Mock Directories

The first step to using mockldap is to define some static LDAP content and install it as a mock LDAPObject. For this, you will use mockldap.MockLdap. Only one instance of this class should exist at a time; setUpClass() is a good place to instantiate it.

MockLdap can mock multiple LDAP directories, identified by URI. You can provide directory content for URIs individually and you can also provide default content for connections to any unrecognized URI. If the code under test is only expected to make one LDAP connection, the simplest option is just to provide default content. If you need multiple directories, you can call set_directory() on your MockLdap instance.

LDAP content takes the form of a Python dictionary. Each key is a distinguished name in string form; each value is a dictionary mapping attributes to lists of values. In other words, directory.items() should take the same form as results from search_s().

directory = {
    'uid=alice,ou=people,o=test': {
        'uid': ['alice'],
        'objectClass': ['person', 'organizationalPerson', 'inetOrgPerson', 'posixAccount'],
        'userPassword': ['password'],
        'uidNumber': ['1000'],
        'gidNumber': ['1000'],
        'givenName': ['Alice'],
        'sn': ['Adams']
    'cn=active,ou=groups,o=test': {
        'cn': ['active'],
        'objectClass': ['groupOfNames'],
        'member': ['uid=alice,ou=people,o=test']

MockLdap is stateful. The overview shows a complete example, but following are the enumerated steps.

For some collection of tests:

  • Instantiate MockLdap. Optionally pass in default directory contents.
  • Add content for any additional directories. This is only necessary if the code under test will connect to multiple LDAP directories.

For each test:

  • Just before an individual test, call start(). This will instantiate your mock directories and patch ldap.initialize(). You may need to call this multiple times if initialize() is accessed by multiple names.
  • Any time during your test, you can access an individual LDAPObject as mockldap[uri]. This will let you seed return values for LDAP operations and recover the record of which operations were performed.
  • After the test, call stop() or stop_all().


The code under test must not keep an LDAP “connection” open across individual test cases. If it does, it will be sharing a mock LDAPObject across tests, so any state mutations will persist.


class mockldap.MockLdap(directory=None)[source]

Top-level class managing directories and patches.

Parameters:directory – Default directory contents.

After calling start(), mockldap[uri] returns an LDAPObject. This is the same object that will be returned by ldap.initialize(uri), so you can use it to seed return values and discover which APIs were called.

MockLdap objects may be used as context managers in place of calling start() and stop() manually. The import path for ldap.initialize can not be overridden in this case.

set_directory(directory, uri='__default__')[source]

Set the mock LDAP content for a given URI.

Parameters:uri (string) – The LDAP URI to associate this content with.

If URI is not given, this will set the default content for all unknown URIs.


Patch ldap.initialize() to return mock LDAPObject instances.

This calls mock.patch(), so under Python 2, you must have the mock library installed.

Parameters:path (string) – The module path to ldap.initialize.

If the code under test looks like:

import ldap

then you can use the default value of path. If the code reads:

from ldap import initialize

then you need to call start(''). See Where to patch for more.


Stop patching ldap.initialize().

Calls to start() and stop() must be balanced. After the final call to stop, we’ll reset all LDAPObject instances.


Remove all patches and reset our state.

If you called start() multiple times, this is the easiest way to reset everything.