1.put("test1a","test1a");
        db2.put("test2a","test2a");
    });
    console.log(db1.get("test1a"));
    console.log(db2.get("test2a"));
    console.log(await db1.getAsync("test1",true));
    console.log(await db2.getAsync("test2",true));
    try {
        await db2.transaction(() => {
            db1.put("test1b","test1b");
            db2.put("test2b","test2b");
            throw new Error("test error");
        });
    } catch (e) {
        console.log(e);
    }
    console.log(db1.get("test1b"));
    console.log(db2.get("test2b"));
    console.log(await db1.getAsync("test1",true));
    console.log(await db2.getAsync("test2",true));
    try {
        await db1.transaction(() => {
            db1.remove("test1a");
            db2.remove("test2a");
            throw new Error("test error");
        });
    } catch(e) {
        console.log(e);
    }
    console.log(db1.get("test1a"));
    console.log(db2.get("test2a"));
    console.log(await db1.getAsync("test1",true));
    console.log(await db2.getAsync("test2",true));
    try {
        await db1.transaction(() => {
            db1.remove("test1a");
            db2.remove("test2a");
        });
    } catch(e) {
        console.log(e);
    }
    console.log(db1.get("test1a"));
    console.log(db2.get("test2a"));
    console.log(await db1.getAsync("test1",true));
    console.log(await db2.getAsync("test2",true));
    await db3.put("test", {name:"test"});
    console.log(db3.get("test"));
    db3.clearSync();
    console.log(db3.get("test"));
    console.log(db3.getAsync("test",true));
    await db3.drop();
</script>

Implementation Notes

API Calls

All API calls have the same sync or async signature as LMDB except open and opendDB, which are asynchronous only due to the nature of IndexedDB. The open option cache is always true. The options encryptionKey, compression, and useVersions are supported.

In order to support a synchronous API, the IndexedDB API is wrapped in a synchronous API that uses an in memory cache of all entries in the local database. This cache is updated on every put and delete call. This means that the cache is always up-to-date unless the database is modified outside of the API.

When an LMDB transaction is started, it is done at a global level. All put and delete calls operate solely against a temporary cache until the user provided callback completes. Then, an IndexedDB transaction is started and all changes are written to IndexedDB. If the user provided callback throws an error or returns ABORT, the transaction is aborted and the primary cache and IndexedDB are not updated. Child transactions are not currently supported.

When synchronous versions of async functions that result in asynchronous IndexedDB activity are called, the IndexedDB functions are invoked but not tracked with the exception of errors, which are logged to the console. As a result, the use of putSync,removeSync and clearSync is DISCOURAGED and warnings are written to the console. The use of the sync methods can be particularly confusing since they will immediately update the local cache and database, but if an error is thrown during the async processing the values will be reverted. The method get, which is synchronous, is safe to use since it only access the cache (all data in a local IndexedDB is loaded when the databases is opened, no IndexedDB activity occurs when get is called).

A non-standard method getEntryAsyc is provided to allow the user to access an IndexedDB entry after opening. The method takes a key and an optional second argument force which, if true, will force the IndexedDB entry to be loaded into the cache. This is useful for debugging or if the database is updated via an alternate mechanism. To load multiple entries you can also use the standard method prefetch.

In addition to version and value, an entry object in the database may have the boolean properties compressed and encrypted and a timestamp mtime. For consistency, these are not returned to the calling code, but they are passed to replication functions. Because these properties are stored, compression and encryption are not all or nothing propositions like they are with the NodeJS version of LMDB. You can have some entries compressed and some not, and the same for encryption. They will be automatically detected and handled appropriately. Note, if you do open a database without an encryption key any entries that are encrypted will have an Int32Array in their value property. Finally, when useVersions is set to false, the version property will not be present in the entry object when returned; however, in the datastore it will be version 1 when the entry is first created.

The method resetReadTxnAsync will reset the cache of a database and load all data from IndexedDB. This is useful if the database is updated via an alternate mechanism. The method resetReadTxnis not implemented and will throw an error.

Encryption and Compression

Theoretically, the same encryption and compression algorithms are used as in the NodeJS version of LMDB. However, compatibility has not yet been tested.

