Pando

Pando is a history-aware state tree serialization system. Using a Pando Repository, you can save a snapshot of your tree hierarchy, and Pando will keep track of the history of snapshots, including branching histories and backtracking.

Pando is intended for use cases where all of the data you want to save is stored in a single tree hierarchy.

It uses Merkle trees to store a complete history of a tree over a series of snapshots while avoiding data duplication (each unique blob is stored only once and referenced by hash).

Usage

The entry point to Pando is the PandoRepository<T> class, where T is the type of the tree that you want to save in this repository. The public interface surface is defined in Pando.Repositories.IRepository<T>.

Creating a PandoRepository

In order to create a PandoRepository, you must define a data source and a serializer. The data source determines the destination to which data will be saved, while the serializer defines how to convert the tree into binary data.

The data source is of type IDataSource and the Pando library offers two built in options to fulfill this dependency (see data sources).

The serializer must be user defined (or source generated) as the logic will be specific to the type that you are serializing (see serializers).

Saving to a PandoRepository

Save a snapshot of your state tree:

ulong snapshotHash = myPandoRepository.SaveSnapshot(myTreeRoot);

The returned hash can be used to identify and retrieve the snapshot at a later time.

Getting data from a PandoRepository

Get a previously saved snapshot:

var myTreeRootAtSnapshot = myPandoRepository.GetSnapshot(snapshotHash);

Getting the full history from a PandoRepository

Get a tree representation of the full history of all snapshots:

SnapshotTree history = myPandoRepository.GetSnapshotTree();

SnapshotTree is a recursive data structure that contains the hash of a snapshot, and a set of child SnapshotTrees.

Credits

• Tree by Diky Setiawan from NounProject.com