Python-Zimbra

Python classes to access Zimbra SOAP backend with a few utilities. Handles creating and sending Zimbra SOAP queries to the backend and adds a few utilities such as a preauth generator.

Compatible with Python 2.7 and 3.x (cPython and PyPy)

This framework is not intended to supply user level access to Zimbra functions. Zimbra tends to be too dynamic and complex for this to work. This is actually a framework to reduce the implementation work if you want to speak to the Zimbra SOAP API.

Please refer to the official SOAP documentation by Zimbra on how to use the SOAP backend.

For details refer to the API-Documentation.

Warning

There are missing sanity checks on purpose. This is because the Zimbra API is due to heavy modifications and to keep up with the current API catalogue is quiet problematic.

So please be aware of this and use the library with caution. Don't expose the library to the public without doing sanity checks on your own.

Tutorial

This tutorial will explain you how to get information about a folder of a specific user. this is how you do it in a few steps:

• Installation
• Imports
• Communication and Token
• Creating your request
• Fault handling
• Batch

Installation

To install, either use :

pip install python-zimbra

or

easy_install python-zimbra

You may add --pre to download the pre-release

Imports

First, import the needed libraries (You will learn about them later):

import pythonzimbra.communication
from pythonzimbra.communication import Communication
import pythonzimbra.tools
from pythonzimbra.tools import auth

Setup

Now, build up the communication object to send the future requests:

# url to your zimbra server
url  = 'https://your-zimbra-server/service/soap'
comm = Communication(url)

Then, we will have to authenticate. Using Zimbra preauth, we can do the following:

usr_token = auth.authenticate(
    url,
    'myuser@mydomain.com',
    'secret-preauth-key'
)

This will return None on failure. If it is so, please check the parameters again. Otherwise, you will get your authentication token for your Zimbra requests.

Request

Time to create our request. This is how to get the content of a folder (located in urn:zimbraMail-namespace).

Here is how initiating the request, adding a request then send it and gather its response.

info_request = comm.gen_request(token=usr_token)
info_request.add_request(
    'GetFolderRequest',
    {
        'folder': {
            'path': '/inbox'
        }
    },
    'urn:zimbraMail'
)
info_response = comm.send_request(info_request)

Now, if the response was successful, we can print the result:

if not info_response.is_fault():
    print (info_response.get_response()['GetFolderResponse']['folder']['n'])

Fault

To verify if there is any fault in your response:

response.is_fault()

If a result is fault, you can get specific information about it.

response.get_fault_code()
response.get_fault_message()

• The fault_code is Zimbra's own fault message code (like mail.NO_SUCH_FOLDER).
• The message is a more elaborate message like (no such folder path: /...).

Batch

Working with batch requests is also possible. To do that, set the parameter "set_batch":

batch_request = comm.gen_request(set_batch=True)

And can afterwards add multiple requests using add_request to it. You'll get the request id of the specific request as a return value. Use that id to retrieve the response later using get_response(id).

Authentication against the administration console

Zimbra currently doesn't support the preauth-method for authentications against the admin-console (URL https://your-zimbra-server:7071/service/admin/soap).

python-zimbra's auth tool can be used to authenticate to this url by specifying the password instead of the preauth-key and setting the parameter admin_auth to True. (see API docs for specifics)

Used dictionary

All requests and responses are built up using a certain dictionary format. This is heavily influenced by the Zimbra json format being:

{
    "RequestName": {
        "_content": "Content of the node",
        "attribute": "value",
        "subnode": {
            "_content": "Content of the subnode"
        }
    }
}

in XML this would look like this:

<RequestName attribute="value">
    <subnode>
        Content of the subnode
    </subnode>
    Content of the node
</RequestName>

All requests should conform to this dictionary format and the responses are also returned in this format. Subnodes can also contain lists of dictionaries, which will create multiple subnodes with the same tag.

Use a Custom SSL Context

In test environments sometimes SSL is not available and we want to disable SSL verification.

To achieve this we can use a custom SSL context:

Not yet released on pypi, so put in your requirements.txt:

-e git+https://github.com/Zimbra-Community/python-zimbra.git#egg=python-zimbra

Example:

import ssl

context = ssl.SSLContext()

...
comm = Communication(url, context=context)

usr_token = auth.authenticate(
    url=url,
    ...,
    context=context
)

Testing

Python-Zimbra includes a testsuite with unittests, that test the supported features.

To enable testing in your environment, copy the config.ini.dist to config.ini in the tests module and configure it to match your environment.

You may need a Zimbra server with an admin and a user account to run all tests. You have to specifically enable these tests.

Test overview:

• test_admin.py
  • Authenticate as admin
  • Add a test account
  • Try logging in using that test account
  • Delete the test account
• test_auth.py
  • Authenticate as user
  • Authenticate as user with wrong preauth key
  • Authenticate as user with password
  • Authenticate as user with wrong password
• test_autoresponse.py
  • Authenticate as user
  • Send NoOpRequest
• test_fault.py
  • Authenticate as user
  • Query a non-existing folder using GetFolderRequest
  • Query a non-existing folder using GetFolderRequest inside a BatchRequest
• test_genrequest.py
  • Authenticate as user
  • Send a NoOpRequest
  • Send a NoOpRequest inside a BatchRequest
  • Send a GetInfoRequest
  • Send a NoOpRequest and a GetInfoRequest inside a BatchRequest

To run the test, enter the tests subdirectory and run

python -m unittest discover -s ..

We thankfully use Travis for continuous integration and Coveralls for code coverage.

The Zimbra server used in CI-testing is kindly hosted by efm.