Total Complexity | 152 |
Total Lines | 1221 |
Duplicated Lines | 0 % |
Complex classes like zipline.TradingAlgorithm often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
1 | # |
||
113 | class TradingAlgorithm(object): |
||
114 | """ |
||
115 | Base class for trading algorithms. Inherit and overload |
||
116 | initialize() and handle_data(data). |
||
117 | |||
118 | A new algorithm could look like this: |
||
119 | ``` |
||
120 | from zipline.api import order, symbol |
||
121 | |||
122 | def initialize(context): |
||
123 | context.sid = symbol('AAPL') |
||
124 | context.amount = 100 |
||
125 | |||
126 | def handle_data(context, data): |
||
127 | sid = context.sid |
||
128 | amount = context.amount |
||
129 | order(sid, amount) |
||
130 | ``` |
||
131 | To then to run this algorithm pass these functions to |
||
132 | TradingAlgorithm: |
||
133 | |||
134 | my_algo = TradingAlgorithm(initialize, handle_data) |
||
135 | stats = my_algo.run(data) |
||
136 | |||
137 | """ |
||
138 | |||
139 | def __init__(self, *args, **kwargs): |
||
140 | """Initialize sids and other state variables. |
||
141 | |||
142 | :Arguments: |
||
143 | :Optional: |
||
144 | initialize : function |
||
145 | Function that is called with a single |
||
146 | argument at the begninning of the simulation. |
||
147 | handle_data : function |
||
148 | Function that is called with 2 arguments |
||
149 | (context and data) on every bar. |
||
150 | script : str |
||
151 | Algoscript that contains initialize and |
||
152 | handle_data function definition. |
||
153 | data_frequency : {'daily', 'minute'} |
||
154 | The duration of the bars. |
||
155 | capital_base : float <default: 1.0e5> |
||
156 | How much capital to start with. |
||
157 | asset_finder : An AssetFinder object |
||
158 | A new AssetFinder object to be used in this TradingEnvironment |
||
159 | equities_metadata : can be either: |
||
160 | - dict |
||
161 | - pandas.DataFrame |
||
162 | - object with 'read' property |
||
163 | If dict is provided, it must have the following structure: |
||
164 | * keys are the identifiers |
||
165 | * values are dicts containing the metadata, with the metadata |
||
166 | field name as the key |
||
167 | If pandas.DataFrame is provided, it must have the |
||
168 | following structure: |
||
169 | * column names must be the metadata fields |
||
170 | * index must be the different asset identifiers |
||
171 | * array contents should be the metadata value |
||
172 | If an object with a 'read' property is provided, 'read' must |
||
173 | return rows containing at least one of 'sid' or 'symbol' along |
||
174 | with the other metadata fields. |
||
175 | identifiers : List |
||
176 | Any asset identifiers that are not provided in the |
||
177 | equities_metadata, but will be traded by this TradingAlgorithm |
||
178 | """ |
||
179 | self.sources = [] |
||
180 | |||
181 | # List of trading controls to be used to validate orders. |
||
182 | self.trading_controls = [] |
||
183 | |||
184 | # List of account controls to be checked on each bar. |
||
185 | self.account_controls = [] |
||
186 | |||
187 | self._recorded_vars = {} |
||
188 | self.namespace = kwargs.pop('namespace', {}) |
||
189 | |||
190 | self._platform = kwargs.pop('platform', 'zipline') |
||
191 | |||
192 | self.logger = None |
||
193 | |||
194 | self.data_portal = None |
||
195 | |||
196 | # If an env has been provided, pop it |
||
197 | self.trading_environment = kwargs.pop('env', None) |
||
198 | |||
199 | if self.trading_environment is None: |
||
200 | self.trading_environment = TradingEnvironment() |
||
201 | |||
202 | # Update the TradingEnvironment with the provided asset metadata |
||
203 | self.trading_environment.write_data( |
||
204 | equities_data=kwargs.pop('equities_metadata', {}), |
||
205 | equities_identifiers=kwargs.pop('identifiers', []), |
||
206 | futures_data=kwargs.pop('futures_metadata', {}), |
||
207 | ) |
||
208 | |||
209 | # set the capital base |
||
210 | self.capital_base = kwargs.pop('capital_base', DEFAULT_CAPITAL_BASE) |
||
211 | self.sim_params = kwargs.pop('sim_params', None) |
||
212 | if self.sim_params is None: |
||
213 | self.sim_params = create_simulation_parameters( |
||
214 | capital_base=self.capital_base, |
||
215 | start=kwargs.pop('start', None), |
||
216 | end=kwargs.pop('end', None), |
||
217 | env=self.trading_environment, |
||
218 | ) |
||
219 | else: |
||
220 | self.sim_params.update_internal_from_env(self.trading_environment) |
||
221 | |||
222 | self.perf_tracker = None |
||
223 | # Pull in the environment's new AssetFinder for quick reference |
||
224 | self.asset_finder = self.trading_environment.asset_finder |
||
225 | |||
226 | # Initialize Pipeline API data. |
||
227 | self.init_engine(kwargs.pop('get_pipeline_loader', None)) |
||
228 | self._pipelines = {} |
||
229 | # Create an always-expired cache so that we compute the first time data |
||
230 | # is requested. |
||
231 | self._pipeline_cache = CachedObject(None, pd.Timestamp(0, tz='UTC')) |
||
232 | |||
233 | self.blotter = kwargs.pop('blotter', None) |
||
234 | if not self.blotter: |
||
235 | self.blotter = Blotter( |
||
236 | slippage_func=VolumeShareSlippage(), |
||
237 | commission=PerShare(), |
||
238 | data_frequency=self.data_frequency, |
||
239 | ) |
||
240 | |||
241 | # The symbol lookup date specifies the date to use when resolving |
||
242 | # symbols to sids, and can be set using set_symbol_lookup_date() |
||
243 | self._symbol_lookup_date = None |
||
244 | |||
245 | self.portfolio_needs_update = True |
||
246 | self.account_needs_update = True |
||
247 | self.performance_needs_update = True |
||
248 | self._portfolio = None |
||
249 | self._account = None |
||
250 | |||
251 | # If string is passed in, execute and get reference to |
||
