Creating a simple LDAP application

Tommi Virtanen

(presented on 2004-06-09)

LDAP presents a distributed tree of information

Writing things down: LDIF

dn: gn=John+sn=Doe,ou=Research & Development,ou=Pe
objectClass: addressbookPerson
gn: John
sn: Doe
street: Back alley
postOfficeBox: 123
postalCode: 54321
postalAddress: Backstreet
st: NY
l: New York City
c: US

Writing things down: LDIF (2)

dn: gn=John+sn=Smith,ou=Marketing,ou=People,
objectClass: addressbookPerson
gn: John
sn: Smith
telephoneNumber: 555-1234
facsimileTelephoneNumber: 555-1235
description: This is a description that can span multi
 ple lines as long as the non-first lines are inden
 ted in the LDIF.

Setting up an LDAP server in 5 seconds


Python, an easy programming language

Batteries included!

Python combines remarkable power with very clear syntax.

Runs on many brands of UNIX, on Windows, OS/2, Mac, Amiga, and many other platforms.

The first step

>>> from ldaptor.protocols.ldap import distinguishedname
>>> dn=distinguishedname.DistinguishedName(
... 'dc=example,dc=com')
>>> dn
attributeType='dc', value='example'),)),
LDAPAttributeTypeAndValue(attributeType='dc', value='com'),))))
>>> str(dn)


Ldaptor is a set of pure-Python LDAP client programs, applications and a programming library.

It is licensed under the GNU LGPL.

Overview of Ldaptor

Preparing to connect

>>> from ldaptor.protocols.ldap import \
... ldapclient, ldapconnector
>>> from twisted.internet import reactor
>>> connector=ldapconnector.LDAPClientCreator(reactor,
... ldapclient.LDAPClient)
>>> connector
instance at 0x40619b6c>


Twisted is an event-driven networking framework written in Python and licensed under the LGPL.

Twisted supports TCP, UDP, SSL/TLS, multicast, Unix sockets, a large number of protocols (including HTTP, NNTP, SSH, IRC, FTP, and others), and much more.

Twisted includes many fullblown applications, such as web, SSH, FTP, DNS and news servers.


>>> d=connector.connectAnonymously(dn,
... {dn: ('localhost', 10389)})
>>> d
<Deferred at 0x402d058c>


  • A promise that a function will at some point have a result.

  • You can attach callback functions to a Deferred. Once it gets a result these callbacks will be called.

  • Also allows you to register a callback for an error, with the default behavior of logging the error.

  • Standard way to handle all sorts of blocking or delayed operations.


>>> from twisted.trial.util import deferredResult
>>> proto=deferredResult(d)
>>> proto
instance at 0x40619dac>
>>> from ldaptor.protocols.ldap import ldapsyntax
>>> baseEntry=ldapsyntax.LDAPEntry(client=proto, dn=dn)
>>> results=deferredResult(d2)


>>> results
dc=example,dc=com', attributes={'description': ['Some text.'],
'facsimileTelephoneNumber': ['555-1235'], 'givenName': ['John'],
'objectClass': ['addressbookPerson'], 'sn': ['Smith'],
'telephoneNumber': ['555-1234']}), LDAPEntry(dn=
attributes={'c': ['US'], 'givenName': ['John'], 'l': ['New York City'],
'objectClass': ['addressbookPerson'], 'postOfficeBox': ['123'],
'postalAddress': ['Backstreet'], 'postalCode': ['54321'],
'sn': ['Doe'], 'st': ['NY'], 'street': ['Back alley']})]

Results one-by-one

>>> results[0]
attributes={'description': ['Some text.'],
'facsimileTelephoneNumber': ['555-1235'], 'givenName': ['John'],
'objectClass': ['addressbookPerson'], 'sn': ['Smith'],
'telephoneNumber': ['555-1234']})
>>> results[3]
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
IndexError: list index out of range

LDIF output

>>> print results[0]
dn: givenName=John+sn=Smith,ou=People,dc=example,dc=com
objectClass: addressbookPerson
description: Some text.
facsimileTelephoneNumber: 555-1235
givenName: John
sn: Smith
telephoneNumber: 555-1234

Closing the connection

>>> proto.unbind()

Access to entry details

>>> smith=results[0]
>>> print smith.dn
>>> smith['givenName']

Object-oriented look at LDAP entries

A lot of similarities with OO programming languages, but some big differences, too.

  • An LDAP entry corresponds with an object.

  • Whereas object are usually instances of a single class, LDAP entries can "implement" multiple objectClasses.

