node.ext.ldap

Overview

node.ext.ldap is a LDAP convenience library for LDAP communication based on python-ldap (version 2.4 or later) and node.

The package contains base configuration and communication objects, a LDAP node object and a LDAP node based user and group management implementation utilizing node.ext.ugm.

This package is the successor of bda.ldap.

API changes compared to 0.9.x

• LDAPNode instances cannot have direct children of subtree any longer. This was a design flaw because of possible duplicate RDN's.
• LDAPNode.search returns DN's instead of RDN's by default.
• Secondary keys and alternative key attribute features have been removed entirely from LDAPNode.
• LDAPProps.check_duplicates setting has been removed.

Usage

LDAP Properties

To define connection properties for LDAP use node.ext.ldap.LDAPProps object:

>>> from node.ext.ldap import LDAPProps

>>> props = LDAPProps(
...     uri='ldap://localhost:12345/',
...     user='cn=Manager,dc=my-domain,dc=com',
...     password='secret',
...     cache=False
... )

Test server connectivity with node.ext.ldap.testLDAPConnectivity:

>>> from node.ext.ldap import testLDAPConnectivity

>>> assert testLDAPConnectivity(props=props) == 'success'

LDAP Connection

For handling LDAP connections, node.ext.ldap.LDAPConnector is used. It expects a LDAPProps instance in the constructor. Normally there is no need to instantiate this object directly, this happens during creation of higher abstractions, see below:

>>> from node.ext.ldap import LDAPConnector
>>> import ldap

>>> connector = LDAPConnector(props=props)

Calling bind creates and returns the LDAP connection:

>>> conn = connector.bind()
>>> assert isinstance(conn, ldap.ldapobject.ReconnectLDAPObject)

Calling unbind destroys the connection:

>>> connector.unbind()

LDAP Communication

For communicating with an LDAP server, node.ext.ldap.LDAPCommunicator is used. It provides all the basic functions needed to search and modify the directory.

LDAPCommunicator expects a LDAPConnector instance at creation time:

>>> from node.ext.ldap import LDAPCommunicator

>>> communicator = LDAPCommunicator(connector)

Bind to server:

>>> communicator.bind()

Adding directory entry:

>>> communicator.add(
...     'cn=foo,ou=demo,dc=my-domain,dc=com',
...     {
...         'cn': 'foo',
...         'sn': 'Mustermann',
...         'userPassword': 'secret',
...         'objectClass': ['person'],
...     }
... )

Set default search DN:

>>> communicator.baseDN = 'ou=demo,dc=my-domain,dc=com'

Search in directory:

>>> import node.ext.ldap

>>> res = communicator.search(
...     '(objectClass=person)',
...     node.ext.ldap.SUBTREE
... )

>>> assert res == [(
...     'cn=foo,ou=demo,dc=my-domain,dc=com',
...     {
...         'objectClass': ['person'],
...         'userPassword': ['secret'],
...         'cn': ['foo'],
...         'sn': ['Mustermann']
...     }
... )]

Modify directory entry:

>>> from ldap import MOD_REPLACE

>>> communicator.modify(
...     'cn=foo,ou=demo,dc=my-domain,dc=com',
...     [(MOD_REPLACE, 'sn', 'Musterfrau')]
... )

>>> res = communicator.search(
...     '(objectClass=person)',
...     node.ext.ldap.SUBTREE,
...     attrlist=['cn']
... )

>>> assert res == [('cn=foo,ou=demo,dc=my-domain,dc=com', {'cn': ['foo']})]

Change the password of a directory entry which represents a user:

>>> communicator.passwd(
...     'cn=foo,ou=demo,dc=my-domain,dc=com',
...     'secret',
...     '12345'
... )

>>> res = communicator.search(
...     '(objectClass=person)',
...     node.ext.ldap.SUBTREE,
...     attrlist=['userPassword']
... )

>>> assert res == [(
...     'cn=foo,ou=demo,dc=my-domain,dc=com',
...     {'userPassword': ['{SSHA}...']}
... )]

Delete directory entry:

>>> communicator.delete('cn=foo,ou=demo,dc=my-domain,dc=com')

>>> res = communicator.search(
...     '(objectClass=person)',
...     node.ext.ldap.SUBTREE
... )

>>> assert res == []

Close connection:

>>> communicator.unbind()

LDAP Session

A more convenient way for dealing with LDAP is provided by node.ext.ldap.LDAPSession. It basically provides the same functionality as LDAPCommunicator, but automatically creates the connectivity objects and checks the connection state before performing actions.