252 | # functions. |
||
253 | self.algoscript = kwargs.pop('script', None) |
||
254 | |||
255 | self._initialize = None |
||
256 | self._before_trading_start = None |
||
257 | self._analyze = None |
||
258 | |||
259 | self.event_manager = EventManager( |
||
260 | create_context=kwargs.pop('create_event_context', None), |
||
261 | ) |
||
262 | |||
263 | if self.algoscript is not None: |
||
264 | filename = kwargs.pop('algo_filename', None) |
||
265 | if filename is None: |
||
266 | filename = '<string>' |
||
267 | code = compile(self.algoscript, filename, 'exec') |
||
268 | exec_(code, self.namespace) |
||
269 | self._initialize = self.namespace.get('initialize') |
||
270 | if 'handle_data' not in self.namespace: |
||
271 | raise ValueError('You must define a handle_data function.') |
||
272 | else: |
||
273 | self._handle_data = self.namespace['handle_data'] |
||
274 | |||
275 | self._before_trading_start = \ |
||
276 | self.namespace.get('before_trading_start') |
||
277 | # Optional analyze function, gets called after run |
||
278 | self._analyze = self.namespace.get('analyze') |
||
279 | |||
280 | elif kwargs.get('initialize') and kwargs.get('handle_data'): |
||
281 | if self.algoscript is not None: |
||
282 | raise ValueError('You can not set script and \ |
||
283 | initialize/handle_data.') |
||
284 | self._initialize = kwargs.pop('initialize') |
||
285 | self._handle_data = kwargs.pop('handle_data') |
||
286 | self._before_trading_start = kwargs.pop('before_trading_start', |
||
287 | None) |
||
288 | self._analyze = kwargs.pop('analyze', None) |
||
289 | |||
290 | self.event_manager.add_event( |
||
291 | zipline.utils.events.Event( |
||
292 | zipline.utils.events.Always(), |
||
293 | # We pass handle_data.__func__ to get the unbound method. |
||
294 | # We will explicitly pass the algorithm to bind it again. |
||
295 | self.handle_data.__func__, |
||
296 | ), |
||
297 | prepend=True, |
||
298 | ) |
||
299 | |||
300 | # If method not defined, NOOP |
||
301 | if self._initialize is None: |
||
302 | self._initialize = lambda x: None |
||
303 | |||
304 | # Alternative way of setting data_frequency for backwards |
||
305 | # compatibility. |
||
306 | if 'data_frequency' in kwargs: |
||
307 | self.data_frequency = kwargs.pop('data_frequency') |
||
308 | |||
309 | # Prepare the algo for initialization |
||
310 | self.initialized = False |
||
311 | self.initialize_args = args |
||
312 | self.initialize_kwargs = kwargs |
||
313 | |||
314 | self.benchmark_sid = kwargs.pop('benchmark_sid', None) |
||
315 | |||
316 | def init_engine(self, get_loader): |
||
317 | """ |
||
318 | Construct and store a PipelineEngine from loader. |
||
319 | |||
320 | If get_loader is None, constructs a NoOpPipelineEngine. |
||
321 | """ |
||
322 | if get_loader is not None: |
||
323 | self.engine = SimplePipelineEngine( |
||
324 | get_loader, |
||
325 | self.trading_environment.trading_days, |
||
326 | self.asset_finder, |
||
327 | ) |
||
328 | else: |
||
329 | self.engine = NoOpPipelineEngine() |
||
330 | |||
331 | def initialize(self, *args, **kwargs): |
||
332 | """ |
||
333 | Call self._initialize with `self` made available to Zipline API |
||
334 | functions. |
||
335 | """ |
||
336 | with ZiplineAPI(self): |
||
337 | self._initialize(self, *args, **kwargs) |
||
338 | |||
339 | def before_trading_start(self, data): |
||
340 | if self._before_trading_start is None: |
||
341 | return |
||
342 | |||
343 | self._before_trading_start(self, data) |
||
344 | |||
345 | def handle_data(self, data): |
||
346 | self._handle_data(self, data) |
||
347 | |||
348 | # Unlike trading controls which remain constant unless placing an |
||
349 | # order, account controls can change each bar. Thus, must check |
||
350 | # every bar no matter if the algorithm places an order or not. |
||
351 | self.validate_account_controls() |
||
352 | |||
353 | def analyze(self, perf): |
||
354 | if self._analyze is None: |
||
355 | return |
||
356 | |||
357 | with ZiplineAPI(self): |
||
358 | self._analyze(self, perf) |
||
359 | |||
360 | def __repr__(self): |
||
361 | """ |
||
362 | N.B. this does not yet represent a string that can be used |
||
363 | to instantiate an exact copy of an algorithm. |
||
364 | |||
365 | However, it is getting close, and provides some value as something |
||
366 | that can be inspected interactively. |
||
367 | """ |
||
368 | return """ |
||
369 | {class_name}( |
||
370 | capital_base={capital_base} |
||
371 | sim_params={sim_params}, |
||
372 | initialized={initialized}, |
||
373 | slippage={slippage}, |
||
374 | commission={commission}, |
||
375 | blotter={blotter}, |
||
376 | recorded_vars={recorded_vars}) |
||
377 | """.strip().format(class_name=self.__class__.__name__, |
||
378 | capital_base=self.capital_base, |
||
379 | sim_params=repr(self.sim_params), |
||
380 | initialized=self.initialized, |
||
381 | slippage=repr(self.blotter.slippage_func), |
||
382 | commission=repr(self.blotter.commission), |
||
383 | blotter=repr(self.blotter), |
||
384 | recorded_vars=repr(self.recorded_vars)) |
||
385 | |||
386 | def _create_clock(self): |
||
387 | """ |
||
388 | If the clock property is not set, then create one based on frequency. |
||
389 | """ |
||
390 | if self.sim_params.data_frequency == 'minute': |
||
391 | env = self.trading_environment |
||
392 | trading_o_and_c = env.open_and_closes.ix[ |
||
393 | self.sim_params.trading_days] |
||
394 | market_opens = trading_o_and_c['market_open'].values.astype( |
||
395 | 'datetime64[ns]').astype(np.int64) |
||
396 | market_closes = trading_o_and_c['market_close'].values.astype( |
||
397 | 'datetime64[ns]').astype(np.int64) |
||
398 | |||
399 | minutely_emission = self.sim_params.emission_rate == "minute" |
||
400 | |||
401 | clock = MinuteSimulationClock( |
