Fully featured python client library for StatPro RiskAPI

This library can be used to communicate programmatically with StatPro RiskAPI from python, it can be embedded in any python 2.7 application and it does not have mandatory external dependencies.

$ pip install git+https://github.com/StatProSA/riskapi-python-client.git

If you want to communicate in msgpack you will also need the optional package msgpack-python::

$ pip install msgpack-python>=0.4

If you want to run the test suite you will also need nose and voluptuous::

$ pip install nose==1.3.4 voluptuous==0.8.7

or just::

$ python setup.py test

Instantiate a client object and call its method. Arguments are just python objects, serialization and deserialization is performed transparently. HTTP errors are wrapped in "HTTPError" exceptions and raised, with the attribute "code" containing the HTTP status code and the exception message containing the error returned by the server, if any.

Please refer to StatPro RiskAPI documentation for further informations.

1. Create a file named .riskapi.conf in your home directory with your own credentials for the RiskAPI service, e.g.:

       [client]
       host=api.risk.statpro.com
       customer=internal
       user=<your user id>
       password=<your password>
   

2. Obtain a client object, the "connect" factory method is the easiest way:

       >>> conn = riskapi_client.connect()
       >>> conn
       <riskapi_client.RiskapiClient object at 0x7f14228f4bd0>
   

3. Get the list of available products, starting with "US" and limited to 2:

       >>> conn.products(search="US", limit=2)
       {u'count': 4963, u'data': [{u'reference_price': 56.48, u'code': u'US0003041052', u'product_type': 1, u'description': u'AAC Technologies Holdings Inc', u'last_update': u'2014-11-12', u'currency': u'USD', u'pricer': u''}, {u'reference_price': 115.337578118, u'code': u'US000324AA15', u'product_type': 7, u'description': u'AAF HOLDINGS 12.00% 01-Jul-2019 CALL', u'last_update': u'2014-11-11', u'currency': u'USD', u'pricer': u'callablecouponbond'}]}
   

4. Build a portfolio and add two holdings with the retrieved products plus a non-existing one:

       >>> pf = riskapi_client.Portfolio("EUR")
       >>> pf.add("US0003041052", quantity=13000)
       Holding('US0003041052', None, 13000, None, [], None, None)
       >>> pf.add("US000324AA15", quantity=10000)
       Holding('US000324AA15', None, 10000, None, [], None, None)
       >>> pf.add("XXX", quantity=13000)
       Holding('XXX', None, 13000, None, [], None, None)
   

5. Call some portfolio-level function:

       >>> conn.portfolio_info(pf)
       {u'errors': [[2, u'uncovered', u"Client code not found: u'XXX'", [u'XXX', None]]], u'results': {u'size': 2, u'exposure': 926549.3167053802, u'full_nominal_exposure': [[912219.776, u''], [14329.5407053803, u'']]}}
   
       >>> conn.risk(pf, [0.99])
       {u'errors': [[2, u'uncovered', u"Client code not found: u'XXX'", [u'XXX', None]]], u'results': [{u'potential_upside': 69280.4843514962, u'average_var': 60192.4303244384, u'expected_shortfall': 66361.2651884222, u'expected_upside': 78593.1966801616, u'lookback_days': 730, u'expected_return': 1075.6266813165, u'percentile': 0.99, u'frequency': 1, u'horizon': 1, u'average_potential_upside': 66453.7067345781, u'var': 60851.9174691098, u'expected_loss': 663.6126518842, u'diversification': 0.9999230482, u'volatility': 22179.0104368703}]}
   

In order to run the test suite you will need a proper ~/riskapi.conf, then run::

$ nosetests tests