Instantiate LDAPSession object. Expects LDAPProps instance:

>>> from node.ext.ldap import LDAPSession

>>> session = LDAPSession(props)

LDAP session has a convenience to check given properties:

>>> res = session.checkServerProperties()

>>> assert res == (True, 'OK')

Set default search DN for session:

>>> session.baseDN = 'ou=demo,dc=my-domain,dc=com'

Search in directory:

>>> res = session.search()

>>> assert res == [
...     ('ou=demo,dc=my-domain,dc=com',
...     {
...         'objectClass': ['top', 'organizationalUnit'],
...         'ou': ['demo'],
...         'description': ['Demo organizational unit']
...     }
... )]

Add directory entry:

>>> session.add(
...     'cn=foo,ou=demo,dc=my-domain,dc=com',
...     {
...         'cn': 'foo',
...         'sn': 'Mustermann',
...         'userPassword': 'secret',
...         'objectClass': ['person'],
...     }
... )

Change the password of a directory entry which represents a user:

>>> session.passwd('cn=foo,ou=demo,dc=my-domain,dc=com', 'secret', '12345')

Authenticate a specific user:

>>> res = session.authenticate('cn=foo,ou=demo,dc=my-domain,dc=com', '12345')

>>> assert res is True

Modify directory entry:

>>> session.modify(
...     'cn=foo,ou=demo,dc=my-domain,dc=com',
...     [(MOD_REPLACE, 'sn', 'Musterfrau')]
... )

>>> res = session.search(
...     '(objectClass=person)',
...     node.ext.ldap.SUBTREE,
...     attrlist=['cn']
... )

>>> assert res == [(
...     'cn=foo,ou=demo,dc=my-domain,dc=com',
...     {'cn': ['foo']}
... )]

Delete directory entry:

>>> session.delete('cn=foo,ou=demo,dc=my-domain,dc=com')

>>> res = session.search('(objectClass=person)', node.ext.ldap.SUBTREE)

>>> assert res == []

Close session:

>>> session.unbind()

LDAP Nodes

One can deal with LDAP entries as node objects. Therefor node.ext.ldap.LDAPNode is used. To get a clue of the complete node API, see node package.

Create a LDAP node. The root Node expects the base DN and a LDAPProps instance:

>>> from node.ext.ldap import LDAPNode

>>> root = LDAPNode('ou=demo,dc=my-domain,dc=com', props=props)

Every LDAP node has a DN and a RDN:

>>> root.DN
u'ou=demo,dc=my-domain,dc=com'

>>> root.rdn_attr
u'ou'

Check whether created node exists in the database:

>>> root.exists
True

Directory entry has no children yet:

>>> root.keys()
[]

Add children to root node:

>>> person = LDAPNode()
>>> person.attrs['objectClass'] = ['person', 'inetOrgPerson']
>>> person.attrs['sn'] = 'Mustermann'
>>> person.attrs['userPassword'] = 'secret'
>>> root['cn=person1'] = person

>>> person = LDAPNode()
>>> person.attrs['objectClass'] = ['person', 'inetOrgPerson']
>>> person.attrs['sn'] = 'Musterfrau'
>>> person.attrs['userPassword'] = 'secret'
>>> root['cn=person2'] = person

If the RDN attribute was not set during node creation, it is computed from node key and set automatically:

>>> person.attrs['cn']
u'person2'

Fetch children DN by key from LDAP node:

>>> root.child_dn('cn=person1')
u'cn=person1,ou=demo,dc=my-domain,dc=com'

Have a look at the tree:

>>> root.printtree()

  
  

The entries have not been written to the directory yet. When modifying a LDAP node tree, everything happens im memory. Persisting is done by calling the tree, or a part of it. You can check sync state of a node with its changed flag. If changed is True it means either that the node attributes or node children has changed:

>>> root.changed
True

>>> root()
>>> root.changed
False

Modify a LDAP node:

>>> person = root['cn=person1']

Modify existing attribute:

>>> person.attrs['sn'] = 'Mustermensch'

Add new attribute:

>>> person.attrs['description'] = 'Mustermensch description'
>>> person()

Delete an attribute:

>>> del person.attrs['description']
>>> person()

Delete LDAP node:

>>> del root['cn=person2']
>>> root()
>>> root.printtree()

  

Searching LDAP

Add some users and groups we'll search for:

>>> for i in range(2, 6):
...     node = LDAPNode()
...     node.attrs['objectClass'] = ['person', 'inetOrgPerson']
...     node.attrs['sn'] = 'Surname %s' % i
...     node.attrs['userPassword'] = 'secret%s' % i
...     node.attrs['description'] = 'description%s' % i
...     node.attrs['businessCategory'] = 'group1'
...     root['cn=person%s' % i] = node

>>> node = LDAPNode()
>>> node.attrs['objectClass'] = ['groupOfNames']
>>> node.attrs['member'] = [
...     root.child_dn('cn=person1'),
...     root.child_dn('cn=person2'),
... ]
... node.attrs['description'] = 'IT'
>>> root['cn=group1'] = node

>>> node = LDAPNode()
>>> node.attrs['objectClass'] = ['groupOfNames']
>>> node.attrs['member'] = [
...     root.child_dn('cn=person4'),
...     root.child_dn('cn=person5'),
... ]
>>> root['cn=group2'] = node

>>> root()
>>> root.printtree()

  
  
  
  
  
  
  