||
402 | self.sim_params.trading_days, |
||
403 | market_opens, |
||
404 | market_closes, |
||
405 | env.trading_days, |
||
406 | minutely_emission |
||
407 | ) |
||
408 | return clock |
||
409 | else: |
||
410 | return DailySimulationClock(self.sim_params.trading_days) |
||
411 | |||
412 | def _create_benchmark_source(self): |
||
413 | return BenchmarkSource( |
||
414 | self.benchmark_sid, |
||
415 | self.trading_environment, |
||
416 | self.sim_params.trading_days, |
||
417 | self.data_portal, |
||
418 | emission_rate=self.sim_params.emission_rate, |
||
419 | ) |
||
420 | |||
421 | def _create_generator(self, sim_params): |
||
422 | if sim_params is not None: |
||
423 | self.sim_params = sim_params |
||
424 | |||
425 | if self.perf_tracker is None: |
||
426 | # HACK: When running with the `run` method, we set perf_tracker to |
||
427 | # None so that it will be overwritten here. |
||
428 | self.perf_tracker = PerformanceTracker( |
||
429 | sim_params=self.sim_params, |
||
430 | env=self.trading_environment, |
||
431 | data_portal=self.data_portal |
||
432 | ) |
||
433 | |||
434 | # Set the dt initially to the period start by forcing it to change. |
||
435 | self.on_dt_changed(self.sim_params.period_start) |
||
436 | |||
437 | if not self.initialized: |
||
438 | self.initialize(*self.initialize_args, **self.initialize_kwargs) |
||
439 | self.initialized = True |
||
440 | |||
441 | self.trading_client = self._create_trading_client(sim_params) |
||
442 | |||
443 | return self.trading_client.transform() |
||
444 | |||
445 | def _create_trading_client(self, sim_params): |
||
446 | return AlgorithmSimulator( |
||
447 | self, |
||
448 | sim_params, |
||
449 | self.data_portal, |
||
450 | self._create_clock(), |
||
451 | self._create_benchmark_source() |
||
452 | ) |
||
453 | |||
454 | def get_generator(self): |
||
455 | """ |
||
456 | Override this method to add new logic to the construction |
||
457 | of the generator. Overrides can use the _create_generator |
||
458 | method to get a standard construction generator. |
||
459 | """ |
||
460 | return self._create_generator(self.sim_params) |
||
461 | |||
462 | def run(self, data_portal=None): |
||
463 | """Run the algorithm. |
||
464 | |||
465 | :Arguments: |
||
466 | source : DataPortal |
||
467 | |||
468 | :Returns: |
||
469 | daily_stats : pandas.DataFrame |
||
470 | Daily performance metrics such as returns, alpha etc. |
||
471 | |||
472 | """ |
||
473 | self.data_portal = data_portal |
||
474 | |||
475 | # Force a reset of the performance tracker, in case |
||
476 | # this is a repeat run of the algorithm. |
||
477 | self.perf_tracker = None |
||
478 | |||
479 | # Create zipline and loop through simulated_trading. |
||
480 | # Each iteration returns a perf dictionary |
||
481 | perfs = [] |
||
482 | for perf in self.get_generator(): |
||
483 | perfs.append(perf) |
||
484 | |||
485 | # convert perf dict to pandas dataframe |
||
486 | daily_stats = self._create_daily_stats(perfs) |
||
487 | |||
488 | self.analyze(daily_stats) |
||
489 | |||
490 | return daily_stats |
||
491 | |||
492 | def _write_and_map_id_index_to_sids(self, identifiers, as_of_date): |
||
493 | # Build new Assets for identifiers that can't be resolved as |
||
494 | # sids/Assets |
||
495 | identifiers_to_build = [] |
||
496 | for identifier in identifiers: |
||
497 | asset = None |
||
498 | |||
499 | if isinstance(identifier, Asset): |
||
500 | asset = self.asset_finder.retrieve_asset(sid=identifier.sid, |
||
501 | default_none=True) |
||
502 | elif isinstance(identifier, Integral): |
||
503 | asset = self.asset_finder.retrieve_asset(sid=identifier, |
||
504 | default_none=True) |
||
505 | if asset is None: |
||
506 | identifiers_to_build.append(identifier) |
||
507 | |||
508 | self.trading_environment.write_data( |
||
509 | equities_identifiers=identifiers_to_build) |
||
510 | |||
511 | # We need to clear out any cache misses that were stored while trying |
||
512 | # to do lookups. The real fix for this problem is to not construct an |
||
513 | # AssetFinder until we `run()` when we actually have all the data we |
||
514 | # need to so. |
||
515 | self.asset_finder._reset_caches() |
||
516 | |||
517 | return self.asset_finder.map_identifier_index_to_sids( |
||
518 | identifiers, as_of_date, |
||
519 | ) |
||
520 | |||
521 | def _create_daily_stats(self, perfs): |
||
522 | # create daily and cumulative stats dataframe |
||
523 | daily_perfs = [] |
||
524 | # TODO: the loop here could overwrite expected properties |
||
525 | # of daily_perf. Could potentially raise or log a |
||
526 | # warning. |
||
527 | for perf in perfs: |
||
528 | if 'daily_perf' in perf: |
||
529 | |||
530 | perf['daily_perf'].update( |
||
531 | perf['daily_perf'].pop('recorded_vars') |
||
532 | ) |
||
533 | perf['daily_perf'].update(perf['cumulative_risk_metrics']) |
||
534 | daily_perfs.append(perf['daily_perf']) |
||
535 | else: |
||
536 | self.risk_report = perf |
||
537 | |||
538 | daily_dts = [np.datetime64(perf['period_close'], utc=True) |
||
539 | for perf in daily_perfs] |
||
540 | daily_stats = pd.DataFrame(daily_perfs, index=daily_dts) |
||
541 | |||
542 | return daily_stats |
||
543 | |||
544 | @api_method |
||
545 | def get_environment(self, field='platform'): |
||
546 | env = { |
||
547 | 'arena': self.sim_params.arena, |
||
548 | 'data_frequency': self.sim_params.data_frequency, |
||
549 | 'start': self.sim_params.first_open, |
||
550 | 'end': self.sim_params.last_close, |
||
551 | 'capital_base': self.sim_params.capital_base, |
||
552 | 'platform': self._platform |
||
553 | } |
||
554 | if field == '*': |
||
555 | return env |
||
556 | else: |
||
557 | return env[field] |
||
558 | |||
559 | @api_method |
||
560 | def fetch_csv(self, url, |
||
561 | pre_func=None, |
||
