conestack / node

Other
24 stars 24 forks source link

Node

.. image:: https://img.shields.io/pypi/v/node.svg :target: https://pypi.python.org/pypi/node :alt: Latest PyPI version

.. image:: https://img.shields.io/pypi/dm/node.svg :target: https://pypi.python.org/pypi/node :alt: Number of PyPI downloads

.. image:: https://github.com/conestack/node/actions/workflows/test.yml/badge.svg :target: https://github.com/conestack/node/actions/workflows/test.yml :alt: Test node

Overview

Node is a library to create nested data models and structures.

These data models are described as trees of nodes, optionally with attributes and schema definitions.

They utilize:

One purpose of this package is to provide a unified API to different backend storages. Specific storage related implementations are for example:

Another usecase is providing interfaces for specific application domains.

E.g. for user and group management, node.ext.ugm <https://pypi.org/project/node.ext.ugm> defines the interfaces. Additional it implements a file based default implementation. Then there are specific implementations of those interfaces in node.ext.ldap <https://pypi.org/project/node.ext.ldap> and cone.sql <https://pypi.org/project/cone.sql>_, to access users and groups in LDAP and SQL databases.

This package is also used to build in-memory models of all sorts.

E.g. yafowil <https://pypi.org/project/yafowil>_ is a HTML form processing and rendering library. It uses node trees for declarative description of the form model.

Another one to mention is cone.app <https://pypi.org/project/cone.app>, a Pyramid <https://pypi.org/project/pyramid> based development environment for web applications, which uses node trees to describe the application model.

Basic Usage

There are two basic node types. Mapping nodes and sequence nodes. This package provides some basic nodes to start from.

Mapping Nodes


Mapping nodes implement ``node.interfaces.IMappingNode``. A mapping in python
is a container object that supports arbitrary key lookups and implements the
methods specified in the ``MutableMapping`` of pythons abstract base classes
respective ``zope.interface.common.mapping.IFullMapping``.

An unordered node. This can be used as base for trees where order of items
doesn't matter:

.. code-block:: python

    from node.base import BaseNode

    root = BaseNode(name='root')
    root['child'] = BaseNode()

An ordered node. The order of items is preserved:

.. code-block:: python

    from node.base import OrderedNode

    root = OrderedNode(name='orderedroot')
    root['foo'] = OrderedNode()
    root['bar'] = OrderedNode()

With ``printtree`` we can do a quick inspection of our node tree:

.. code-block:: pycon

    >>> root.printtree()
    <class 'node.base.OrderedNode'>: orderedroot
      <class 'node.base.OrderedNode'>: foo
      <class 'node.base.OrderedNode'>: bar

Sequence Nodes

Sequence nodes implement node.interfaces.ISequenceNode. In the context of this library, a sequence is an implementation of the methods specified in the MutableSequence of pythons abstract base classes respective zope.interface.common.collections.IMutableSequence.

Using a list node:

.. code-block:: python

from node.base import BaseNode
from node.base import ListNode

root = ListNode(name='listroot')
root.insert(0, BaseNode())
root.insert(1, BaseNode())

Check tree structure with printtree:

.. code-block:: pycon

>>> root.printtree()
<class 'node.base.ListNode'>: listnode
  <class 'node.base.BaseNode'>: 0
  <class 'node.base.BaseNode'>: 1

.. note::

Sequence nodes are introduced as of node 1.0 and are not as feature rich
as mapping nodes (yet). If you find inconsistencies or missing features,
please file an issue or create a pull request at github.

Behaviors


``node`` utilizes the `plumber <http://pypi.python.org/pypi/plumber>`_ package.

