visualeyes / cabinet Goto Github PK

Methods to check the space remaining total space will enable us to be smart about replication. For example, we don't want to replicate files and cause the disk to completely fill up.

• IFileCabinet - long? GetStorageSpace() // size of underlying store - null would indicated unlimited
• IFIleCabinet - long GetSize() // total size of files in cabinet
• IFileCabinet - long? GetAvailableSize() // space available in underlying store - null would indicate unlimited

Clients should be able to configure a key prefix for the Amazon S3 provider. All operations would then be based on keyPrefix + key

Create a project with an Google Cloud Storage provider to store files in the Google Cloud.

Add an option to create a S3 bucket if it does not exist. This will need to consider what permissions should be set on the bucket.

Hi,

First, this is awesome, looks like just what I need, however I noticed a couple of things which I need which are missing and I wanted to check if you'd considered them

• Directory Management (Create/Move/Copy).
  I realize on key/value stores like S3 there is no such thing as a directory, but I need a way to create a browsable directory hierarchy, no matter what storage provider.
• Ability to get the url to a file.
  For the filesystem provider this doesn't make a whole lot of sense but for S3 et al it does. Maybe the filesystem provider could return the full path? Or you can provide a config setting which specifies a base url?
• Azure support.
  I see you have an old in-progress PR - are you continuing with that or is that somewhat abandoned?

Anyway, great work, thanks!

Provide support from migrating from one FileCabinet to another. For example, on disk to s3 or on disk on one location to another location.

This will require:

• A IFileCabinet that lazily migrates files from one File Cabinet to another.
• A replication task to move files from one provider to another.
• Leave old provider as is during migration (safer) - ability to delete files from old provider once safe to do so.

The UploadKeyProvider current uses Path.Combine to join the guid and filename into a key. This causes issues for non-filesystem providers that may have a different delimiter. It should require the delimiter and use KeyUtils.JoinKeys

The coverage needs to be increased to increase confidence in changes.

It's important to have the file information in the upload save result to ensure the client can display information about uploaded files.

Create a project with an Azure provider to store files in the Azure blob store.

Allows developers to easily see the size of items they are about to operate on.

• Create interface
• Add optional total bytes to interface
• Add percentage done extensions
• add current key to progress

Create a MultiPartStreamProvider that can be configured with a file Cabinet. See http://www.azurefromthetrenches.com/uploading-an-image-to-a-blob-container-via-web-api/.

This will allow clients to stream files directly to a provider S3 / Azure without hitting the file system.

The migration task runner executes all the task but never completes. This is likely due to the use of the TransformBlock rather than an ActionBlock.

When checking file extensions jpg is allowed and JPG is denied. Cabinet should ignore case when checking file extensions.

Convert to .NET Core to support .NET Core and .NET 4.x frameworks

To allow easier configuration create providers that can load config from JSON stored on disk.

Example config file.

{
    "ondisk": {
        "type": "FileSystem",
        "config": {
            "dir": "~/App_Data/Uploads",
            "createIfNotExists": true
        }
    },
    "amazon": {
        "type": "AmazonS3",
        "config": {
            "credentials": {
                "type":  "profile", // environment, environment-variable, basic
                "name": "some-name"
            },
            "region": "ap-southeast-2",
            "bucket": "test-bucket"
        }
    }
}