562 | post_func=None, |
||
563 | date_column='date', |
||
564 | date_format=None, |
||
565 | timezone=pytz.utc.zone, |
||
566 | symbol=None, |
||
567 | mask=True, |
||
568 | symbol_column=None, |
||
569 | special_params_checker=None, |
||
570 | **kwargs): |
||
571 | |||
572 | # Show all the logs every time fetcher is used. |
||
573 | csv_data_source = PandasRequestsCSV( |
||
574 | url, |
||
575 | pre_func, |
||
576 | post_func, |
||
577 | self.trading_environment, |
||
578 | self.sim_params.period_start, |
||
579 | self.sim_params.period_end, |
||
580 | date_column, |
||
581 | date_format, |
||
582 | timezone, |
||
583 | symbol, |
||
584 | mask, |
||
585 | symbol_column, |
||
586 | data_frequency=self.data_frequency, |
||
587 | special_params_checker=special_params_checker, |
||
588 | **kwargs |
||
589 | ) |
||
590 | |||
591 | # ingest this into dataportal |
||
592 | self.data_portal.handle_extra_source(csv_data_source.df, |
||
593 | self.sim_params) |
||
594 | |||
595 | return csv_data_source |
||
596 | |||
597 | def add_event(self, rule=None, callback=None): |
||
598 | """ |
||
599 | Adds an event to the algorithm's EventManager. |
||
600 | """ |
||
601 | self.event_manager.add_event( |
||
602 | zipline.utils.events.Event(rule, callback), |
||
603 | ) |
||
604 | |||
605 | @api_method |
||
606 | def schedule_function(self, |
||
607 | func, |
||
608 | date_rule=None, |
||
609 | time_rule=None, |
||
610 | half_days=True): |
||
611 | """ |
||
612 | Schedules a function to be called with some timed rules. |
||
613 | """ |
||
614 | date_rule = date_rule or DateRuleFactory.every_day() |
||
615 | time_rule = ((time_rule or TimeRuleFactory.market_open()) |
||
616 | if self.sim_params.data_frequency == 'minute' else |
||
617 | # If we are in daily mode the time_rule is ignored. |
||
618 | zipline.utils.events.Always()) |
||
619 | |||
620 | self.add_event( |
||
621 | make_eventrule(date_rule, time_rule, half_days), |
||
622 | func, |
||
623 | ) |
||
624 | |||
625 | @api_method |
||
626 | def record(self, *args, **kwargs): |
||
627 | """ |
||
628 | Track and record local variable (i.e. attributes) each day. |
||
629 | """ |
||
630 | # Make 2 objects both referencing the same iterator |
||
631 | args = [iter(args)] * 2 |
||
632 | |||
633 | # Zip generates list entries by calling `next` on each iterator it |
||
634 | # receives. In this case the two iterators are the same object, so the |
||
635 | # call to next on args[0] will also advance args[1], resulting in zip |
||
636 | # returning (a,b) (c,d) (e,f) rather than (a,a) (b,b) (c,c) etc. |
||
637 | positionals = zip(*args) |
||
638 | for name, value in chain(positionals, iteritems(kwargs)): |
||
639 | self._recorded_vars[name] = value |
||
640 | |||
641 | @api_method |
||
642 | def set_benchmark(self, benchmark_sid): |
||
643 | if self.initialized: |
||
644 | raise SetBenchmarkOutsideInitialize() |
||
645 | |||
646 | self.benchmark_sid = benchmark_sid |
||
647 | |||
648 | @api_method |
||
649 | @preprocess(symbol_str=ensure_upper_case) |
||
650 | def symbol(self, symbol_str): |
||
651 | """ |
||
652 | Default symbol lookup for any source that directly maps the |
||
653 | symbol to the Asset (e.g. yahoo finance). |
||
654 | """ |
||
655 | # If the user has not set the symbol lookup date, |
||
656 | # use the period_end as the date for sybmol->sid resolution. |
||
657 | _lookup_date = self._symbol_lookup_date if self._symbol_lookup_date is not None \ |
||
658 | else self.sim_params.period_end |
||
659 | |||
660 | return self.asset_finder.lookup_symbol( |
||
661 | symbol_str, |
||
662 | as_of_date=_lookup_date, |
||
663 | ) |
||
664 | |||
665 | @api_method |
||
666 | def symbols(self, *args): |
||
667 | """ |
||
668 | Default symbols lookup for any source that directly maps the |
||
669 | symbol to the Asset (e.g. yahoo finance). |
||
670 | """ |
||
671 | return [self.symbol(identifier) for identifier in args] |
||
672 | |||
673 | @api_method |
||
674 | def sid(self, a_sid): |
||
675 | """ |
||
676 | Default sid lookup for any source that directly maps the integer sid |
||
677 | to the Asset. |
||
678 | """ |
||
679 | return self.asset_finder.retrieve_asset(a_sid) |
||
680 | |||
681 | @api_method |
||
682 | @preprocess(symbol=ensure_upper_case) |
||
683 | def future_symbol(self, symbol): |
||
684 | """ Lookup a futures contract with a given symbol. |
||
685 | |||
686 | Parameters |
||
687 | ---------- |
||
688 | symbol : str |
||
689 | The symbol of the desired contract. |
||
690 | |||
691 | Returns |
||
692 | ------- |
||
693 | Future |
||
694 | A Future object. |
||
695 | |||
696 | Raises |
||
697 | ------ |
||
698 | SymbolNotFound |
||
699 | Raised when no contract named 'symbol' is found. |
||
700 | |||
701 | """ |
||
702 | return self.asset_finder.lookup_future_symbol(symbol) |
||
703 | |||
704 | @api_method |
||
705 | @preprocess(root_symbol=ensure_upper_case) |
||
706 | def future_chain(self, root_symbol, as_of_date=None): |
||
707 | """ Look up a future chain with the specified parameters. |
||
708 | |||
709 | Parameters |
||
710 | ---------- |
||
711 | root_symbol : str |
||
712 | The root symbol of a future chain. |
||
713 | as_of_date : datetime.datetime or pandas.Timestamp or str, optional |
||
714 | Date at which the chain determination is rooted. I.e. the |
||
715 | existing contract whose notice date is first after this date is |
||
716 | the primary contract, etc. |
||
717 | |||
718 | Returns |
||
719 | ------- |
||
720 | FutureChain |
||
721 | The future chain matching the specified parameters. |
||
722 | |||
723 | Raises |
||
724 | ------ |
||
725 | RootSymbolNotFound |
||
726 | If a future chain could not be found for the given root symbol. |
||
727 | """ |
||
728 | if as_of_date: |
||
729 | try: |
||
