Tags: advanced-user, tutorial, schema

Creating a Data Model#

This tutorial covers creating a simple LinkAhead data model using PyLinkAhead, including Properties, RecordTypes, and inheritance hierarchies.

Concept	PyLinkAhead
Create a Property	`db.Property(name=..., datatype=...)`
Create a RecordType	`db.RecordType(name=...)`
Add a Property to a RecordType	`.add_property()`
Insert a single entity into LinkAhead	`.insert()`
Insert multiple entities into LinkAhead	`db.Container().extend([...]).insert()`
Create a child RecordType	`.add_parent(name=..., inheritance="all")`
Check inherited property importance	`.get_importance()`
Declare a list-typed Property	`db.LIST()`

Prerequisites#

Make sure you have completed First Steps with PyLinkAhead and have access to a writable LinkAhead instance. Import the client and configure your connection as shown in First Steps.

Properties and RecordTypes#

A Data Model in LinkAhead is built from RecordTypes and Properties; for a conceptual overview see LinkAhead Data Model.

Create a Property of datatype double:

temperature = db.Property(name="temperature", datatype=db.DOUBLE)

To insert multiple entities at once, add them to a Container and call .insert() once. The following example builds a RecordType with three properties:

pressure = db.Property(name="pressure", datatype=db.DOUBLE)
duration = db.Property(name="duration", datatype=db.DOUBLE)
recordtype = db.RecordType(name="Experiment")
recordtype.add_property(temperature)
recordtype.add_property(pressure)
recordtype.add_property(duration)
container = db.Container()
container.extend([temperature, pressure, duration, recordtype])
container.insert()
print(recordtype.id)

Output:#

Objects are created locally in Python and become part of the LinkAhead instance only after .insert() is called. The printed value is the server-assigned ID; the actual value depends on the state of your instance.

Inheritance of Properties#

For a conceptual explanation, see Property Inheritance.

To create a child RecordType that inherits all properties from a parent, use .add_parent() with inheritance="all":

rt = db.RecordType(name="ControlledExperiment",
                   description="Experiment with controlled conditions")
rt.add_parent(name="Experiment", inheritance="all")
rt.insert()
print(rt.get_importance("temperature"))

Output:#

RECOMMENDED

.get_importance() returns the importance level the server assigned to the inherited property. RECOMMENDED is the default for properties inherited with inheritance="all".

Note

The inherited properties are only visible after insertion, since they are set by the LinkAhead server, not by the Python client.

List Properties#

To declare a Property that holds multiple values, use db.LIST() wrapping the element datatype:

p = db.Property(name="Measurements", datatype=db.LIST(db.DOUBLE))
p.insert()

The same db.LIST() syntax works in .add_property() when building a RecordType that uses the property:

rt = db.RecordType(name="Experiment")
rt.add_property(name="Measurements", datatype=db.LIST(db.DOUBLE))
rt.insert()

The element type can be any scalar datatype (db.DOUBLE, db.INTEGER, db.TEXT, etc.) or a RecordType name for properties that hold references to multiple records of that type. For example, a Study RecordType whose experiments property references multiple Experiment records:

experiments_prop = db.Property(name="experiments", datatype=db.LIST("Experiment"))
experiments_prop.insert()
study_rt = db.RecordType(name="Study")
study_rt.add_property(experiments_prop)
study_rt.insert()

Summary#

This tutorial covered creating Properties and RecordTypes, inserting them as a group, setting up RecordType hierarchies using inheritance, and declaring list-typed Properties.

Continue with Inserting Records to insert, reference, and update records in LinkAhead.