bloc-0.1.0.0: Binary Large Objects with Concurrency

StabilityExperimental
Portabilitynon-portable (requires POSIX)
Safe HaskellNone
LanguageHaskell2010

Data.Blob

Description

This module provides interface for handling large objects - also known as blobs.

One of the use cases for bloc is storing large objects in databases. Instead of storing the entire blob in the database, you can just store the BlobId of the blob.

Multiple values in the database can share a BlobId and we provide an interface for garbage collection for such cases.

Synopsis

Documentation

newtype Blob Source

Wrapper around strict ByteString.

Constructors

Blob ByteString 

Instances

data BlobId Source

BlobId is used to uniquely identify any blob.

data BlobStore Source

This is used to store the base directory.

Each application should maintain there own BlobStores.

data WriteContext Source

WriteContext maintains the blob's state during writing.

data ReadContext Source

ReadContext maintains the reading state of a blob.

openBlobStore :: FilePath -> IO BlobStore Source

All the blobs of an application are stored in the same directory. openBlobStore returns the BlobStore corresponding to a given directory.

newBlob :: BlobStore -> IO WriteContext Source

Creates an empty blob in the given BlobStore.

Use writePartial to write contents to the newly created blob.

writePartial :: WriteContext -> Blob -> IO WriteContext Source

writePartial appends the given blob to the blob referenced by the WriteContext.

endWrite :: WriteContext -> IO BlobId Source

Finalize the write to the given blob.

After calling endWrite no more updates are possible on the blob.

createBlob :: BlobStore -> Blob -> IO BlobId Source

Create a blob from the given contents.

Use createBlob only for small contents. For large contents, use the partial write interface (newBlob followed by calls to writePartial).

startRead :: BlobId -> IO ReadContext Source

Open blob for reading.

readPartial :: ReadContext -> Int -> IO Blob Source

Read given number of bytes from the blob.

skipBytes :: ReadContext -> Integer -> IO () Source

Skip given number of bytes ahead in the blob.

endRead :: ReadContext -> IO () Source

Complete reading from a blob.

readBlob :: BlobId -> IO Blob Source

readBlob reads an entire blob.

Use readBlob only for small blobs. For large blobs, use readPartial instead.

deleteBlob :: BlobId -> IO () Source

Deletes the given blob.

Use deleteBlob only when you are sure that the given blob is not accessible by anyone. If the blob is shared, you should use the GC interface instead.