Skip to content

Latest commit

 

History

History
192 lines (133 loc) · 6.11 KB

migration_to_api2.1.md

File metadata and controls

192 lines (133 loc) · 6.11 KB
Raw

  • Every API method that accesses the cloud storage service returns a Promise that resolves to an object:

    type ResultObject = {
  error: string | null;
  value: string | number | Array<[string, number]> | Array<String> | Readable; // depends on method
};

  • The storage instance will also no longer hold a reference to all available buckets; a call to listBuckets will access the cloud storage service every time it is called; this is handy in case another process or user has created or deleted a new bucket.

  • createBucket resolves with an error if that bucket already exists

  • removeFile has an additional optional boolean argument allVersions; if set to true all versions of the specified file will be removed. Default: false

  • addFile is added to the API; you can use this method instead of addFileFromPath, addFileFromBuffer or addFileFromReadable

  • Extended and updated the introspect API of the adapter:

    • getConfig() and getType() are implemented as getter as well, resp.: storage.config and storage.type
    • added configError and storage.configError
    • added getServiceClient and storage.serviceClient
    • added getter and setter for (de)selecting the bucket; storage.bucket = "the-buck" and console.log(storage.bucket)

  • Supported storages:

    • Amazon S3
    • Cubbit
    • Cloudflare R2
    • Backblaze B2 S3 compatible
    • Backblaze B2
    • Google Cloud
    • Azure Blob
    • MinIO
    • local storage

Configuration URLs

The format of the configuration URLs is more consistent across the storage services. The @region part of the config URL of the AmazonS3 adapter has been moved to the query string.

// old format:
const u = "s3://accessKeyId:secretAccessKey@region/bucketName?option1=value1&option2=value2...";

// new format:
const v = "s3://accessKeyId:secretAccessKey@bucketName?region=us-west-1&option2=value2...";

The general format is:

const u = "protocol://part1:part2@bucketName?option1=value1&option2=value2...";

Which translates for most storage services into:

const u = "storage_type://credential1:credential2@bucketName?option1=value1&option2=value2...";

Old API (1.5.x) compared to new API (2.x)

init

init(config):Promise<boolean>
N/A

test

test():Promise<string>
N/A

selectBucket

selectBucket(name: string | null): Promise<string>
selectBucket(name: string | null): void

getSelectedBucket

getSelectedBucket(): string
getSelectedBucket(): string | null

validateName

validateName(name: string): string
N/A

listBuckets

listBuckets(): Promise<string[]>
listBuckets(): Promise<ResultObjectBuckets>

createBucket

createBucket(name: string, options?: object): Promise<string>
createBucket(name: string, options?: object): Promise<ResultObject>

clearBucket

clearBucket(name?: string): Promise<string>
clearBucket(name: string): Promise<ResultObject>

deleteBucket

deleteBucket(name?: string): Promise<string>
deleteBucket(name: string): Promise<ResultObject>

removeFile

removeFile(fileName: string): Promise<string>
removeFile(bucketName: string, fileName: string): Promise<ResultObject>

listFiles

listFiles(): Promise<[string, number][]>
listFiles(bucketName: string): Promise<ResultObjectFiles>

sizeOf

sizeOf(name: string): Promise<number>
sizeOf(bucketName: string, fileName: string): Promise<ResultObject>

fileExists

fileExists(name: string): Promise<boolean>
fileExists(bucketName: string, fileName: string): Promise<ResultObjectBoolean>

getFileAsReadable

getFileAsReadable(
    name: string,
    options?: { start?: number; end?: number }
  ): Promise<Readable>
getFileAsStream(
    bucketName: string,
    fileName: string,
    options?: { [id: string]: any }
  ): Promise<ResultObjectStream>

addFileFromPath

addFileFromPath(origPath: string, targetPath: string, options: object = {}): Promise<string>
addFileFromPath(params: FilePathParams): Promise<ResultObject>

addFileFromBuffer

addFileFromBuffer(buffer: Buffer, targetPath: string, options: object = {}): Promise<string>
addFileFromBuffer(params: FileBufferParams): Promise<ResultObject>

addFileFromReadable

addFileFromReadable(stream: Readable, targetPath: string, options: object = {}): Promise<string>
addFileFromStream(FileStreamParams): Promise<ResultObject>

The init function is not required anymore

Only Backblaze B2 Native API storage requires initial authorization by calling the async authorize function. This authorization step was performed once by calling the init method. Although it would yield an error, it was still possible to call API methods without calling init prior to that. In the new version every API call checks if the initial authorization has been performed.

Other storage services do not require initial authorization but their init method was used to select and/or create the bucket that was provided in the config.

Because in the new API seeks to be more transparent, there will be no more 'magic behind the screen'. So if you want to create a bucket (provided you have the access rights to do so) you have to call createBucket explicitly.

Also the new version tries to keep as little local state as possible so selectBucket and getSelectedBucket have been removed.

Because of all aforementioned changes the init is no longer required! You can start calling API methods right after instantiating a storage:

const b2 = new Storage("b2://applicationKeyId=your-key-id&applicationKey=your-key");
await b2.listBuckets();

The bucket in the config is no longer automatically selected or created

However, the bucket name that you've provided with the configuration url or object is available by calling getConfig:

const  s3 = new Storage("s3://key=key&secret=secret&region=eu-west-2&bucketName=erwe");
await s3.listFiles(s3.config.bucketName, "your-file.jpg')