730 | as_of_date = pd.Timestamp(as_of_date, tz='UTC') |
||
731 | except ValueError: |
||
732 | raise UnsupportedDatetimeFormat(input=as_of_date, |
||
733 | method='future_chain') |
||
734 | return FutureChain( |
||
735 | asset_finder=self.asset_finder, |
||
736 | get_datetime=self.get_datetime, |
||
737 | root_symbol=root_symbol, |
||
738 | as_of_date=as_of_date |
||
739 | ) |
||
740 | |||
741 | def _calculate_order_value_amount(self, asset, value): |
||
742 | """ |
||
743 | Calculates how many shares/contracts to order based on the type of |
||
744 | asset being ordered. |
||
745 | """ |
||
746 | last_price = self.trading_client.current_data[asset].price |
||
747 | |||
748 | if tolerant_equals(last_price, 0): |
||
749 | zero_message = "Price of 0 for {psid}; can't infer value".format( |
||
750 | psid=asset |
||
751 | ) |
||
752 | if self.logger: |
||
753 | self.logger.debug(zero_message) |
||
754 | # Don't place any order |
||
755 | return 0 |
||
756 | |||
757 | if isinstance(asset, Future): |
||
758 | value_multiplier = asset.contract_multiplier |
||
759 | else: |
||
760 | value_multiplier = 1 |
||
761 | |||
762 | return value / (last_price * value_multiplier) |
||
763 | |||
764 | @api_method |
||
765 | def order(self, sid, amount, |
||
766 | limit_price=None, |
||
767 | stop_price=None, |
||
768 | style=None): |
||
769 | """ |
||
770 | Place an order using the specified parameters. |
||
771 | """ |
||
772 | # Truncate to the integer share count that's either within .0001 of |
||
773 | # amount or closer to zero. |
||
774 | # E.g. 3.9999 -> 4.0; 5.5 -> 5.0; -5.5 -> -5.0 |
||
775 | amount = int(round_if_near_integer(amount)) |
||
776 | |||
777 | # Raises a ZiplineError if invalid parameters are detected. |
||
778 | self.validate_order_params(sid, |
||
779 | amount, |
||
780 | limit_price, |
||
781 | stop_price, |
||
782 | style) |
||
783 | |||
784 | # Convert deprecated limit_price and stop_price parameters to use |
||
785 | # ExecutionStyle objects. |
||
786 | style = self.__convert_order_params_for_blotter(limit_price, |
||
787 | stop_price, |
||
788 | style) |
||
789 | return self.blotter.order(sid, amount, style) |
||
790 | |||
791 | def validate_order_params(self, |
||
792 | asset, |
||
793 | amount, |
||
794 | limit_price, |
||
795 | stop_price, |
||
796 | style): |
||
797 | """ |
||
798 | Helper method for validating parameters to the order API function. |
||
799 | |||
800 | Raises an UnsupportedOrderParameters if invalid arguments are found. |
||
801 | """ |
||
802 | |||
803 | if not self.initialized: |
||
804 | raise OrderDuringInitialize( |
||
805 | msg="order() can only be called from within handle_data()" |
||
806 | ) |
||
807 | |||
808 | if style: |
||
809 | if limit_price: |
||
810 | raise UnsupportedOrderParameters( |
||
811 | msg="Passing both limit_price and style is not supported." |
||
812 | ) |
||
813 | |||
814 | if stop_price: |
||
815 | raise UnsupportedOrderParameters( |
||
816 | msg="Passing both stop_price and style is not supported." |
||
817 | ) |
||
818 | |||
819 | if not isinstance(asset, Asset): |
||
820 | raise UnsupportedOrderParameters( |
||
821 | msg="Passing non-Asset argument to 'order()' is not supported." |
||
822 | " Use 'sid()' or 'symbol()' methods to look up an Asset." |
||
823 | ) |
||
824 | |||
825 | for control in self.trading_controls: |
||
826 | control.validate(asset, |
||
827 | amount, |
||
828 | self.updated_portfolio(), |
||
829 | self.get_datetime(), |
||
830 | self.trading_client.current_data) |
||
831 | |||
832 | @staticmethod |
||
833 | def __convert_order_params_for_blotter(limit_price, stop_price, style): |
||
834 | """ |
||
835 | Helper method for converting deprecated limit_price and stop_price |
||
836 | arguments into ExecutionStyle instances. |
||
837 | |||
838 | This function assumes that either style == None or (limit_price, |
||
839 | stop_price) == (None, None). |
||
840 | """ |
||
841 | # TODO_SS: DeprecationWarning for usage of limit_price and stop_price. |
||
842 | if style: |
||
843 | assert (limit_price, stop_price) == (None, None) |
||
844 | return style |
||
845 | if limit_price and stop_price: |
||
846 | return StopLimitOrder(limit_price, stop_price) |
||
847 | if limit_price: |
||
848 | return LimitOrder(limit_price) |
||
849 | if stop_price: |
||
850 | return StopOrder(stop_price) |
||
851 | else: |
||
852 | return MarketOrder() |
||
853 | |||
854 | @api_method |
||
855 | def order_value(self, sid, value, |
||
856 | limit_price=None, stop_price=None, style=None): |
||
857 | """ |
||
858 | Place an order by desired value rather than desired number of shares. |
||
859 | If the requested sid exists, the requested value is |
||
860 | divided by its price to imply the number of shares to transact. |
||
861 | If the Asset being ordered is a Future, the 'value' calculated |
||
862 | is actually the exposure, as Futures have no 'value'. |
||
863 | |||
864 | value > 0 :: Buy/Cover |
||
865 | value < 0 :: Sell/Short |
||
866 | Market order: order(sid, value) |
||
867 | Limit order: order(sid, value, limit_price) |
||
868 | Stop order: order(sid, value, None, stop_price) |
||
869 | StopLimit order: order(sid, value, limit_price, stop_price) |
||
870 | """ |
||
871 | amount = self._calculate_order_value_amount(sid, value) |
||
872 | return self.order(sid, amount, |
||
873 | limit_price=limit_price, |
||
874 | stop_price=stop_price, |
||
875 | style=style) |
||
876 | |||
877 | @property |
||
878 | def recorded_vars(self): |
||
879 | return copy(self._recorded_vars) |
||
880 | |||
881 | @property |
||
882 | def portfolio(self): |
||
883 | return self.updated_portfolio() |
||
884 | |||
885 | def updated_portfolio(self): |
||
886 | if self.portfolio_needs_update: |
||
887 | self._portfolio = \ |
||
