bento_meta and Metamodel Database (MDB)

bento_meta provides an object representation of property graph based data models, as embodied in the structure of the Bento Metamodel Database, or MDB. The MDB can store multiple data models in terms of their nodes, relationships, and properties. The MDB links these entities according to the structure of the individual models:

For example, model nodes are represented as metamodel nodes of type node, model relationships as metamodel nodes of type relationship, and metamodel relationships link to the appropriate “source” and “destination” nodes. Classes in bento_meta can create, read, update, and link these entities together according to the MDB structure. bento_meta can push and pull them to and from a backing Neo4j database.

The MDB also provides entities for defining and maintaining terminology associated with the stored models. These include the terms themselves, their origin, and associated concepts and predicates. Each of these entities can be created, read, and updated using bento_meta.

Installation

Run:

pip install bento-meta

Contents:

• Metamodel Database (MDB) Principles
  • MDB Motivation and Rationale
  • MDB Design Decisions
  • MDB Entities and Structures
  • Models and Meanings
• MDB Conventions and Patterns
  • MDB entity properties required for consistency
  • MDB indexes needed for additional functionality
  • Conventions for consistent and idempotent updates
  • Graph Patterns for Representation
• The Object Model
  • Working with Models
  • Component Objects
  • Model as a Container
  • Model as an Interface
  • Database Interaction
  • Objects and their Attributes
• bento_meta and MDF
  • Model Description Files
  • Slurping MDF into bento_meta
  • Squirting a Model into MDF YAML
• Model Versioning
  • Using Model Versioning
• Object Map Specifications
  • Defining and Updating Objects
• API Reference
  • bento_meta

A Note on “Nodes”

The metamodel is a property graph, designed to store specific property graph models, in a database built for property graphs. The word “node” is therefore used in different contexts and can be confusing, especially since the Cancer Research Data Commons is also set up in terms of “nodes”, which are central repositories of cancer data of different kinds. This and related documentation will attempt to distinguish these concepts as follows.

• A “graph node” is a instance of the node concept in the property graph model, that usually represents a category or item of interest in the real world, and has associate properties that distinguish it from other instances.

• A “model node” is a graph node within a specific data model, and represents groups of data items (properties) and can be related to other model nodes via model relationships.

• A “metamodel node” is a graph node that represents a model node, model relationship, or model property, in the metamodel database.

• A “Neo4j node” refers generically to the representation of a node in the Neo4j database engine.

• A “CRDC node” refers to a data commons repository that is part of the CRDC, such as the ICDC.

A Note on Objects, Properties, and Attributes

bento_meta creates a mapping between Neo4j nodes and Python objects. Of course, the objects have data associated with them, accessed via setters and getters. These object-associated data are referred to exclusively as attributes in the documentation.

Thus, a Node object has an attribute props (that is, properties), which is an associative array of Property objects. The props attribute is a representation of the has_property relationships between the metamodel node-type node to its metamodel property-type nodes.

See Object attributes for more details.

Indices and tables

• Index

• Module Index

• Search Page