Synchronizing With Remote Databases

Thereplicate method that takes putRawEntry, getRawEntryAsync, getKeys, getRange, remove functions as options that either get values from a remote database or put and remove them from a remote database. These functions are implemented by a library user, typically with HTTPS. Support for functions like clear and drop are not currently planned.

Version numbers and an mtime modification time stamp are used to resolve conflicts.

To avoid conflicts with the remote database, if mtime of an entry on the remote server is greater than the machine time on which lmdb-indexeddb is running, the change is deferred with a timeout until the local time is greater than or equal to the remote entry time. This means that getEntryAsync, getAsync, put, and remove all do a call to getRawEntryAsync to uncover the mtime of the current value for the entry (if any) on the remote machine. Hence, they will not return until a server request has been made and responded to. It is assumed that remote server time is more accurate than local time. If there is gap larger than 30 seconds, an error is thrown.

In the highly unlikely case the mtime for a remote entry is identical to the time for the local action, then the action is deferred one JavaScript event cycle. If the mtime is identical to the local action time again, an error is thrown. Otherwise, these rules are applied:

  • For a get action, if useVersions is false the entry with the highest mtime is returned; otherwise, the entry with the highest version number is returned, and if versions are the same, the entry with the highest mtime is returned. In ties, the entry with the highest lexicographic value is returned. If necessary, the local cache and IndexedDB are updated to reflect the remote entry and a request is made to update the remote entry to match the local entry. Depending on its own conflict resolution or access control rules, the server may or may not update its version. If it rejects the update, then the get returns the server version, the local cache and indexedDB are updated and if inside a transaction, the transaction will be aborted.
  • For a put action, if useVersions is false the entry with the highest mtime is kept; otherwise the entry with the highest version number is kept, and if versions are the same, the entry with the highest mtime is kept. In ties, the entry with the highest lexicographic value is kept. If necessary, a request of the server is made to update the remote entry. Depending on its own conflict resolution or access control rules, the server may or may not update its version. If it refuses to make a requested update and the local version was to be used, the put action will return false and if inside a transaction, the transaction will be aborted. If the server version is to be used locally, then the put action will return false and if inside a transaction, the transaction will be aborted since the put entry is not the one that was requested.
  • For a remove action, if useVersions is false and the mtime on the client entry is greater than or equal to that of the server or useVersions is true and the local version number is greater than or equal to the server version, a request is made to delete the version on the server. Depending on its own conflict resolution or access control rules, the server may or may not delete its version. If is refuses to delete its version or the server version is newer than the local version, the delete action will return false and if inside a transaction, the transaction will be aborted.

Note: If the server fails to return version and mtime, the local activity will always result in an update request to the server.

Note: If you are implementing a server, you should also be checking mtime to make sure that the mtime sent by a client is not in the future. If it is, you should use a timeout to defer changes and response. If the mtime is in the past relative to the mtime of an impacted entry, you can simply ignore the request.

Future Enhancements

Web sockets for push and pull replication. When a push is made from the server, the in memory cache will be updated accordingly. In order to avoid overload, the client will only be pushed values for keys on which it has done a get or getRange (effectively a subscribe) and will only get new keys that are within the range it has requested using getKeys.

Updates (Reverse Chronological Order)

2023-04-13 v0.0.9 Ehanced documentation.

2023-04-07 v0.0.8 Documentation updates.

2023-04-05 v0.0.7 LMDB treats put("hello","world") and put(["hello"],"world") at the same key, i.e. "hello". Made this library consistent. Simplified core code for remove`. Added more unit tests for replication and adjusted code accordingly.

2023-04-04 v0.0.6 Basic replication for remove working.

2023-04-03 v0.0.5 Basic replication for put and get working.

2023-04-03 v0.0.4 Unit tests for transactions, compression, encryption, and versions. Fixed issue with conditional versioning not working. Enhanced implementation notes.

2023-04-02 v0.0.3 Unit tests for keys and entries. Fixed issue with serializing symbols.

2023-04-01 v0.0.2 Unit tests for open, put, putSync, get, remove, removeSync, clearSync, clearAsync, drop, dropSync, transaction.

2023-03-29 v0.0.1 Initial release. open, put, get, remove, transaction all at basic level of functionality