For defining search criteria LDAP filters are used, which can be combined by bool operators '&' and '|':

>>> from node.ext.ldap import LDAPFilter

>>> filter = LDAPFilter('(objectClass=person)')
>>> filter |= LDAPFilter('(objectClass=groupOfNames)')

>>> res = sorted(root.search(queryFilter=filter))

>>> assert res == [
...     u'cn=group1,ou=demo,dc=my-domain,dc=com',
...     u'cn=group2,ou=demo,dc=my-domain,dc=com',
...     u'cn=person1,ou=demo,dc=my-domain,dc=com',
...     u'cn=person2,ou=demo,dc=my-domain,dc=com',
...     u'cn=person3,ou=demo,dc=my-domain,dc=com',
...     u'cn=person4,ou=demo,dc=my-domain,dc=com',
...     u'cn=person5,ou=demo,dc=my-domain,dc=com'
... ]

Define multiple criteria LDAP filter:

>>> from node.ext.ldap import LDAPDictFilter

>>> filter = LDAPDictFilter({
...     'objectClass': ['person'],
...     'cn': 'person1'
... })

>>> res = root.search(queryFilter=filter)

>>> assert res == [u'cn=person1,ou=demo,dc=my-domain,dc=com']

Define a relation LDAP filter. In this case we build a relation between group 'cn' and person 'businessCategory':

>>> from node.ext.ldap import LDAPRelationFilter

>>> filter = LDAPRelationFilter(root['cn=group1'], 'cn:businessCategory')

>>> res = root.search(queryFilter=filter)

>>> assert res == [
...     u'cn=person2,ou=demo,dc=my-domain,dc=com',
...     u'cn=person3,ou=demo,dc=my-domain,dc=com',
...     u'cn=person4,ou=demo,dc=my-domain,dc=com',
...     u'cn=person5,ou=demo,dc=my-domain,dc=com'
... ]

Different LDAP filter types can be combined:

>>> filter &= LDAPFilter('(cn=person2)')
>>> str(filter)
'(&(businessCategory=group1)(cn=person2))'

The following keyword arguments are accepted by LDAPNode.search. If multiple keywords are used, combine search criteria with '&' where appropriate.

If attrlist is given, the result items consists of 2-tuples with a dict containing requested attributes at position 1:

queryFilter
Either a LDAP filter instance or a string. If given argument is string type, a LDAPFilter instance is created.

criteria
A dictionary containing search criteria. A LDAPDictFilter instance is created.

attrlist
List of attribute names to return. Special attributes rdn and dn are allowed.

relation
Either LDAPRelationFilter instance or a string defining the relation. If given argument is string type, a LDAPRelationFilter instance is created.

relation_node
In combination with relation argument, when given as string, use relation_node instead of self for filter creation.

exact_match
Flag whether 1-length result is expected. Raises an error if empty result or more than one entry found.

or_search
In combination with criteria, this parameter is passed to the creation of LDAPDictFilter. This flag controls whether to combine criteria keys and values with '&' or '|'.

or_keys
In combination with criteria, this parameter is passed to the creation of LDAPDictFilter. This flag controls whether criteria keys are combined with '|' instead of '&'.

or_values
In combination with criteria, this parameter is passed to the creation of LDAPDictFilter. This flag controls whether criteria values are combined with '|' instead of '&'.

page_size
Used in conjunction with cookie for querying paged results.

cookie
Used in conjunction with page_size for querying paged results.

get_nodes
If True result contains LDAPNode instances instead of DN's

You can define search defaults on the node which are always considered when calling search on this node. If set, they are always '&' combined with any (optional) passed filters.

Define the default search scope:

>>> from node.ext.ldap import SUBTREE

>>> root.search_scope = SUBTREE

Define default search filter, could be of type LDAPFilter, LDAPDictFilter, LDAPRelationFilter or string:

>>> root.search_filter = LDAPFilter('objectClass=groupOfNames')

>>> res = root.search()

>>> assert res == [
...     u'cn=group1,ou=demo,dc=my-domain,dc=com',
...     u'cn=group2,ou=demo,dc=my-domain,dc=com'
... ]

>>> root.search_filter = None

Define default search criteria as dict:

>>> root.search_criteria = {'objectClass': 'person'}

>>> res = root.search()

>>> assert res == [
...     u'cn=person1,ou=demo,dc=my-domain,dc=com',
...     u'cn=person2,ou=demo,dc=my-domain,dc=com',
...     u'cn=person3,ou=demo,dc=my-domain,dc=com',
...     u'cn=person4,ou=demo,dc=my-domain,dc=com',
...     u'cn=person5,ou=demo,dc=my-domain,dc=com'
... ]

Define default search relation:

>>> root.search_relation = LDAPRelationFilter(
...     root['cn=group1'],
...     'cn:businessCategory'
... )

>>> res = root.search()

