TypeDB is a polymorphic database with a conceptual data model, a strong subtyping system, a symbolic reasoning engine, and a beautiful and elegant type-theoretic language TypeQL.

IMPORTANT NOTE: > > TypeDB & TypeQL are in the process of being rewritten in Rust. There will be significant refinement to the language, and minor breaks in backwards compatibility. Learn about the changes on our roadmap issue on GitHub. The biggest change to TypeDB 3.0 will be our storage data structure and architecture that significantly boosts performance. We’re aiming to release 3.0 in the summer this year, along with preliminary benchmarks of TypeDB.

Polymorphic databases

Why TypeDB was built

Data frequently exhibits polymorphic features in the form of inheritance hierarchies and interface dependencies. TypeDB was crafted to solve the inability of current database paradigms to natively express these polymorphic features.

• Relational schemas have no native capability for modeling polymorphic data.
• Unstructured databases eliminate the schemas entirely, but this prevents declarative data retrieval.
• ORMs work around the fundamental problem by trading off performance.

Providing full support for polymorphism

In order to fully support polymorphism, a database needs to implement three key components:

• Support for polymorphic schemas that can express inheritance hierarchies and interface implementations.
• Implementation of a fully variablizable query language to support powerful parametric database operations.
• Integration of an inference engine to interpret variables in the semantic context given by the schema.

The TypeDB database

The schema

TypeDB schemas are based on a modern type system that natively supports inheritance and interfaces, and follows a conceptual data modeling approach, in which user-defined types subtype (based on their function) three root types: entities, relations, and attributes.

• Entities are independent objects,
• Relations depend on their role interfaces played by either entities or relations,
• Attributes are properties with a value that can be owned by entities or relations.

Interface and inheritance for these types can be combined in many ways, resulting in highly expressive ways of modeling data.

define

full-name sub attribute, value string;
id sub attribute, value string;
email sub id;
employee-id sub id;

user sub entity,
    owns full-name,
    owns email @unique,
    plays mentorship:trainee;
employee sub user,
    owns employee-id @key,
    plays mentorship:mentor;

mentorship sub relation,
    relates mentor,
    relates trainee;

The query language

The query language of TypeDB is TypeQL. The syntax of TypeQL is fully variablizable and provides native support for polymorphic queries. The language is based on fully declarative and composable patterns, mirroring the structure of natural language.

match $user isa user,
    has full-name $name,
    has email $email;
# This returns all users of any type

match $user isa employee,
    has full-name $name,
    has email $email,
    has employee-id $id;
# This returns only users who are employees

match $user-type sub user;
$user isa $user-type,
    has full-name $name,
    has email $email;
# This returns all users and their type

The inference engine

Any query in TypeDB is semantically validated by TypeDB’s inference engine for consistency with the database schema. This prevents invalid schema updates and data inserts before they can affect the integrity of the database.

TypeDB can also work with data that is not physically stored in the database, but instead logically inferred based on user-specified rules. This enables developers to cleanly separate their source data from their application logic, often allowing for complex systems to be described by combinations of simple rules.

define
rule transitive-team-membership:
    when {
        (team: $team-1, member: $team-2) isa team-membership;
        (team: $team-2, member: $member) isa team-membership;
    } then {
        (team: $team-1, member: $member) isa team-membership;
    };

insert
$john isa user, has email "john@vaticle.com";
$eng isa team, has name "Engineering ";
$cloud isa team, has name "Cloud";
(team: $eng, member: $cloud) isa team-membership;
(team: $cloud, member: $john) isa team-membership;

match
$john isa user, has email "john@vaticle.com";
(team: $team, member: $john) isa team-membership;
# This will return both Cloud and Engineering for $team due to the defined rule

Effective database engineering

TypeDB breaks down the patchwork of existing database paradigms into three fundamental ingredients: types, inheritance, and interfaces. This provides a unified way of working with data across all database applications, that directly impacts development:

• Make use of full object model parity when working with OOP
• Ensure continuous extensibility of your data model
• Work with high-level logical abstractions eliminating the need for physical data modeling
• Let TypeDB's inference engine guarantee data-consistency at all times
• Write high-clarity code with TypeQL's near-natural queries even for the most complex databases
• Unleash the power of fully declarative and composable patterns onto your data

Installation and editions

TypeDB editions

• TypeDB Cloud — multi-cloud DBaaS
• TypeDB Cloud self-hosted — allows you to deploy TypeDB Cloud in your own environment
• TypeDB Core — Open-source edition of TypeDB ← This repository

For a comparison of all three editions, see the Deploy page on our website.

Download and run TypeDB Core

You can download TypeDB from the GitHub Releases.

Check our Installation guide to get started.

Compiling TypeDB Core from source

Note: You DO NOT NEED to compile TypeDB from the source if you just want to use TypeDB. See the "Download and Run TypeDB Core" section above.

1. Make sure you have the following dependencies installed on your machine:

   • Java JDK 11 or higher
   • Bazel 6 or higher.

2. You can build TypeDB with either one of the following commands, depending on the targeted architecture and Operation system:

   $ bazel build //:assemble-linux-x86_64-targz
   $ bazel build //:assemble-linux-arm64-targz
   $ bazel build //:assemble-mac-x86_64-zip
   $ bazel build //:assemble-mac-arm64-zip
   $ bazel build //:assemble-windows-x86_64-zip
   

   Outputs to: bazel-bin/.

3. If you're on a Mac and would like to run any bazel test commands, you will need to install:

   • snappy: brew install snappy
   • jemalloc: brew install jemalloc

Resources

Developer resources

• Documentation: https://typedb.com/docs
• Discussion Forum: https://forum.typedb.com/
• Discord Chat Server: https://typedb.com/discord
• Community Projects: https://github.com/typedb-osi

Useful links

If you want to begin your journey with TypeDB, you can explore the following resources:

• More on TypeDB's features
• In-depth dive into TypeDB's philosophy
• Our TypeDB quickstart
• TypeDB in 25 queries
• TypeQL
• TypeDB Studio

Contributions

TypeDB and TypeQL are built using various open-source frameworks and technologies throughout its evolution. Today TypeDB and TypeQL use Speedb, pest, SCIP, Bazel, gRPC, ZeroMQ, and Caffeine.

Thank you!

In the past, TypeDB was enabled by various open-source products and communities that we are hugely thankful to: RocksDB, ANTLR, Apache Cassandra, Apache Hadoop, Apache Spark, Apache TinkerPop, and JanusGraph.

Package hosting

Package repository hosting is graciously provided by Cloudsmith. Cloudsmith is the only fully hosted, cloud-native, universal package management solution, that enables your organization to create, store and share packages in any format, to any place, with total confidence.

Licensing

This software is developed by Vaticle.
It's released under the Mozilla Public License 2.0 (MPL 2.0). For license information, please see LICENSE.

Vaticle also provides a commercial license for TypeDB Cloud self-hosted - get in touch with our team at commercial@vaticle.com.

Copyright (C) 2023 Vaticle.