The different functionalities of nodes are provided as plumbing behaviors:

.. code-block:: python

    from node.behaviors import DefaultInit
    from node.behaviors import MappingNode
    from node.behaviors import OdictStorage
    from plumber import plumbing

    @plumbing(
        DefaultInit,
        MappingNode,
        OdictStorage)
    class CustomNode:
        pass

When inspecting the ``CustomNode`` class, we can see it was plumbed using given
behaviors, now representing a complete node implementation:

.. code-block:: pycon

    >>> dir(CustomNode)
    ['__bool__', '__class__', '__contains__', '__delattr__', '__delitem__',
    '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__',
    '__getattribute__', '__getitem__', '__gt__', '__hash__', '__implemented__',
    '__init__', '__init_subclass__', '__iter__', '__le__', '__len__', '__lt__',
    '__module__', '__name__', '__ne__', '__new__', '__nonzero__', '__parent__',
    '__plumbing__', '__plumbing_stacks__', '__provides__', '__reduce__',
    '__reduce_ex__', '__repr__', '__setattr__', '__setitem__', '__sizeof__',
    '__str__', '__subclasshook__', '__weakref__', 'acquire', 'clear', 'copy',
    'deepcopy', 'detach', 'filtereditems', 'filtereditervalues', 'filteredvalues',
    'get', 'has_key', 'items', 'iteritems', 'iterkeys', 'itervalues', 'keys',
    'name', 'noderepr', 'parent', 'path', 'pop', 'popitem', 'printtree', 'root',
    'setdefault', 'storage', 'treerepr', 'update', 'values']

Please read the documentation of ``plumber`` for detailed information about the
plumbing system.

Attributes

While it is not strictly necessary, it's a good idea to separate the hierarchical structure of a model from the node related attributes to avoid naming conflicts. Attributes are provided via node.behaviors.Attributes plumbing behavior:

.. code-block:: python

from node.behaviors import Attributes
from node.behaviors import DefaultInit
from node.behaviors import DictStorage
from node.behaviors import MappingNode
from plumber import plumbing

@plumbing(
    Attributes,
    DefaultInit,
    MappingNode,
    DictStorage)
class NodeWithAttributes:
    pass

The node now provides an attrs attribute. Node attributes are itself just a node:

.. code-block:: pycon

>>> node = NodeWithAttributes()
>>> attrs = node.attrs
>>> attrs
<NodeAttributes object 'None' at ...>

>>> attrs['foo'] = 'foo'

If it's desired to access attribute members via python attribute access, attribute_access_for_attrs must be set on node:

.. code-block:: pycon

>>> node.attribute_access_for_attrs = True
>>> attrs = node.attrs
>>> attrs.foo = 'bar'
>>> attrs.foo
'bar'

A custom attributes implementation can be set by defining attributes_factory on the node:

.. code-block:: python

from node.behaviors import NodeAttributes

class CustomAttributes(NodeAttributes):
    pass

class CustomAttributesNode(NodeWithAttributes):
    attributes_factory = CustomAttributes

This factory is then used to instantiate the attributes:

.. code-block:: pycon

>>> node = CustomAttributesNode()
>>> node.attrs
<CustomAttributes object 'None' at ...>

Schema


To describe the data types of node members, this package provides a mechanism
for defining schemata.

This can happen in different ways. One is to define the schema for node members
directly. This is useful for nodes representing a leaf in the hierarchy or for
node attribute nodes:

.. code-block:: python

    from node import schema
    from node.base import BaseNode
    from node.behaviors import DefaultInit
    from node.behaviors import DictStorage
    from node.behaviors import MappingNode
    from node.behaviors import Schema
    from plumber import plumbing

    @plumbing(
        DefaultInit,
        MappingNode,
        DictStorage,
        Schema)
    class SchemaNode:
        schema = {
            'int': schema.Int(),
            'float': schema.Float(default=1.),
            'str': schema.Str(),
            'bool': schema.Bool(default=False),
            'node': schema.Node(BaseNode)
        }

Children defined in the schema provide a default value. If not explicitely
defined, the default value is always ``node.utils.UNSET``:

.. code-block:: pycon

    >>> node = SchemaNode()
    >>> node['int']
    <UNSET>

    >>> node['float']
    1.0

    >>> node['bool']
    False

Children defined in the schema are validated against the defined type when
setting it's value:

.. code-block:: pycon

    >>> node = SchemaNode()
    >>> node['int'] = 'A'
    Traceback (most recent call last):
      ...
    ValueError: A is no <class 'int'> type

For accessing members defined in the schema as node attributes,
``SchemaAsAttributes`` plumbing behavior can be used:

.. code-block:: python

    from node.behaviors import SchemaAsAttributes

    @plumbing(SchemaAsAttributes)
    class SchemaAsAttributesNode(BaseNode):
        schema = {
            'int': schema.Int(default=1),
        }

Node ``attrs`` now provides access to the schema members:

.. code-block:: pycon

    >>> node = SchemaAsAttributesNode()
    >>> node.attrs['int']
    1

Schema members can also be defined as class attributes. This is syntactically
the most elegant way, but comes with the tradeoff of possible naming conflicts:

.. code-block:: python

    from node.behaviors import SchemaProperties

    @plumbing(
        DefaultInit,
        MappingNode,
        DictStorage,
        SchemaProperties)
    class SchemaPropertiesNode:
        text = schema.Str(default='Text')

Here we access ``text`` as class attribute:

.. code-block:: pycon

    >>> node = SchemaPropertiesNode()
    >>> node.text
    'Text'

    >>> node.text = 1
    Traceback (most recent call last):
      ...
    ValueError: 1 is no <class 'str'> type

Plumbing Behaviors
------------------

General Behaviors

node.behaviors.DefaultInit Plumbing behavior providing default __init__ function on node. This behavior is going to be deprecated in future versions. Use node.behaviors.NodeInit instead. See node.interfaces.IDefaultInit.

node.behaviors.NodeInit Plumbing behavior for transparent setting of __name__ and __parent__ at object initialization time. See node.interfaces.INodeInit.

node.behaviors.Node Fill in gaps for full INode API. See node.interfaces.INode.

node.behaviors.ContentishNode A node which can contain children. See node.interfaces.IContentishNode. Concrete implementations are node.behaviors.MappingNode and node.behaviors.SequenceNode.

node.behaviors.Attributes Provide attributes on node. See node.interfaces.IAttributes. If node.behaviors.Nodespaces is applied on node, the attributes instance gets stored internally in __attrs__ nodespace, otherwise its set on __attrs__ attribute.

node.behaviors.Events Provide an event registration and dispatching mechanism. See node.interfaces.IEvents.

node.behaviors.BoundContext Mechanism for scoping objects to interfaces and classes. See node.interfaces.IBoundContext.

node.behaviors.NodeReference Plumbing behavior holding an index of nodes contained in the tree. See node.interfaces.INodeReference.

node.behaviors.WildcardFactory Plumbing behavior providing factories by wildcard patterns. See node.interfaces.IWildcardFactory.

Mapping Behaviors


**node.behaviors.MappingNode**
    Turn an object into a mapping node. Extends ``node.behaviors.Node``.
    See ``node.interfaces.IMappingNode``.

**node.behaviors.MappingAdopt**
    Plumbing behavior that provides ``__name__`` and ``__parent__`` attribute
    adoption on child nodes of mapping.
    See ``node.interfaces.IMappingAdopt``.

**node.behaviors.MappingConstraints**
    Plumbing behavior for constraints on mapping nodes.
    See ``node.interfaces.IMappingConstraints``.

**node.behaviors.UnicodeAware**
    Plumbing behavior to ensure unicode for keys and string values.
    See ``node.interfaces.IUnicodeAware``.

**node.behaviors.Alias**
    Plumbing behavior that provides aliasing of child keys.
    See ``node.interfaces.IAlias``.

**node.behaviors.AsAttrAccess**
    Plumbing behavior to get node as IAttributeAccess implementation.
    See ``node.interfaces.IAsAttrAccess``.

**node.behaviors.ChildFactory**
    Plumbing behavior providing child factories.
    See ``node.interfaces.IChildFactory``.

**node.behaviors.FixedChildren**
    Plumbing Behavior that initializes a fixed dictionary as children.
    See ``node.interfaces.IFixedChildren``.