888 | self.perf_tracker.get_portfolio(self.performance_needs_update, |
||
889 | self.datetime) |
||
890 | self.portfolio_needs_update = False |
||
891 | self.performance_needs_update = False |
||
892 | return self._portfolio |
||
893 | |||
894 | @property |
||
895 | def account(self): |
||
896 | return self.updated_account() |
||
897 | |||
898 | def updated_account(self): |
||
899 | if self.account_needs_update: |
||
900 | self._account = \ |
||
901 | self.perf_tracker.get_account(self.performance_needs_update, |
||
902 | self.datetime) |
||
903 | self.account_needs_update = False |
||
904 | self.performance_needs_update = False |
||
905 | return self._account |
||
906 | |||
907 | def set_logger(self, logger): |
||
908 | self.logger = logger |
||
909 | |||
910 | def on_dt_changed(self, dt): |
||
911 | """ |
||
912 | Callback triggered by the simulation loop whenever the current dt |
||
913 | changes. |
||
914 | |||
915 | Any logic that should happen exactly once at the start of each datetime |
||
916 | group should happen here. |
||
917 | """ |
||
918 | assert isinstance(dt, datetime), \ |
||
919 | "Attempt to set algorithm's current time with non-datetime" |
||
920 | assert dt.tzinfo == pytz.utc, \ |
||
921 | "Algorithm expects a utc datetime" |
||
922 | |||
923 | self.datetime = dt |
||
924 | self.perf_tracker.set_date(dt) |
||
925 | self.blotter.set_date(dt) |
||
926 | |||
927 | self.portfolio_needs_update = True |
||
928 | self.account_needs_update = True |
||
929 | self.performance_needs_update = True |
||
930 | |||
931 | @api_method |
||
932 | def get_datetime(self, tz=None): |
||
933 | """ |
||
934 | Returns the simulation datetime. |
||
935 | """ |
||
936 | dt = self.datetime |
||
937 | assert dt.tzinfo == pytz.utc, "Algorithm should have a utc datetime" |
||
938 | |||
939 | if tz is not None: |
||
940 | # Convert to the given timezone passed as a string or tzinfo. |
||
941 | if isinstance(tz, string_types): |
||
942 | tz = pytz.timezone(tz) |
||
943 | dt = dt.astimezone(tz) |
||
944 | |||
945 | return dt # datetime.datetime objects are immutable. |
||
946 | |||
947 | def update_dividends(self, dividend_frame): |
||
948 | """ |
||
949 | Set DataFrame used to process dividends. DataFrame columns should |
||
950 | contain at least the entries in zp.DIVIDEND_FIELDS. |
||
951 | """ |
||
952 | self.perf_tracker.update_dividends(dividend_frame) |
||
953 | |||
954 | @api_method |
||
955 | def set_slippage(self, slippage): |
||
956 | if not isinstance(slippage, SlippageModel): |
||
957 | raise UnsupportedSlippageModel() |
||
958 | if self.initialized: |
||
959 | raise SetSlippagePostInit() |
||
960 | self.blotter.slippage_func = slippage |
||
961 | |||
962 | @api_method |
||
963 | def set_commission(self, commission): |
||
964 | if not isinstance(commission, (PerShare, PerTrade, PerDollar)): |
||
965 | raise UnsupportedCommissionModel() |
||
966 | |||
967 | if self.initialized: |
||
968 | raise SetCommissionPostInit() |
||
969 | self.blotter.commission = commission |
||
970 | |||
971 | @api_method |
||
972 | def set_symbol_lookup_date(self, dt): |
||
973 | """ |
||
974 | Set the date for which symbols will be resolved to their sids |
||
975 | (symbols may map to different firms or underlying assets at |
||
976 | different times) |
||
977 | """ |
||
978 | try: |
||
979 | self._symbol_lookup_date = pd.Timestamp(dt, tz='UTC') |
||
980 | except ValueError: |
||
981 | raise UnsupportedDatetimeFormat(input=dt, |
||
982 | method='set_symbol_lookup_date') |
||
983 | |||
984 | # Remain backwards compatibility |
||
985 | @property |
||
986 | def data_frequency(self): |
||
987 | return self.sim_params.data_frequency |
||
988 | |||
989 | @data_frequency.setter |
||
990 | def data_frequency(self, value): |
||
991 | assert value in ('daily', 'minute') |
||
992 | self.sim_params.data_frequency = value |
||
993 | |||
994 | @api_method |
||
995 | def order_percent(self, sid, percent, |
||
996 | limit_price=None, stop_price=None, style=None): |
||
997 | """ |
||
998 | Place an order in the specified asset corresponding to the given |
||
999 | percent of the current portfolio value. |
||
1000 | |||
1001 | Note that percent must expressed as a decimal (0.50 means 50\%). |
||
1002 | """ |
||
1003 | value = self.portfolio.portfolio_value * percent |
||
1004 | return self.order_value(sid, value, |
||
1005 | limit_price=limit_price, |
||
1006 | stop_price=stop_price, |
||
1007 | style=style) |
||
1008 | |||
1009 | @api_method |
||
1010 | def order_target(self, sid, target, |
||
1011 | limit_price=None, stop_price=None, style=None): |
||
1012 | """ |
||
1013 | Place an order to adjust a position to a target number of shares. If |
||
1014 | the position doesn't already exist, this is equivalent to placing a new |
||
1015 | order. If the position does exist, this is equivalent to placing an |
||
1016 | order for the difference between the target number of shares and the |
||
1017 | current number of shares. |
||
1018 | """ |
||
1019 | if sid in self.portfolio.positions: |
||
1020 | current_position = self.portfolio.positions[sid].amount |
||
1021 | req_shares = target - current_position |
||
1022 | return self.order(sid, req_shares, |
||
1023 | limit_price=limit_price, |
||
1024 | stop_price=stop_price, |
||
1025 | style=style) |
||
1026 | else: |
||
1027 | return self.order(sid, target, |
||
1028 | limit_price=limit_price, |
||
1029 | stop_price=stop_price, |
||
1030 | style=style) |
||
1031 | |||
1032 | @api_method |
||
1033 | def order_target_value(self, sid, target, |
||
1034 | limit_price=None, stop_price=None, style=None): |
||
1035 | """ |
||
1036 | Place an order to adjust a position to a target value. If |
||
1037 | the position doesn't already exist, this is equivalent to placing a new |
||