>>> assert res == [
...     u'cn=person2,ou=demo,dc=my-domain,dc=com',
...     u'cn=person3,ou=demo,dc=my-domain,dc=com',
...     u'cn=person4,ou=demo,dc=my-domain,dc=com',
...     u'cn=person5,ou=demo,dc=my-domain,dc=com'
... ]

Again, like with the keyword arguments, multiple defined defaults are '&' combined:

# empty result, there are no groups with group 'cn' as 'description'
>>> root.search_criteria = {'objectClass': 'group'}

>>> res = root.search()

>>> assert res == []

JSON Serialization

Serialize and deserialize LDAP nodes:

>>> root = LDAPNode('ou=demo,dc=my-domain,dc=com', props=props)

Serialize children:

>>> from node.serializer import serialize

>>> json_dump = serialize(root.values())

Clear and persist root:

>>> root.clear()

>>> root()

Deserialize JSON dump:

>>> from node.serializer import deserialize

>>> deserialize(json_dump, root=root)
[,
,
,
,
,
,
]

Since root has been given, created nodes were added:

>>> root()
>>> root.printtree()

  
  
  
  
  
  
  

Non simple vs simple mode. Create container with children:

>>> container = LDAPNode()
>>> container.attrs['objectClass'] = ['organizationalUnit']
>>> root['ou=container'] = container

>>> person = LDAPNode()
>>> person.attrs['objectClass'] = ['person', 'inetOrgPerson']
>>> person.attrs['sn'] = 'Mustermann'
>>> person.attrs['userPassword'] = 'secret'
>>> container['cn=person1'] = person

>>> root()

Serialize in default mode contains type specific information. Thus JSON dump can be deserialized later:

>>> serialized = serialize(container)

>>> assert serialized == (
...     '{'
...         '"__node__": {'
...             '"attrs": {'
...                 '"objectClass": ["organizationalUnit"], '
...                 '"ou": "container"'
...             '}, '
...             '"children": [{'
...                 '"__node__": {'
...                     '"attrs": {'
...                         '"objectClass": ["person", "inetOrgPerson"], '
...                         '"userPassword": "secret", '
...                         '"sn": "Mustermann", '
...                         '"cn": "person1"'
...                     '},'
...                     '"class": "node.ext.ldap._node.LDAPNode", '
...                     '"name": "cn=person1"'
...                 '}'
...             '}], '
...             '"class": "node.ext.ldap._node.LDAPNode", '
...             '"name": "ou=container"'
...         '}'
...     '}'
... )

Serialize in simple mode is better readable, but not deserialzable any more:

>>> serialized = serialize(container, simple_mode=True)

>>> assert serialized == (
...     '{'
...         '"attrs": {'
...             '"objectClass": ["organizationalUnit"], '
...             '"ou": "container"'
...         '}, '
...         '"name": "ou=container", '
...         '"children": [{'
...             '"name": "cn=person1", '
...             '"attrs": {'
...                 '"objectClass": ["person", "inetOrgPerson"], '
...                 '"userPassword": "secret", '
...                 '"sn": "Mustermann", '
...                 '"cn": "person1"'
...             '}'
...         '}]'
...     '}'
... )

User and Group management

LDAP is often used to manage Authentication, thus node.ext.ldap provides an API for User and Group management. The API follows the contract of node.ext.ugm:

>>> from node.ext.ldap import ONELEVEL
>>> from node.ext.ldap.ugm import UsersConfig
>>> from node.ext.ldap.ugm import GroupsConfig
>>> from node.ext.ldap.ugm import RolesConfig
>>> from node.ext.ldap.ugm import Ugm

Instantiate users, groups and roles configuration. They are based on PrincipalsConfig class and expect this settings:

baseDN
Principals container base DN.

attrmap
Principals Attribute map as odict.odict. This object must contain the mapping between reserved keys and the real LDAP attribute, as well as mappings to all accessible attributes for principal nodes if instantiated in strict mode, see below.

scope
Search scope for principals.

queryFilter
Search Query filter for principals

objectClasses
Object classes used for creation of new principals. For some objectClasses default value callbacks are registered, which are used to generate default values for mandatory attributes if not already set on principal vessel node.

defaults
Dict like object containing default values for principal creation. A value could either be static or a callable accepting the principals node and the new principal id as arguments. This defaults take precedence to defaults detected via set object classes.

strict
Define whether all available principal attributes must be declared in attmap, or only reserved ones. Defaults to True.

memberOfSupport
Flag whether to use 'memberOf' attribute (AD) or memberOf overlay (openldap) for Group membership resolution where appropriate.

