Mariner Backtesting - Strategy.on_start()

Pre-Start initiating event

on_start() is used for pre-running activity. Data initialization should happen here, including reading of historic bars, and creation of any internal data structures. The strategy should also set its calculate and evaluate triggers with the appropriate calls to service.

Interface:

 def on_start(self,
             md,
             order,
             service,
             account):
    pass

Parameters:

Same for on_finish()

Name	Type	Default	Information
self	object	required	The instance of Strategy (used by the system).
md	object	required	Market data object. It should not be saved, as it may change in future calls.
order	object	required	Order object. It should not be saved, as it may change in future calls.
service	object	required	Service object. It should not be saved, as it may change in future calls.
account	object	required	Account object. It should not be saved, as it may change in future calls.

Remarks:

Best place to read daily bars (md.bar.daily(-10)), since the data isn't changing.
service.terminate() can be called in on_start() (e.g. a symbol no longer qualifies for model).
Good place to schedule on_timer() calls

Working Example:

 from cloudquant.interfaces import Strategy


class OnStartExample(Strategy):
    @classmethod
    def is_symbol_qualified(cls, symbol, md, service, account):
        return symbol == "AAL"

    def on_start(self, md, order, service, account):
        print(self.symbol +"\n"+ service.time_to_string(service.system_time) +"\n\n")

        print ("on_start has begun")
        self.ep = 0  # establish variables
        print ("on_start has completed")

Console

AAL
2016-08-09 09:24:32.989000

on_start has begun
on_start has completed