Python Library for accessing the Springserve API
For more information on using the API, please refer to our Wiki:
https://wiki.springserve.com/display/SSD/API+-+Getting+Started
Installation
To install from source:
python setup.py install
To install from pip:
pip install springserve
To use the Reporting API <https://springserve.atlassian.net/wiki/spaces/SSD/pages/1588035603/Reporting+API>__
with this package, also install pandas <https://pypi.org/project/pandas/>__.
Springserve is using link to handle it's configuration. Link is a way to centrally configure your database, api handles. It has support for Springserve api connections. For more, see the link documentation. https://link-docs.readthedocs.org/en/latest/
Link will be installed when you install springserve
To configure link for springserve:
Open ipython and run the following. This will edit your link.config. By default this will be ~/.link/link.config. You can change this directory location by setting the environment variable LNK_DIR
Run the following to set up your config:
In [1]: import springserve
In [2]: springserve.setup_config()
Enter a user name: {enter email here}
Enter password:
Would you like write[Y/n] y
writing config to: /Users/{username}/.link/link.config
done: refreshing config
You can now configure your username and password directly in code instead of using link. Here is an example:
In [1]: import springserve
In [2]: springserve.set_credentials(user='blah@blah.com', password="mypassword")
If you have many users and you want to store their tokens and inject the token in for each call you can do that using set_token. Here is an example:
In [1]: import springserve
In [2]: springserve.set_token('56385e9ae7ef6cc35e5c34153f479b858')
The python library was built to work seamlessly with tools like IPython. IPython is an interactive shell that is well suited for adhoc data analysis as well as general python debugging. One of it's best features is tab completion and documentation
In [1]: import springserve
In [2]: springserve.<tab>
springserve.API springserve.demand_tags springserve.domain_lists springserve.raw_get springserve.supply_tags
# see documentation on the get function of supply_tags
In [3]: springserve.supply_tags.get?
Signature: springserve.supply_tags.get(path_param=None, reauth=False, **query_params)
Docstring:
Make a get request to this api service. Allows you to pass in arbitrary query paramaters.
Examples::
# get all supply_tags
tags = springserve.supply_tags.get()
for tag in tags:
print tag.id, tag.name
# get one supply tag
tag = springserve.supply_tag.get(1)
print tag.id, tag.name
# get by many ids
tags = springserve.supply_tags.get(ids=[1,2,3])
# get users that are account_contacts (ie, using query string # params)
users = springserve.users.get(account_contact=True)
File: /usr/local/lib/python2.7/site-packages/springserve/__init__.py
Type: instancemethod
# get a supply_tag by it's id
In [4]: tag = springserve.supply_tags.get(1234)
# see what fields exist on the supply_tag
In [5]: tag.<tab>
tag.active tag.domain_list_ids tag.player_size_targeting
tag.supply_partner_id
tag.allowed_player_sizes tag.domain_targeting tag.rate
tag.supply_type
tag.country_codes tag.id tag.raw
tag.country_targeting tag.name tag.save
tag.demand_tag_priorities tag.payment_terms tag.supply_group_id
# see the contents of a field
In [5]: tag.name
Out[7]: "My Test Tag"
# change the contents and save it
In [6]: tag.name = "My New Test Tag"
In [7]: resp = tag.save()
In addition to working with single responses. This simple interface makes it easy to make calls that will return more than one result.
In [8]: tags = springserve.demand_tags.get()
In [9]: for tag in tags:
...: print tag.name
...:
My Tag 1
My Tag 2
Below is the documentation for and an example of using SpringServe reporting.
This module requires the pandas package to be installed.
In [10]: springserve.reports.run?
Signature: springserve.reports.run(start_date=None, end_date=None, interval=None,
dimensions=None, account_id=None, **kwargs)
Docstring:
parameter options (if applicable) notes
===================================================
start_date: "2015-12-01 00:00:00" or "2015-12-01"
end_date: "2015-12-02 00:00:00" or "2015-12-01"
interval: "hour", "day", "cumulative"
timezone: "UTC", "America/New_York" defaults to America/New_York
date_range: Today, Yesterday, Last 7 Days date_range takes precedence over
start_date/end_date
dimensions: supply_tag_id, demand_tag_id, detected_domain, declared_domain,
demand_type, supply_type, supply_partner_id, demand_partner_id, supply_group
domain is only available when using date_range of Today, Yesterday, or Last 7 Days
the following parameters act as filters; pass an array of values (usually IDs)
=================================================================================
supply_tag_ids: [22423,22375, 25463]
demand_tag_ids: [22423,22375, 25463]
detected_domains: ["nytimes.com", "weather.com"]
declared_domains: ["nytimes.com", "weather.com"]
supply_types ["Syndicated","Third-Party"]
supply_partner_ids: [30,42,41]
supply_group_ids: [13,15,81]
demand_partner_ids: [3,10,81]
demand_types: ["Vast Only","FLASH"]
In[11]: report = springserve.reports.run(state_date="2016-09-19", end_date="2016-09-19",
dimensions=["supply_tag_id"], declared_domains=["nytimes.com", "weather.com"])
In[12]: report.ok
Out[12]: True
Getting next pages of your report
=================================================================================
In[11]: report = springserve.reports.run(state_date="2016-09-19", end_date="2016-09-19",
dimensions=["supply_tag_id"], declared_domains=["nytimes.com", "weather.com"])
In[11]: report_df = report.to_dataframe()
In[12]: report.get_next_page()
Out[12]: True
This method returns True if it got the next page and False if not, indicating that you
are already at the last page.
Note that if you call get_next_page this overwrite the current data and when you call
to_dataframe the results will only contain the data from the last page you got.
In[12]: report.get_all_pages()
Out[12]:
This method gets all the remaining pages of the report (so if you're currently at page 2 it
will get page 2 onwards, if you're at page 1 it will get everything) then if you call to_dataframe
on the report you will get all the data. Note that if this is a very large report it's best
to get one page at a time so you don't run out of memory.