1038 | order. If the position does exist, this is equivalent to placing an |
||
1039 | order for the difference between the target value and the |
||
1040 | current value. |
||
1041 | If the Asset being ordered is a Future, the 'target value' calculated |
||
1042 | is actually the target exposure, as Futures have no 'value'. |
||
1043 | """ |
||
1044 | target_amount = self._calculate_order_value_amount(sid, target) |
||
1045 | return self.order_target(sid, target_amount, |
||
1046 | limit_price=limit_price, |
||
1047 | stop_price=stop_price, |
||
1048 | style=style) |
||
1049 | |||
1050 | @api_method |
||
1051 | def order_target_percent(self, sid, target, |
||
1052 | limit_price=None, stop_price=None, style=None): |
||
1053 | """ |
||
1054 | Place an order to adjust a position to a target percent of the |
||
1055 | current portfolio value. If the position doesn't already exist, this is |
||
1056 | equivalent to placing a new order. If the position does exist, this is |
||
1057 | equivalent to placing an order for the difference between the target |
||
1058 | percent and the current percent. |
||
1059 | |||
1060 | Note that target must expressed as a decimal (0.50 means 50\%). |
||
1061 | """ |
||
1062 | target_value = self.portfolio.portfolio_value * target |
||
1063 | return self.order_target_value(sid, target_value, |
||
1064 | limit_price=limit_price, |
||
1065 | stop_price=stop_price, |
||
1066 | style=style) |
||
1067 | |||
1068 | @api_method |
||
1069 | def get_open_orders(self, sid=None): |
||
1070 | if sid is None: |
||
1071 | return { |
||
1072 | key: [order.to_api_obj() for order in orders] |
||
1073 | for key, orders in iteritems(self.blotter.open_orders) |
||
1074 | if orders |
||
1075 | } |
||
1076 | if sid in self.blotter.open_orders: |
||
1077 | orders = self.blotter.open_orders[sid] |
||
1078 | return [order.to_api_obj() for order in orders] |
||
1079 | return [] |
||
1080 | |||
1081 | @api_method |
||
1082 | def get_order(self, order_id): |
||
1083 | if order_id in self.blotter.orders: |
||
1084 | return self.blotter.orders[order_id].to_api_obj() |
||
1085 | |||
1086 | @api_method |
||
1087 | def cancel_order(self, order_param): |
||
1088 | order_id = order_param |
||
1089 | if isinstance(order_param, zipline.protocol.Order): |
||
1090 | order_id = order_param.id |
||
1091 | |||
1092 | self.blotter.cancel(order_id) |
||
1093 | |||
1094 | @api_method |
||
1095 | @require_initialized(HistoryInInitialize()) |
||
1096 | def history(self, sids, bar_count, frequency, field, ffill=True): |
||
1097 | if self.data_portal is None: |
||
1098 | raise Exception("no data portal!") |
||
1099 | |||
1100 | return self.data_portal.get_history_window( |
||
1101 | sids, |
||
1102 | self.datetime, |
||
1103 | bar_count, |
||
1104 | frequency, |
||
1105 | field, |
||
1106 | ffill, |
||
1107 | ) |
||
1108 | |||
1109 | #################### |
||
1110 | # Account Controls # |
||
1111 | #################### |
||
1112 | |||
1113 | def register_account_control(self, control): |
||
1114 | """ |
||
1115 | Register a new AccountControl to be checked on each bar. |
||
1116 | """ |
||
1117 | if self.initialized: |
||
1118 | raise RegisterAccountControlPostInit() |
||
1119 | self.account_controls.append(control) |
||
1120 | |||
1121 | def validate_account_controls(self): |
||
1122 | for control in self.account_controls: |
||
1123 | control.validate(self.updated_portfolio(), |
||
1124 | self.updated_account(), |
||
1125 | self.get_datetime(), |
||
1126 | self.trading_client.current_data) |
||
1127 | |||
1128 | @api_method |
||
1129 | def set_max_leverage(self, max_leverage=None): |
||
1130 | """ |
||
1131 | Set a limit on the maximum leverage of the algorithm. |
||
1132 | """ |
||
1133 | control = MaxLeverage(max_leverage) |
||
1134 | self.register_account_control(control) |
||
1135 | |||
1136 | #################### |
||
1137 | # Trading Controls # |
||
1138 | #################### |
||
1139 | |||
1140 | def register_trading_control(self, control): |
||
1141 | """ |
||
1142 | Register a new TradingControl to be checked prior to order calls. |
||
1143 | """ |
||
1144 | if self.initialized: |
||
1145 | raise RegisterTradingControlPostInit() |
||
1146 | self.trading_controls.append(control) |
||
1147 | |||
1148 | @api_method |
||
1149 | def set_max_position_size(self, |
||
1150 | sid=None, |
||
1151 | max_shares=None, |
||
1152 | max_notional=None): |
||
1153 | """ |
||
1154 | Set a limit on the number of shares and/or dollar value held for the |
||
1155 | given sid. Limits are treated as absolute values and are enforced at |
||
1156 | the time that the algo attempts to place an order for sid. This means |
||
1157 | that it's possible to end up with more than the max number of shares |
||
1158 | due to splits/dividends, and more than the max notional due to price |
||
1159 | improvement. |
||
1160 | |||
1161 | If an algorithm attempts to place an order that would result in |
||
1162 | increasing the absolute value of shares/dollar value exceeding one of |
||
1163 | these limits, raise a TradingControlException. |
||
1164 | """ |
||
1165 | control = MaxPositionSize(asset=sid, |
||
1166 | max_shares=max_shares, |
||
1167 | max_notional=max_notional) |
||
1168 | self.register_trading_control(control) |
||
1169 | |||
1170 | @api_method |
||
1171 | def set_max_order_size(self, sid=None, max_shares=None, max_notional=None): |
||
1172 | """ |
||
1173 | Set a limit on the number of shares and/or dollar value of any single |
||
1174 | order placed for sid. Limits are treated as absolute values and are |
||
1175 | enforced at the time that the algo attempts to place an order for sid. |
||
1176 | |||
1177 | If an algorithm attempts to place an order that would result in |
||
1178 | exceeding one of these limits, raise a TradingControlException. |
