def test_register_simulated_price(self):
price_time = datetime.datetime(2011, 1, 1)
price_value = numpy.array([1.1, 1.2, 1.367345987359734598734598723459872345987235698237459862345])
simulation_id = create_uuid4()
self.assertRaises(KeyError, self.app.simulated_price_repo.__getitem__, simulation_id)
price = register_simulated_price(simulation_id, '#1', price_time, price_value)
assert isinstance(price, SimulatedPrice), price
assert price.id
price = self.app.simulated_price_repo[make_simulated_price_id(simulation_id, '#1', price_time)]
assert isinstance(price, SimulatedPrice)
numpy.testing.assert_equal(price.value, price_value)
def generate_simulated_prices(market_calibration, market_simulation):
for market_name, fixing_date, price_value in simulate_future_prices(market_simulation, market_calibration):
register_simulated_price(market_simulation.id, market_name, fixing_date, price_value)
def generate_simulated_prices(market_simulation, market_calibration):
for commodity_name, fixing_date, delivery_date, price_value in simulate_future_prices(
market_simulation, market_calibration):
yield register_simulated_price(market_simulation.id, commodity_name,
fixing_date, delivery_date, price_value)
def _test_distributed_dependency_graph_runner(self):
# Setup the contract.
# - branching function calls
dsl_source = """
def Swing(starts, ends, underlying, quantity):
if (quantity == 0) or (starts >= ends):
0
else:
Wait(starts, Choice(
Swing(starts + TimeDelta('1d'), ends, underlying, quantity - 1) + Fixing(starts, underlying),
Swing(starts + TimeDelta('1d'), ends, underlying, quantity)
))
Swing(Date('2011-01-01'), Date('2011-01-03'), 10, 50)
"""
# Generate the dependency graph.
dependency_graph = dsl_compile(dsl_source, is_parallel=True)
assert isinstance(dependency_graph, DependencyGraph)
# Remember the number of stubbed exprs - will check it after the evaluation.
actual_len_stubbed_exprs = len(dependency_graph.call_requirements)
kwds = {
'interest_rate': 0,
'present_time': datetime.datetime(2011, 1, 1, tzinfo=utc),
'simulation_id': create_uuid4(),
}
app = get_quantdsl_app()
market_simulation = app.register_market_simulation({})
market_names = ['#1']
for market_name in market_names:
# NB Need enough days to cover the date range in the dsl_source.
for i in range(0, 10):
dt = datetime.datetime(2011, 1, 1, tzinfo=utc) + datetime.timedelta(1) * i
value = numpy.array([10] * 2000)
register_simulated_price(market_simulation.id, market_name, fixing_date=dt)
# Check we've got a path to the 'celery' command line program (hopefully it's next to this python executable).
celery_script_path = os.path.join(os.path.dirname(sys.executable), 'celery')
self.assertTrue(os.path.exists(celery_script_path), celery_script_path)
# Check the example task returns correct result (this assumes the celery worker IS running).
# - invoke a celery worker process as a subprocess
worker_cmd = [celery_script_path, 'worker', '-A', 'quantdsl.infrastructure.celery.tasks', '-l', 'info']
# - its usage as a context manager causes a wait for it to finish after it has been terminated
with Popen(worker_cmd) as worker:
try:
# Evaluate the dependency graph.
runner = DistributedDependencyGraphRunner(dependency_graph, app=app)
dsl_value = runner.evaluate(**kwds)
# Get the mean of the value, if it has one.
if isinstance(dsl_value, numpy.ndarray):
dsl_value = dsl_value.mean()
# Check the value is expected.
expected_value = 20
self.assertEqual(dsl_value, expected_value)
# Check the number of stubbed exprs is expected.
expected_len_stubbed_exprs = 7
self.assertEqual(actual_len_stubbed_exprs, expected_len_stubbed_exprs)
finally:
# Shutdown the celery worker.
worker.terminate()
def generate_simulated_prices(market_simulation, market_calibration):
for commodity_name, fixing_date, delivery_date, price_value in simulate_future_prices(market_simulation, market_calibration):
yield register_simulated_price(market_simulation.id, commodity_name, fixing_date, delivery_date, price_value)