The googleads Python client library
Note: DoubleClick for Publishers was recently renamed to Google Ad Manager, and this library was modified to follow suit. Please see the migration guide for more details.
This client library simplifies accessing Google's SOAP Ads APIs - AdWords, and DoubleClick for Publishers. The library provides easy ways to store your authentication and create SOAP web service clients. It also contains example code to help you get started integrating with our APIs.
Download and install the library
setuptools is a pre-requisite for installing the googleads library
It is recommended that you install the library and its dependencies from PyPI using pip. This can be accomplished with a single command:
$ pip install googleads
As an alternative, you can download the library as a tarball. To start the installation, navigate to the directory that contains your downloaded unzipped client library and run the "setup.py" script as follows:
$ python setup.py build install
Copy the googleads.yaml file to your home directory.
This will be used to store credentials and other settings that can be loaded to initialize a client.
Set up your OAuth2 credentials
The AdWords and DoubleClick for Publishers APIs use OAuth2 as the authentication mechanism. Follow the appropriate guide below based on your use case.
If you're accessing an API using your own credentials...
If you're accessing an API on behalf of clients...
Where can I find samples?
You can find code examples for the latest versions of AdWords or Ad Manager on the releases page.
Where can I find the pydocs?
Our pydocs can be found here.
Caching authentication information
It is possible to cache your API authentication information. The library
includes a sample file showing how to do this named
in the fields for the API and features you plan to use. The library's
LoadFromStorage methods default to looking for a file with this name in your
home directory, but you can pass in a path to any file with the correct yaml
# Use the default location - your home directory: adwords_client = adwords.AdWordsClient.LoadFromStorage() # Alternatively, pass in the location of the file: ad_manager_client = ad_manager.AdManagerClient.LoadFromStorage('C:\My\Directory\googleads.yaml')
How do I change the Client Customer Id at runtime?
You can change the Client Customer Id with the following:
adwords_client = AdWordsClient.LoadFromStorage() adwords_client.SetClientCustomerId('my_client_customer_id')
Where do I submit bug reports and/or feature requests?
If you have issues directly related to the client library, use the issue tracker.
If you have issues pertaining to a specific API, use the product support forums:
How do I log SOAP interactions?
The library uses Python's built in logging framework. You can specify your configuration via the configuration file; see googleads.yaml for an example.
Alternatively, you can manually specify your logging configuration. For example, if you want to log your SOAP interactions to stdout, and you are using the Zeep soap library, you can do the following:
logging.basicConfig(level=logging.INFO, format=googleads.util.LOGGER_FORMAT) logging.getLogger('googleads.soap').setLevel(logging.DEBUG)
If you are using suds, you can achieve the same like this:
logging.basicConfig(level=logging.INFO, format=googleads.util.LOGGER_FORMAT) logging.getLogger('suds.transport').setLevel(logging.DEBUG)
If you wish to log to a file, you'll need to attach a log handler to this source which is configured to write the output to a file.
How do I disable log filters?
The zeep plugin used for logging strips sensitive data from its output. If you would like this data included in logs, you'll need to implement your own simple logging plugin. For example:
class DangerousZeepLogger(zeep.Plugin): def ingress(self, envelope, http_headers, operation): logging.debug('Incoming response: \n%s', etree.tostring(envelope, pretty_print=True)) return envelope, http_headers def egress(self, envelope, http_headers, operation, binding_options): logging.debug('Incoming response: \n%s', etree.tostring(envelope, pretty_print=True)) return envelope, http_headers adwords_client.zeep_client.plugins.append(DangerousZeepLogger())
When using suds, this library will apply log filters to the
suds.transport loggers in order to omit sensitive data. If
you need to see this data in your logs, you can disable the filters with the
logging.getLogger('googleads.common').removeFilter( googleads.util.GetGoogleAdsCommonFilter()) logging.getLogger('suds.client').removeFilter( googleads.util.GetSudsClientFilter()) logging.getLogger('suds.mx.core').removeFilter( googleads.util.GetSudsMXCoreFilter()) logging.getLogger('suds.mx.literal').removeFilter( googleads.util.GetSudsMXLiteralFilter()) logging.getLogger('suds.transport').removeFilter( googleads.util.GetSudsTransportFilter())
I'm familiar with suds/zeep. Can I use those features in the library?
Yes, you can. The services returned by the
client.GetService() functions all
have a reference to the underlying client stored in the
attributes. You can retrieve the client and use it in familiar ways:
client = AdWordsClient.LoadFromStorage() campaign_service = client.GetService('CampaignService') suds_client = campaign_service.suds_client campaign = suds_client.factory.create('Campaign') # Set any attributes on the campaign object which you need. campaign.name = 'My Campaign' campaign.status = 'PAUSED' operation = suds_client.factory.create('CampaignOperation') operation.operator = 'ADD' operation.operand = campaign # The service object returned from the client.GetService() call accepts suds # objects and will set the SOAP and HTTP headers for you. campaign_service.mutate([operation]) # Alternatively, if you wish to set the headers yourself, you can use the # suds_client.service directly. soap_header = suds_client.factory.create('SoapHeader') soap_header.clientCustomerId = client.client_customer_id soap_header.developerToken = client.developer_token soap_header.userAgent = client.user_agent suds_client.set_options( soapheaders=soap_header, headers=client.oauth2_client.CreateHttpHeader()) suds_client.service.mutate([operation])
How can I configure or disable caching?
By default, clients are cached because reading and digesting the WSDL can be expensive. However, the default caching method requires permission to access a local file system that may not be available in certain hosting environments such as App Engine.
You can pass an implementation of
zeep.cache.Base to the
AdManagerClient initializer to modify the default caching behavior.
For example, configuring a different location and duration of the cache file with zeep
doc_cache = zeep.cache.SqliteCache(path=cache_path) adwords_client = adwords.AdWordsClient( developer_token, oauth2_client, user_agent, client_customer_id=client_customer_id, cache=doc_cache)
And with suds:
doc_cache = suds.cache.DocumentCache(location=cache_path, days=2) adwords_client = adwords.AdWordsClient( developer_token, oauth2_client, user_agent, client_customer_id=client_customer_id, cache=doc_cache)
You can also disable caching in similar fashion with zeep
adwords_client = adwords.AdWordsClient( developer_token, oauth2_client, user_agent, client_customer_id=client_customer_id, cache=googleads.common.ZeepServiceProxy.NO_CACHE)
And with suds:
adwords_client = adwords.AdWordsClient( developer_token, oauth2_client, user_agent, client_customer_id=client_customer_id, cache=suds.cache.NoCache())
This library supports both Python 2 and 3. To use this library, you will need to have Python 2.7.9 (or higher) or Python 3.4 (or higher) installed.
- httplib2 -- https://pypi.python.org/pypi/httplib2/ - oauth2client -- https://pypi.python.org/pypi/oauth2client/ - suds-jurko -- https://pypi.python.org/pypi/suds-jurko/ - pysocks -- https://pypi.python.org/pypi/PySocks/ - pytz -- https://pypi.python.org/pypi/pytz - pyYAML -- https://pypi.python.org/pypi/pyYAML/ - xmltodict -- https://pypi.python.org/pypi/xmltodict/ - zeep -- https://pypi.python.org/pypi/zeep - mock -- https://pypi.python.org/pypi/mock (only needed to run unit tests) - pyfakefs -- https://pypi.python.org/pypi/pyfakefs (only needed to run unit tests)
Mark Saniscalchi David Wihl Ben Karl