Welcome! This is just a little experimental project on deeply-embedded graph databases.
The main goal here is to attempt to prototype a graph database system with three main goals:
- Be as size-efficient and fast as possible, and
- Be as easy to use as possible,
- Depend on non-built-in libraries as little as possible, for ease of deployment.
Bakdoh is just a hobby project (well, for now); it won't be big and professional, unlike Neo4j or OrientDB.
At the moment there are only two parts to this project:
The TAGS is an API and format for graph databases. Contents are stored as individually-accessible values called Anchors interconnected by Relations. No complete objects are stored in the database; Relations serve to reconstruct the objects, as well as maintain links between objects.
This allows sharing (normalisation) of values between objects.
The API is found in the tags.py module.
SQLiteRepo stores TAGS databases inside a two-table SQLite database: one for the graph and the other for configuration. The repository code is kept in the same module as the TAGS API, tags.py.