Propertree

Map Python data structures to classes and create tree of resolvable objects.

Overview

This library provides a set of classes that can be used to take Python data structures and map them to a set of properties organised in a tree-like structure. The basic idea is that a Python object with one or more attributes can be represented by a unique key in a dictionary whereby the the value of that key represents the internal makeup of the object i.e. attributes etc.

The basic design is as follows:

A PTreeSection object takes Python data structures dict/list/literal after having been converted from YAML or JSON. This input is treated like a tree of properties whereby each level of a structure equates to a branch containing a set of "property overrides" which are identified by a unique root key that maps to an implementation of PTreeOverrideBase. Descendant lists, dictionaries or unknown literals are treated new branches.

Once the whole tree has been mapped, objects can be retreived by either iterating over the root object or accessing them directly as attributes.

For example lets say we have the following YAML:

config:
  fire:
    danger:
      level: high
  banana:
    danger:
      level: low

And the following accompanying Python code:

from propertree.propertree2 import PTreeOverrideBase, PTreeSection

class Config(PTreeOverrideBase):
  override_keys = ['config']


root = PTreeSection(MYYAML)

We can then access the config as follows:

print(root.config.fire.danger.level)
print(root.config.banana.danger.level)

Property Inheritance

Property inheritance is supported by passing down all properties identified at a branch level to all descendent branches. This allows property objects to access any property within it's call chain although it would be the most recently overriden version if one exists. For example:

input: /etc/foo
checks:
  chk1:
    input: /etc/bar
    condition: C1
  chk2:
    condition: C2

In the above, checks.chk1.input would be "/etc/bar" whereas checks.chk2.input would be "/etc/foo".

Mapped Properties

It is possible to compose complex properties made up of one or more "member" properties. These are called mapped properties and are provided by the PTreeMappedOverrideBase class. A mapped property has two parts; a primary and its members. A property can only be a member of one mapped property (i.e. be associated with a single primary). These properties also have the special feature that allow them to be defined either explicitly using their full construct i.e. primary and members or implicitly using just their members. If the latter form is used, the primary is implicitly created such that when the members are accessed it is always done through the primary.

For example, here is some code to define a mapped property:

from propertree.propertree2 import PTreeOverrideBase, PTreeMappedOverrideBase, PTreeSection

class MapPrimary(PTreeMappedOverrideBase):
  override_keys = ['mapprimary']
  override_members = ['member1']


class Member1(PTreeOverrideBase):
  override_keys = ['member1']

...

the following will both behave the same when accessed:

The explicit declaration:

mapprimary:
  member1:
    attr1:

And the counterpart implicit declaration:

member1:
  attr1:

Are both accessed as follows:

mapprimary.member1.attr1

Logical Groupings

Sometimes it might be useful to represent properties or content within a property as a logical function. To achieve this, propertree has builtin logical operator properties which map to the PTreeLogicalGrouping class. This class provides a default implementation of logical operations that can be used by any property. Grouped items are expected to have a result attribute that returns a boolean result such that the respective logical operator of that grouping is then applied to the set of all results.

In the following example we have a top-level operator with two items, each of which is itself an operator, the first having two property items and second having one. This equates to AND(OR(P1, P2), NOT(P3)).

and:
  or: [P1, P2]
  not: P3

Example code for this would look like:

from propertree.propertree2 import PTreeSection

root = PTreeSection(MYYAML)
result = root.and.result