Reserved attrmap keys for Users, Groups and roles:

id
The attribute containing the user id (mandatory).

rdn
The attribute representing the RDN of the node (mandatory) XXX: get rid of, should be detected automatically

Reserved attrmap keys for Users:

login
Alternative login name attribute (optional)

Create config objects:

>>> ucfg = UsersConfig(
...     baseDN='ou=demo,dc=my-domain,dc=com',
...     attrmap={
...         'id': 'cn',
...         'rdn': 'cn',
...         'login': 'sn',
...     },
...     scope=ONELEVEL,
...     queryFilter='(objectClass=person)',
...     objectClasses=['person'],
...     defaults={},
...     strict=False,
... )

>>> gcfg = GroupsConfig(
...     baseDN='ou=demo,dc=my-domain,dc=com',
...     attrmap={
...         'id': 'cn',
...         'rdn': 'cn',
...     },
...     scope=ONELEVEL,
...     queryFilter='(objectClass=groupOfNames)',
...     objectClasses=['groupOfNames'],
...     defaults={},
...     strict=False,
...     memberOfSupport=False,
... )

Roles are represented in LDAP like groups. Note, if groups and roles are mixed up in the same container, make sure that query filter fits. For our demo, different group object classes are used. Anyway, in real world it might be worth considering a seperate container for roles:

>>> rcfg = GroupsConfig(
...     baseDN='ou=demo,dc=my-domain,dc=com',
...     attrmap={
...         'id': 'cn',
...         'rdn': 'cn',
...     },
...     scope=ONELEVEL,
...     queryFilter='(objectClass=groupOfUniqueNames)',
...     objectClasses=['groupOfUniqueNames'],
...     defaults={},
...     strict=False,
... )

Instantiate Ugm object:

>>> ugm = Ugm(props=props, ucfg=ucfg, gcfg=gcfg, rcfg=rcfg)

The Ugm object has 2 children, the users container and the groups container. The are accessible via node API, but also on users respective groups attribute:

>>> ugm.keys()
['users', 'groups']

>>> ugm.users


>>> ugm.groups

Fetch user:

>>> user = ugm.users['person1']
>>> user

User attributes. Reserved keys are available on user attributes:

>>> user.attrs['id']
u'person1'

>>> user.attrs['login']
u'Mustermensch'

'login' maps to 'sn':

>>> user.attrs['sn']
u'Mustermensch'

>>> user.attrs['login'] = u'Mustermensch1'
>>> user.attrs['sn']
u'Mustermensch1'

>>> user.attrs['description'] = 'Some description'
>>> user()

Check user credentials:

>>> user.authenticate('secret')
True

Change user password:

>>> user.passwd('secret', 'newsecret')
>>> user.authenticate('newsecret')
True

Groups user is member of:

>>> user.groups
[]

Add new User:

>>> user = ugm.users.create('person99', sn='Person 99')
>>> user()

>>> res = ugm.users.keys()

>>> assert res == [
...     u'person1',
...     u'person2',
...     u'person3',
...     u'person4',
...     u'person5',
...     u'person99'
... ]

Delete User:

>>> del ugm.users['person99']
>>> ugm.users()

>>> res = ugm.users.keys()

>>> assert res == [
...     u'person1',
...     u'person2',
...     u'person3',
...     u'person4',
...     u'person5'
... ]

Fetch Group:

>>> group = ugm.groups['group1']

Group members:

>>> res = group.member_ids

>>> assert res == [u'person1', u'person2']

>>> group.users
[, ]

Add group member:

>>> group.add('person3')

>>> member_ids = group.member_ids

>>> assert member_ids == [u'person1', u'person2', u'person3']

Delete group member:

>>> del group['person3']

>>> member_ids = group.member_ids

>>> assert member_ids == [u'person1', u'person2']

Group attribute manipulation works the same way as on user objects.

Manage roles for users and groups. Roles can be queried, added and removed via ugm or principal object. Fetch a user:

>>> user = ugm.users['person1']

Add role for user via ugm:

>>> ugm.add_role('viewer', user)

Add role for user directly:

>>> user.add_role('editor')

Query roles for user via ugm:

>>> roles = sorted(ugm.roles(user))

>>> assert roles == ['editor', 'viewer']

Query roles directly:

>>> roles = sorted(user.roles)

>>> assert roles == ['editor', 'viewer']

Call UGM to persist roles:

>>> ugm()

Delete role via ugm:

>>> ugm.remove_role('viewer', user)

>>> roles = user.roles

>>> assert roles == ['editor']

Delete role directly:

>>> user.remove_role('editor')

>>> roles = user.roles

>>> assert roles == []

Call UGM to persist roles:

>>> ugm()

Same with group. Fetch a group:

>>> group = ugm.groups['group1']

Add roles:

>>> ugm.add_role('viewer', group)

>>> group.add_role('editor')

>>> roles = sorted(ugm.roles(group))

>>> assert roles == ['editor', 'viewer']

>>> roles = sorted(group.roles)

>>> assert roles == ['editor', 'viewer']

>>> ugm()

Remove roles:

>>> ugm.remove_role('viewer', group)

>>> group.remove_role('editor')

>>> roles = group.roles

>>> assert roles == []

>>> ugm()

Character Encoding

LDAP (v3 at least, RFC 2251) uses utf-8 string encoding only. LDAPNode does the encoding for you. Consider it a bug, if you receive anything else than unicode from LDAPNode, except attributes configured as binary. The LDAPSession, LDAPConnector and LDAPCommunicator are encoding-neutral, they do no decoding or encoding.