**node.behaviors.Nodespaces**
    Plumbing behavior for providing nodespaces on node.
    See ``node.interfaces.INodespaces``.

**node.behaviors.Lifecycle**
    Plumbing behavior taking care of lifecycle events.
    See ``node.interfaces.ILifecycle``.

**node.behaviors.AttributesLifecycle**
    Plumbing behavior for handling lifecycle events on attribute manipulation.
    See ``node.interfaces.IAttributesLifecycle``.

**node.behaviors.Invalidate**
    Plumbing behavior for node invalidation.
    See ``node.interfaces.Invalidate``.

**node.behaviors.VolatileStorageInvalidate**
    Plumbing behavior for invalidating nodes using a volatile storage.
    See ``node.interfaces.Invalidate``.

**node.behaviors.Cache**
    Plumbing behavior for caching.
    See ``node.interfaces.ICache``.

**node.behaviors.MappingOrder**
    Plumbing behavior for ordering support on mapping nodes.
    See ``node.interfaces.IMappingOrder``.

**node.behaviors.UUIDAware**
    Plumbing behavior providing a uuid on nodes.
    See ``node.interfaces.IUUIDAware``.

**node.behaviors.MappingReference**
    Plumbing behavior to provide ``node.interfaces.INodeReference`` on mapping
    nodes. See ``node.interfaces.IMappingReference``.

**node.behaviors.MappingStorage**
    Provide abstract mapping storage access.
    See ``node.interfaces.IMappingStorage``.

**node.behaviors.DictStorage**
    Provide dictionary storage. Extends ``node.behaviors.MappingStorage``.
    See ``node.interfaces.IMappingStorage``.

**node.behaviors.OdictStorage**
    Provide ordered dictionary storage. Extends
    ``node.behaviors.MappingStorage``. See ``node.interfaces.IMappingStorage``.

**node.behaviors.Fallback**
    Provide a way to fall back to values by subpath stored on another node.
    See ``node.interfaces.IFallback``.

**node.behaviors.Schema**
    Provide schema validation and value serialization on node values.
    See ``node.interfaces.ISchema``.

**node.behaviors.SchemaAsAttributes**
    Provide schema validation and value serialization on node values via
    dedicated attributes object.
    See ``node.interfaces.ISchemaAsAttributes``.

**node.behaviors.SchemaProperties**
    Provide schema fields as class properties.
    See ``node.interfaces.ISchemaProperties``.

**node.behaviors.MappingFilter**
    Filter mapping children by class or interface.
    See ``node.interfaces.IChildFilter``.

Sequence Behaviors

node.behaviors.SequenceNode Turn an object into a sequence node. Extends node.behaviors.Node. See node.interfaces.IMappingNode.

node.behaviors.SequenceAdopt Plumbing behavior that provides __name__ and __parent__ attribute adoption on child nodes of sequence. See node.interfaces.ISequenceAdopt.

node.behaviors.SequenceConstraints Plumbing behavior for constraints on sequence nodes. See node.interfaces.ISequenceConstraints.

node.behaviors.SequenceStorage Provide abstract sequence storage access. See node.interfaces.ISequenceStorage.

node.behaviors.ListStorage Provide list storage. See node.interfaces.ISequenceStorage.

node.behaviors.SequenceReference Plumbing behavior to provide node.interfaces.INodeReference on sequence nodes. See node.interfaces.ISequenceReference.

node.behaviors.SequenceFilter Filter sequence children by class or interface. See node.interfaces.IChildFilter.

node.behaviors.SequenceOrder Plumbing behavior for ordering support on sequence nodes. See node.interfaces.ISequenceOrder.

JSON Serialization

Nodes can be serialized to and deserialized from JSON:

.. code-block:: pycon

>>> from node.serializer import serialize
>>> json_dump = serialize(BaseNode(name='node'))

>>> from node.serializer import deserialize
>>> deserialize(json_dump)
<BaseNode object 'node' at ...>

For details on serialization API please read file in docs/archive/serializer.rst.

Python Versions

Contributors