||
1179 | """ |
||
1180 | control = MaxOrderSize(asset=sid, |
||
1181 | max_shares=max_shares, |
||
1182 | max_notional=max_notional) |
||
1183 | self.register_trading_control(control) |
||
1184 | |||
1185 | @api_method |
||
1186 | def set_max_order_count(self, max_count): |
||
1187 | """ |
||
1188 | Set a limit on the number of orders that can be placed within the given |
||
1189 | time interval. |
||
1190 | """ |
||
1191 | control = MaxOrderCount(max_count) |
||
1192 | self.register_trading_control(control) |
||
1193 | |||
1194 | @api_method |
||
1195 | def set_do_not_order_list(self, restricted_list): |
||
1196 | """ |
||
1197 | Set a restriction on which sids can be ordered. |
||
1198 | """ |
||
1199 | control = RestrictedListOrder(restricted_list) |
||
1200 | self.register_trading_control(control) |
||
1201 | |||
1202 | @api_method |
||
1203 | def set_long_only(self): |
||
1204 | """ |
||
1205 | Set a rule specifying that this algorithm cannot take short positions. |
||
1206 | """ |
||
1207 | self.register_trading_control(LongOnly()) |
||
1208 | |||
1209 | ############## |
||
1210 | # Pipeline API |
||
1211 | ############## |
||
1212 | @api_method |
||
1213 | @require_not_initialized(AttachPipelineAfterInitialize()) |
||
1214 | def attach_pipeline(self, pipeline, name, chunksize=None): |
||
1215 | """ |
||
1216 | Register a pipeline to be computed at the start of each day. |
||
1217 | """ |
||
1218 | if self._pipelines: |
||
1219 | raise NotImplementedError("Multiple pipelines are not supported.") |
||
1220 | if chunksize is None: |
||
1221 | # Make the first chunk smaller to get more immediate results: |
||
1222 | # (one week, then every half year) |
||
1223 | chunks = iter(chain([5], repeat(126))) |
||
1224 | else: |
||
1225 | chunks = iter(repeat(int(chunksize))) |
||
1226 | self._pipelines[name] = pipeline, chunks |
||
1227 | |||
1228 | # Return the pipeline to allow expressions like |
||
1229 | # p = attach_pipeline(Pipeline(), 'name') |
||
1230 | return pipeline |
||
1231 | |||
1232 | @api_method |
||
1233 | @require_initialized(PipelineOutputDuringInitialize()) |
||
1234 | def pipeline_output(self, name): |
||
1235 | """ |
||
1236 | Get the results of pipeline with name `name`. |
||
1237 | |||
1238 | Parameters |
||
1239 | ---------- |
||
1240 | name : str |
||
1241 | Name of the pipeline for which results are requested. |
||
1242 | |||
1243 | Returns |
||
1244 | ------- |
||
1245 | results : pd.DataFrame |
||
1246 | DataFrame containing the results of the requested pipeline for |
||
1247 | the current simulation date. |
||
1248 | |||
1249 | Raises |
||
1250 | ------ |
||
1251 | NoSuchPipeline |
||
1252 | Raised when no pipeline with the name `name` has been registered. |
||
1253 | |||
1254 | See Also |
||
1255 | -------- |
||
1256 | :meth:`zipline.pipeline.engine.PipelineEngine.run_pipeline` |
||
1257 | """ |
||
1258 | # NOTE: We don't currently support multiple pipelines, but we plan to |
||
1259 | # in the future. |
||
1260 | try: |
||
1261 | p, chunks = self._pipelines[name] |
||
1262 | except KeyError: |
||
1263 | raise NoSuchPipeline( |
||
1264 | name=name, |
||
1265 | valid=list(self._pipelines.keys()), |
||
1266 | ) |
||
1267 | return self._pipeline_output(p, chunks) |
||
1268 | |||
1269 | def _pipeline_output(self, pipeline, chunks): |
||
1270 | """ |
||
1271 | Internal implementation of `pipeline_output`. |
||
1272 | """ |
||
1273 | today = normalize_date(self.get_datetime()) |
||
1274 | try: |
||
1275 | data = self._pipeline_cache.unwrap(today) |
||
1276 | except Expired: |
||
1277 | data, valid_until = self._run_pipeline( |
||
1278 | pipeline, today, next(chunks), |
||
1279 | ) |
||
1280 | self._pipeline_cache = CachedObject(data, valid_until) |
||
1281 | |||
1282 | # Now that we have a cached result, try to return the data for today. |
||
1283 | try: |
||
1284 | return data.loc[today] |
||
1285 | except KeyError: |
||
1286 | # This happens if no assets passed the pipeline screen on a given |
||
1287 | # day. |
||
1288 | return pd.DataFrame(index=[], columns=data.columns) |
||
1289 | |||
1290 | def _run_pipeline(self, pipeline, start_date, chunksize): |
||
1291 | """ |
||
1292 | Compute `pipeline`, providing values for at least `start_date`. |
||
1293 | |||
1294 | Produces a DataFrame containing data for days between `start_date` and |
||
1295 | `end_date`, where `end_date` is defined by: |
||
1296 | |||
1297 | `end_date = min(start_date + chunksize trading days, |
||
1298 | simulation_end)` |
||
1299 | |||
1300 | Returns |
||
1301 | ------- |
||
1302 | (data, valid_until) : tuple (pd.DataFrame, pd.Timestamp) |
||
1303 | |||
1304 | See Also |
||
1305 | -------- |
||
1306 | PipelineEngine.run_pipeline |
||
1307 | """ |
||
1308 | days = self.trading_environment.trading_days |
||
1309 | |||
1310 | # Load data starting from the previous trading day... |
||
1311 | start_date_loc = days.get_loc(start_date) |
||
1312 | |||
1313 | # ...continuing until either the day before the simulation end, or |
||
1314 | # until chunksize days of data have been loaded. |
||
1315 | sim_end = self.sim_params.last_close.normalize() |
||
1316 | end_loc = min(start_date_loc + chunksize, days.get_loc(sim_end)) |
||
1317 | end_date = days[end_loc] |
||
1318 | |||
1319 | return \ |
||
1320 | self.engine.run_pipeline(pipeline, start_date, end_date), end_date |
||
1321 | |||
1322 | ################## |
||
1323 | # End Pipeline API |
||
1324 | ################## |
||
1325 | |||
1326 | @classmethod |
||
1327 | def all_api_methods(cls): |
||
1328 | """ |
||
1329 | Return a list of all the TradingAlgorithm API methods. |
||
1330 | """ |
||
1331 | return [ |
||
1332 | fn for fn in itervalues(vars(cls)) |
||
1333 | if getattr(fn, 'is_api_method', False) |
||
1334 | ] |
||
1335 |