Unicode strings you pass to nodes or sessions are automatically encoded as uft8 for LDAP, except if configured binary. If you feed them ordinary strings they are decoded as utf8 and reencoded as utf8 to make sure they are utf8 or compatible, e.g. ascii.

If you have an LDAP server that does not use utf8, monkey-patch node.ext.ldap._node.CHARACTER_ENCODING.

Caching Support

node.ext.ldap can cache LDAP searches using bda.cache. You need to provide a cache factory utility in you application in order to make caching work. If you don't, node.ext.ldap falls back to use bda.cache.NullCache, which does not cache anything and is just an API placeholder.

To provide a cache based on Memcached install memcached server and configure it. Then you need to provide the factory utility:

>>> from zope.interface import registry

>>> components = registry.Components('comps')

>>> from node.ext.ldap.cache import MemcachedProviderFactory

>>> cache_factory = MemcachedProviderFactory()

>>> components.registerUtility(cache_factory)

In case of multiple memcached backends on various IPs and ports initialization of the factory looks like this:

>>> components = registry.Components('comps')

>>> cache_factory = MemcachedProviderFactory(servers=[
...     '10.0.0.10:22122',
...     '10.0.0.11:22322'
... ])

>>> components.registerUtility(cache_factory)

Dependencies

• python-ldap
• passlib
• argparse
• plumber
• node
• node.ext.ugm
• bda.cache

Contributors

• Robert Niederreiter
• Florian Friesdorf
• Jens Klein
• Georg Bernhard
• Johannes Raggam
• Alexander Pilz
• Domen Kožar
• Daniel Widerin
• Asko Soukka
• Alex Milosz Sielicki
• Manuel Reinhardt
• Philip Bauer

History

1.2 (2022-12-05)

• Implement expires and expired properties on node.ext.ldap.ugm._api.LDAPUser as introduced on node.ext.ugm.interfaces.IUser as of node.ext.ugm 1.1. [rnix]
• Introduce node.ext.ldap.ugm.expires.AccountExpiration and use it for account expiration management. [rnix]
• Remove node.ext.ldap.ugm._api.AccountExpired singleton. LDAPUsers.authenticate always returns False if authentication fails. [rnix]
• node >= 1.1 is required by node.behaviors.suppress_lifecycle_events support [mamico]
• Backward compatibility with pas.plugins.ldap <= 1.8.1 where LdapProps does not have timeout properties. [mamico]

1.1 (2022-10-06)

• Add properties conn_timeout and op_timeout (both not set by default) to configure ReconnectLDAPObject. [mamico]
• Adopt lifecycle related changes from node 1.1. [rnix]
• Move ensure_connection from LDAPSession to LDAPCommunicator to prevent binds on searches that return cached results. [enfold-josh]

1.0 (2022-03-19)

• Call ensure_connection in LDAPSession.delete. [rnix]
• Remove usage of Nodespaces behavior. [rnix]
• Replace deprecated use of Storage by MappingStorage. [rnix]
• Replace deprecated use of IStorage by IMappingStorage. [rnix]
• Replace deprecated use of Nodify by MappingNode. [rnix]
• Replace deprecated use of NodeChildValidate by MappingConstraints. [rnix]
• Replace deprecated use of Adopt by MappingAdopt. [rnix]
• Replace deprecated use of allow_non_node_children by child_constraints. [rnix]

1.0rc2 (2022-03-01)

• Fix #61: Close open connections to LDAP on GC. [jensens]

1.0rc1 (2021-11-08)

• Rename deprecated allow_non_node_childs to allow_non_node_children on PrincipalAliasedAttributes. [rnix]
• Allow to generate MD5 hashes in FIPS enabled environments. [frapell]
• Fix DN comparison in LDAPStorage.node_by_dn to ignore case sensitivity. [rnix]

1.0b12 (2020-05-28)

• Make sure LDAPPrincipals._login_attr has a value. This way LDAPUsers.id_for_login always returns the principal id as stored in the database. [rnix]
• Improve value comparison in LDAPAttributesBehavior.__setitem__ to avoid unicode warnings. [rnix]
• Implement invalidate on node.ext.ldap.ugm._api.Ugm. [rnix]
• Support for group DNs in memberOf attribute that are outside of the UGMs configured group. [jensens]

1.0b11 (2019-09-08)

• Return empty search result list when an LDAP error occurs. Fixes issue #50. [maurits]
• Skip objects that were found in LDAP while searching on several attributes but don't contain the required attribute. [fredvd, maurits]

1.0b10 (2019-06-30)

• Fix cache key generation. [rnix, pbauer]

1.0b9 (2019-05-07)