  • objectClasses can inherit zero, one or many objectClasses, just like programming classes.

  • objectClasses have a root class, known as top; many object oriented programming languages have a root class, e.g. named Object.

  • objectClasses are either STRUCTURAL or AUXILIARY; entries can only implement one STRUCTURAL objectClass.

  • The objectClasses of an entry can be changed at will; you only need to take care that the entry has all the MUST attribute types, and no attribute types outside of the ones that are MUST or MAY.

    Note that e.g. OpenLDAP doesn't implement this.

  • Attributes of an entry closely match attributes of objects in programming languages; however, LDAP attributes may have multiple values.

Search inputs

  • The base DN for the search

  • A search filter, specifying criteria an entry must fulfill to match.

  • Scope of the search, either look at the base DN only, only look one level below it, or look at the whole subtree rooted at the base DN.

  • Size limit of at most how many matching entries to return.

  • Attributes to return, or none for all attributes the matching entries happen to have.

An example search filter:

(cn=John Smith)

Our first Python program

from twisted.internet import reactor, defer

from ldaptor.protocols.ldap import ldapclient, ldapsyntax, ldapconnector, \
from ldaptor import ldapfilter

def search(config):
    c=ldapconnector.LDAPClientCreator(reactor, ldapclient.LDAPClient)

    def _doSearch(proto, config):
        searchFilter = ldapfilter.parseFilter('(gn=j*)')
        baseEntry = ldapsyntax.LDAPEntry(client=proto, dn=config['base'])
        return d

    d.addCallback(_doSearch, config)
    return d

def main():
    import sys
    from twisted.python import log
    log.startLogging(sys.stderr, setStdout=0)

    config = {
        'serviceLocationOverrides': {
            ('localhost', 10389),

    d = search(config)
    def _show(results):
        for item in results:
            print item
    d.addBoth(lambda _: reactor.stop())

if __name__ == '__main__':

Phases of the protocol chat

  1. Open and bind
  2. Search (possibly many times)
  3. Unbind and close

Opening and binding

Doing multiple searches

Unbinding and closing


  • A web application framework for building highly interactive web applications.

  • Separates HTML templates from page-generation logic.

  • Uses the Model-View-Controller (MVC) pattern to create dynamic HTML on the fly.

A Web App: Code

import os
from twisted.internet import reactor
from nevow import rend, appserver, inevow, compy, \
     stan, loaders
from formless import annotate, webform, iformless

from ldaptor.protocols.ldap import ldapclient, ldapsyntax, ldapconnector, \
from ldaptor import ldapfilter
from ldaptor.protocols import pureldap

class ILDAPConfig(compy.Interface):
    """Addressbook configuration retrieval."""

    def getBaseDN(self):
        """Get the LDAP base DN, as a DistinguishedName."""

    def getServiceLocationOverrides(self):
        Get the LDAP service location overrides, as a mapping of
        DistinguishedName to (host, port) tuples.

class LDAPConfig(object):
    __implements__ = ILDAPConfig

    def __init__(self,
        self.baseDN = distinguishedname.DistinguishedName(baseDN)
        self.serviceLocationOverrides = {}
        if serviceLocationOverrides is not None:
            for k,v in serviceLocationOverrides.items():
                dn = distinguishedname.DistinguishedName(k)

    def getBaseDN(self):
        return self.baseDN

    def getServiceLocationOverrides(self):
        return self.serviceLocationOverrides

class IAddressBookSearch(annotate.TypedInterface):
    def search(self,
               sn = annotate.String(label="Last name"),
               givenName = annotate.String(label="First name"),
               telephoneNumber = annotate.String(),
               description = annotate.String()):
    search = annotate.autocallable(search)

class CurrentSearch(object):
    __implements__ = IAddressBookSearch, inevow.IContainer
    data = {}

    def _getSearchFilter(self):
        filters = []
        for attr,value in
            if value is not None:
                f = ldapfilter.parseMaybeSubstring(attr, value)
        if not filters:
            return None
        searchFilter = pureldap.LDAPFilter_and(
            + filters)
        return searchFilter

    def search(self, **kw):
        for k,v in kw.items():
            if v is None:
                del kw[k] = kw
        return self

    def __nonzero__(self):
        return bool(

    def __iter__(self):
        if is None:
        for k,v in
            yield (k,v)

    def child(self, context, name):
        if name == 'searchFilter':
            return self._getSearchFilter()
        if name != 'results':
            return None
        config = context.locate(ILDAPConfig)

        c=ldapconnector.LDAPClientCreator(reactor, ldapclient.LDAPClient)

        def _search(proto, base, searchFilter):
            baseEntry = ldapsyntax.LDAPEntry(client=proto, dn=base)
            return d

        d.addCallback(_search, config.getBaseDN(), self._getSearchFilter())
        return d

def LDAPFilterSerializer(original, context):
    return original.asText()

