SUPERKV - Supercharge Clouflare KeyValue store (experimental)
Store Cloudflare KV with more control and power.
Features
- Create complex queries within KV :) - KVTable
- Upload KeyValue bigger the 25MB limit - KVBig
- Manage namespaces (create, rename, delete) - KVNamespaceApi
- Interact with CF KV API v4 - KVApi
- Insert / delete KeyValue in bulk - KVBatch
- You can use it outside of workers and manage KV in your own server.
- Trying to be subrequest efficient (always using bulk insert for prefix keys instead of
kv.put
...) avoid hitting the 50 subrequest to fast.
Why / Limits / Warning
- I made this because I simply wanted to have more control over the KV storage.
- If you have ideas or you know how to optimize/fix things please help :)
- KV uses ascending lexicographical order use ULID to sort descending https://www.npmjs.com/package/ulidx
- I have no prior knowledge in making a tool for data management and this is the best I could come up with.
How to use it
For more example check test
folder
Simple usage
import { KVNamespaceApi } from 'superkv'
import { ulid } from 'ulid'
interface User {
username: string
points: number
active: boolean
timestamp: number
}
const userTableDefinition = {
name: 'users',
properties: ['username', 'points', 'active', 'timestamp']
prefix: {
'username': {
keyValue: (metadata) => metadata.username
},
'>500points': {
filter: (metadata) => metadata.points > 500,
},
'isActive': {
sortValue: (metadata) => metadata.active ? 'true' : 'false',
},
'timestamp_desc': {
sortValue: (metadata) => ulid(timestamp),
},
}
} as KVTableDefinition<User>
const main = async () => {
const accountId = '', authToken = ''
const kvNamespaceApi = new KVNamespaceApi({ accountId }, { authToken })
const namespace = await kvNamespaceApi.createOrGetNamespace('worker_namespace')
const kvApi = kvNamespaceApi.useKVApi()
const kvUsers = kvApi.useKVTable(userTableDefinition)
// Get user by key
const user = await kvUsers.getMetadata('{key}')
// Get user by username
const user = await kvUsers.getMetadata('{username}'. 'username')
// Get users with lease 500 points
const prefix = kvUsers.createPrefixKey('>500points')
const list = await kvUsers.list({ prefix })
// Get active users
const prefix = kvUsers.createPrefixKey('isActive', 'true')
const list = await kvUsers.list({ prefix })
// Get latest users
const prefix = kvUsers.createPrefixKey('timestamp_desc')
const list = await kvUsers.list({ prefix })
}
main()
KVNamespaceApi
Initialize
import { KVNamespaceApi } from 'superkv'
const accountId = '', authToken = ''
const kvNamespaceApi = new KVNamespaceApi({ accountId }, { authToken })
Create a new namespace
const res = await kvNamespaceApi.createNamespace('{title}')
const namespace = res.result
// { id, title, supports_url_encoding }
Rename namespace
const res = await kvNamespaceApi.renameNamespace('{namespaceId}', '{title}')
Remove namespace
const res = await kvNamespaceApi.removeNamespace('{namespaceId}')
List namespaces
const res = await kvNamespaceApi.listNamespaces()
const res = await kvNamespaceApi.listNamespaces({ direction: 'desc', order: 'id', page: 0, per_page: 1 })
Create or get namespace
const namespace = await kvNamespaceApi.createOrGetNamespace('{namespaceTitle}')
Reset or get namespace
I usually use this for testing and want to purge data every run
const namespace = await kvNamespaceApi.resetAndGetNamespace('{namespaceTitle}')
KVApi
Initialize
const namespace = await kvNamespaceApi.createOrGetNamespace('{namespaceTitle}')
const kvApi = kvNamespaceApi.useKVApi(namespace.id)
// or
const kvApi = new KVApi({ kvNamespaceApi, namespaceId: namespace.id })
Read keyValue
const res = await kvApi.readKeyValuePair('{key}')
const value = res.result
Write keyValue
const res = await kvApi.writeKeyValuePair('{key}', '{value}')
const res = await kvApi.writeKeyValuePair('{key}', '{value}', metadata, { expiration: 1578435000, expiration_ttl: 300 })
Write multiple keyValue
const res = await kvApi.writeMultipleKeyValuePairs([
{ key: '{key}', value: '{value}' }
{ key: '{key}', value: '{value}', metadata: {} },
{ key: '{key}', value: '{value}', metadata: {}, expiration: 1578435000, expiration_ttl: 300 }
])
Delete keyValue
const res = await kvApi.deleteKeyValuePair('{key}')
Delete multiple keyValue
const res = await kvApi.deleteMultipleKeyValuePairs(['{key}', '{key}'])
List keyValues
const res = await kvApi.listKeys({ prefix: '', limit: 10, cursor: '' })
const result = res.result
const result_info = res.result_info
KVTable
Initialize
const tableDefinition = { name, properties, prefix, ... }
const kvUser = kvApi.useKVTable<Metadata>(tableDefinition)
const kvUser = new KVTable<Metadata>({ kvApi, tableDefinition })
Set user
const user = { username: 'fred', emai: 'fred@mail.com', ... }
await kvUsers.set('{key}', user)
Del user
await kvUsers.del('{key}')
Get user
await kvUsers.del('{key}')
List users
const list = await kvUsers.list()
const result = list.result
const cursor = list.cursor
const prefix = kvUsers.createPrefixKey('{prefixName}')
const list = await kvUsers.list({ prefix })
Iterator
const prefix = kvUsers.createPrefixKey('{prefixName}')
const it = kvUsers.iterator(prefix)
let result = await it.next()
while (!result.done) {
const list = result.value
}
Update prefix
await kvUsers.updatePrefix()
Delete prefix
await kvUsers.delPrefix('{prefixName}')
KVBatch
Initialize
const kvBatch = kvApi.useKVBig()
const kvBatch = new KVBig({ kvApi })
Set keyValue in bulk
for (const user of users) {
await kvBatch.set('{key}', metadata, '{value}')
}
await kvBatch.finish()
Delete keyValue in bulk
for (const user of users) {
await kvBatch.del('{key}')
}
await kvBatch.finish()
KVBig
Initialize
const kvBig = kvApi.useKVBig()
const kvBig = new KVBig({ kvApi })
Write kv file from readStream
import { open } from 'fs/promises'
const fd = await open('./data.txt', 'r')
await kvBig.set('{key}', fd.createReadStream())
Get KV file to writeStream
import { open } from 'fs/promises'
const fd = await open('./data.txt', 'a')
await kvBig.get('{key}', fd.createWriteStream())
Previous package
https://github.com/g45t345rt/cf-kvprefix
cf-kvprefix
is deprecated in a favor of superkv
- Only works in workers
- Does not use any bulk operations
- Can't update or add new prefix
- Is not subrequest efficient
- Can't upload KeyValue bigger than 25MB