• Refactor mapping from object-class to format and attributes to increase readability. [jensens]
• Increase Exception verbosity to ease debugging. [jensens]
• Add missing object classes from principal config when persisting principals. [rnix]
• Remove attribute from entry if setting it's value to node.utils.UNSET or empty string. Most LDAP implementations not allow setting empty values, thus we delete the entire attribute in this case. [rnix]
• Add debug-level logging if search fails with no-such-object. [jensens]
• Fix problem with missing LDAP batching cookie in search. [jensens, rnix]
• Remove smbpasswd dependency. Use passlib instead. [rnix]
• Use bytes_mode=False when using python-ldap. This is the default behavior in python 3 and handles everything as unicode/text except entry attribute values. For more details see [rnix]
• Add ensure_bytes_py2 in node.ext.ldap.base. [rnix]
• Rename decode_utf8 to ensure_text in node.ext.ldap.base. [rnix]
• Rename encode_utf8 to ensure_bytes in node.ext.ldap.base. [rnix]
• Python 3 Support. [rnix, reinhardt]
• Convert doctests to unittests. [rnix]

1.0b8 (2018-10-22)

• Use ldap.ldapobject.ReconnectLDAPObject instead of SimpleLDAPObject to create the connection object. This makes the connection more robust. Add properties retry_max (default 1) and retry_delay (default 10) to node.ext.ldap.properties.LDAPServerProperties to configure ReconnectLDAPObject. [joka]
• Use explode_dn in LDAPPrincipals.__getitem__ to prevent KeyError if DN contains comma. [dmunicio]

1.0b7 (2017-12-15)

• Do not catch ValueError in node.ext.ldap._node.LDAPStorage.batched_search. [rnix]
• Use property decorators for node.ext.ldap._node.LDAPStorage.changed and node.ext.ldap.session.LDAPSession.baseDN. [rnix]
• Fix signature of node.ext.ldap.interfaces.ILDAPStorage.search to match the actual implementation in node.ext.ldap._node.LDAPStorage.search. [rnix]
• Fix signature of node.ext.ldap.ugm.LDAPPrincipals.search according to node.ext.ugm.interfaces.IPrincipals.search. The implementation exposed LDAP related arguments and has been renamed to raw_search. [rnix]
• Add exists property to LDAPStorage. [rnix]
• Add objectSid and objectGUID from Active Directory schema to properties.BINARY_DEFAULTS. [rnix]
• Fix default value of LDAPStorage._multivalued_attributes and LDAPStorage._binary_attributes. [rnix]

1.0b6 (2017-10-27)

• Switch to use mdb as default db for slapd i testing layer. [jensens]
• fix tests, where output order could be random. [jensens]

1.0b5 (2017-10-27)

• make db-type in test layer configurable [jensens]

1.0b4 (2017-06-07)

• Turning referrals off to fix problems with MS AD if it contains aliases. [alexsielicki]
• Fix search to check list of binary attributes directly from the root node data (not from attr behavior) to avoid unnecessarily initializing attribute behavior just a simple search [datakurre]
• Fix to skip group DNs outside the base DN to allow users' memberOf attribute contain groups outside the group base DN [datakurre]

1.0b3 (2016-10-18)

• Add a batched_search generator function, which do the actual batching for us. Use this function internally too. [jensens, rnix]
• In testing set size_limit to 3 in slapd.conf in order to catch problems with batching. [jensens, rnix]
• Fix missing paging in UGM group mapping method member_ids. [jensens]

1.0b2 (2016-09-09)

• Minor code cleanup [jensens]
• Paginate LDAP node __iter__. [jensens, rnix]

1.0b1 (31.12.2015)

• Remove ILDAPProps.check_duplicates respective LDAPProps.check_duplicates. [rnix]
• rdn can be queried via attrlist in LDAPNode.search explicitely. [rnix]
• Introduce get_nodes keyword argument in LDAPNode.search. When set, search result contains LDAPNode instances instead of DN's in result. [rnix]
• LDAPNode.search returns DN's instead of RDN's in result. This fixes searches with scope SUBTREE where result items can potentially contain duplicate RDN's. [rnix]
• Introduce node_by_dn on LDAPNode. [rnix]
• remove bbb code: no python 2.4 support (2.7+ now), usage of LDAPProperties mandatory now. [jensens]
• Overhaul LDAP UGM implementation. [rnix]
• LDAP Node only returns direct children in __iter__, even if search scope subtree. [rnix]
• LDAPNode keys cannot be aliased any longer. Removed _key_attr and _rdn_attr. child.
• LDAPNode does not provide secondary keys any longer. Removed _seckey_attrs. [rnix]
• Deprecate node.ext.ldap._node.AttributesBehavior in favor of node.ext.ldap._node.LDAPAttributesBehavior. [rnix]
• Remove deprecated node.ext.ldap._node.AttributesPart. [rnix]
• Don't fail on UNWILLING_TO_PERFORM exceptions when authenticating. That might be thrown, if the LDAP server disallows us to authenticate an admin user, while we are interested in the local admin user. [thet]
• Add ignore_cert option to ignore TLS/SSL certificate errors for self signed certificates when using the ldaps uri schema. [thet]
• Housekeeping. [rnix]

0.9.7

• Added possibility to hook external LDIF layer for testldap server via buildout configuration. [rnix]
• Update openldap version in buildout configs. [rnix]

0.9.6

• Add new property to allow disable check_duplicates. This avoids following Exception when connecting ldap servers with non-unique attributes used as keys. [saily] :

  Traceback (most recent call last):
  ...
  RuntimeError: Key not unique: =''.
  