# TODO need to make this pretty some day.
for c in [

class AddressBookResource(rend.Page):
    docFactory = loaders.xmlfile(

    def configurable_(self, context):
            i = context.locate(inevow.IHand)
        except KeyError:
            i = CurrentSearch()
        return i

    def data_search(self, context, data):
        configurable = self.locateConfigurable(context, '')
        cur = configurable.original
        return cur

    def child_form_css(self, request):
        return webform.defaultCSS

    def render_input(self, context, data):
        formDefaults = context.locate(iformless.IFormDefaults)
        methodDefaults = formDefaults.getAllDefaults('search')
        conf = self.configurable_(context)
        for k,v in conf:
            methodDefaults[k] = v
        return webform.renderForms()

    def render_haveSearch(self, context, data):
        return context.tag.clear()[r]

    def render_searchFilter(self, context, data):
        return data.asText()

    def render_iterateMapping(self, context, data):
        headers = context.allPatterns('header')
        keyPattern = context.patternGenerator('key')
        valuePattern = context.patternGenerator('value')
        divider = context.patternGenerator('divider', default=stan.invisible)
        content = [(keyPattern(data=key),
                   for key, value in data.items()]
        if not content:
            content = context.allPatterns('empty')
            # No divider after the last thing.
            content[-1] = content[-1][:-1]
        footers = context.allPatterns('footer')

        return context.tag.clear()[ headers, content, footers ]

def getSite(config):
    form = AddressBookResource()
    form.remember(config, ILDAPConfig)
    site = appserver.NevowSite(form)
    return site

A Web App: Template

<?xml version="1.0"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"

<html xmlns=""
    <title>Address book search</title>
    <link href="/form_css" rel="stylesheet" />
    <h1>Address book search</h1>

    <form nevow:render="input" action="post">
      <p nevow:render="Remove">Search form will be here.</p>

    <div nevow:render="haveSearch" nevow:data="search">
      <p nevow:pattern="False">No search given.</p>
      <div nevow:pattern="True">
	<ol nevow:render="sequence" nevow:data="results">
	  <p nevow:pattern="empty">No entries found.</p>
	  <li nevow:pattern="item" style="margin-bottom: 2em;">
	    <dl nevow:render="iterateMapping">
	      <dt nevow:pattern="key" nevow:render="string">
		Attribute type goes here
	      <nevow:invisible nevow:pattern="value" nevow:render="sequence">
		<dd nevow:pattern="item" nevow:render="string">
		  Attribute value goes here
	  Used search filter:
	  <span nevow:render="searchFilter" nevow:data="searchFilter"/>

A Web App: Startup

# -*- python -*-
from twisted.application import service, internet

import addressbook

config = addressbook.LDAPConfig(
    'dc=example,dc=com': ('localhost', 10389),

application = service.Application("LDAPressBook")
site = addressbook.getSite(config)
webServer = internet.TCPServer(8088, site)

A complex search filter



  • Special attribute objectClass lists all the objectclasses an LDAP entry manifests.

  • Objectclass defines

    • what attributetypes an entry MUST have
    • what attributetypes an entry MAY have
  • An entry in a phonebook must have a name and a telephone number, and may have a fax number and street address.


  • a configuration file included in the LDAP server configuration.

  • a combination of attribute type and object class definitions.

  • stored as plain text

  • can be requested over an LDAP connection

Attribute type

attributetype ( NAME ( 'sn' 'surname' )
	DESC 'RFC2256: last (family) name(s) for 
			  which the entity is known by'
	SUP name )

Can also contain

  • content data type
  • comparison and sort mechanism
  • substring search mechanism
  • whether multiple values are allowed

Object class

objectclass ( NAME 'person'
	DESC 'RFC2256: a person'
	MUST ( sn $ cn )
	MAY ( userPassword $ telephoneNumber
		$ seeAlso $ description ) )

Creating schemas

  • Anyone can create their own schema
  • Need to be globally unique
  • But try to use already existing ones

Demo of ldaptor-webui


Where to go from here?

Install OpenLDAP .

Install Ldaptor , play around with ldaptor-webui.

Learn Python .

Learn Twisted. Write a client application for a simple protocol. Read the HOWTOs .

Thank You