• ensure attrlist values are strings [rnix, 2013-12-03]

0.9.5

• Add expired property to node.ext.ldap.ugm._api.LDAPUser. [rnix, 2012-12-17]
• Introduce node.ext.ldap.ugm._api.calculate_expired helper function. [rnix, 2012-12-17]
• Lookup expired attribut from LDAP in node.ext.ldap.ugm._api.LDAPUser.authenticate. [rnix, 2012-12-17]

0.9.4

• Encode DN in node.ext.ldap._node.LDAPStorage._ldap_modify. [rnix, 2012-11-08]
• Encode DN in node.ext.ldap._node.LDAPStorage._ldap_delete. [rnix, 2012-11-08]
• Encode DN in node.ext.ldap.ugm._api.LDAPUsers.passwd. [rnix, 2012-11-08]
• Encode DN in node.ext.ldap.ugm._api.LDAPUsers.authenticate. [rnix, 2012-11-07]
• Encode baseDN in LDAPPrincipal.member_of_attr. [rnix, 2012-11-06]
• Encode baseDN in AttributesBehavior.load. [rnix, 2012-11-06]
• Python 2.7 compatibility. [rnix, 2012-10-16]
• PEP-8. [rnix, 2012-10-16]
• Fix LDAPPrincipals.idbydn handling UTF-8 DN's properly. [rnix, 2012-10-16]
• Rename parts to behaviors. [rnix, 2012-07-29]
• adopt to node 0.9.8. [rnix, 2012-07-29]
• Adopt to plumber 1.2. [rnix, 2012-07-29]
• Do not convert cookie to unicode in LDAPSession.search. Cookie value is no utf-8 string but octet string as described in . [rnix, 2012-07-27]
• Add User.group_ids. [rnix, 2012-07-26]

0.9.3

• Fix schema to not bind to test BaseDN only and make binding deferred. [jensens, 2012-05-30]

0.9.2

• Remove escape_queries property from node.ext.ldap.properties.LDAPProps. [rnix, 2012-05-18]
• Use zope.interface.implementer instead of zope.interface.implements. [rnix, 2012-05-18]
• Structural object class inetOrgPerson instead of account on posix users and groups related test LDIF's [rnix, 2012-04-23]
• session no longer magically decodes everything and prevents binary data from being fetched from ldap. LDAP-Node has semantic knowledge to determine binary data LDAP-Node converts all non binary data and all keys to unicode. [jensens, 2012-04-04]
• or_values and or_keys for finer control of filter criteria [iElectric, chaoflow, 2012-03-24]
• support paged searching [iElectric, chaoflow, 2012-03-24]

0.9.1

• added is_multivalued to properties and modified node to use this list instead of the static list. prepare for binary attributes. [jensens, 2012-03-19]
• added schema_info to node. [jensens, 2012-03-19]
• shadowInactive defaults to 0. [rnix, 2012-03-06]
• Introduce expiresAttr and expiresUnit in principals config. Considered in Users.authenticate. [rnix, 2012-02-11]
• Do not throw KeyError if secondary key set but attribute not found on entry. In case, skip entry. [rnix, 2012-02-10]
• Force unicode ids and keys in UGM API. [rnix, 2012-01-23]
• Add unicode support for filters. [rnix, 2012-01-23]
• Add LDAPUsers.id_for_login. [rnix, 2012-01-18]
• Implement memberOf Support for openldap memberof overlay and AD memberOf behavior. [rnix, 2011-11-07]
• Add LDAPProps.escape_queries for ActiveDirectory. [rnix, 2011-11-06]
• Add group object class to member attribute mapping for ActiveDirectory. [rnix, 2011-11-06]
• Make testlayer and testldap more flexible for usage outside this package. [jensens, 2010-09-30]

0.9

• refactor form bda.ldap. [rnix, chaoflow]

TODO

• Consider search_st with timeout.
• Investigate ReconnectLDAPObject.set_cache_options.
• Check/implement silent sort on only the keys LDAPNode.sortonkeys.
• Interactive configuration showing live how many users/groups are found with the current config and what a selected user/group would look like.
• Configuration validation for UGM. Add some checks in Ugm.__init__ which tries to block stupid configuration.
• Group in group support.
• Rework ldap testsetup to allow for multiple servers in order to test with different overlays it would be nice to start different servers or have one server with multiple databases. whatever feels better.
• Rework tests and ldifs to target isolated aspects.
• Potentially multi-valued attrs always as list.

License

Copyright (c) 2006-2021, BlueDynamics Alliance, Austria, Germany, Switzerland Copyright (c) 2021-2022, Node Contributors All rights reserved.

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

• Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
• Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.