search

mementum / bta-lib

Technical Analysis library in pandas for backtesting algotrading and quantitative analysis
MIT License
449 stars 106 forks source link

dm and dx indicators seem broken #10

Open JavierAntoran opened 3 years ago

JavierAntoran commented 3 years ago

When I run

btalib.dx(btc_df.high, btc_df.low, btc_df.close)

or 

btalib.dm(btc_df.high, btc_df.low)

with btc_df.high, btc_df.low and btc_df.close being columns from a pandas data frame (series of floats).

I get:

TypeError Traceback (most recent call last)

in 1 # import talib 2 ----> 3 btalib.dx(btc_df.high, btc_df.low, btc_df.close) ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/btalib/indicator.py in __call__(cls, *args, **kwargs) 150 # Auto-call base classes 151 for b_init in reversed(list(dict.fromkeys(b.__init__ for b in bases))): --> 152 b_init(self, *args, **kwargs) 153 154 # delete old aliases only meant for operational purposes ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/btalib/indicators/directionalmove.py in __init__(self) 95 if self._plus: # upvmoe > 0 and where upmove > downmove 96 pdm = upmove.clip(lower=0.0) * (upmove > downmove) ---> 97 self._pdm = self._smoother(pdm, **smoothargs) 98 99 if self._minus: # downmove > 0 and where downmove > upmove ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/btalib/indicator.py in __call__(cls, *args, **kwargs) 150 # Auto-call base classes 151 for b_init in reversed(list(dict.fromkeys(b.__init__ for b in bases))): --> 152 b_init(self, *args, **kwargs) 153 154 # delete old aliases only meant for operational purposes ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/btalib/indicators/directionalmove.py in __init__(self) 50 p = self.p.period 51 _ewm = self.i0._ewm(span=p, _pearly=self.p._pearly, _seed=self.p._seed) ---> 52 self.o.smacc = _ewm._lfilter(alpha=1.0, beta=(p - 1) / p) 53 54 def _talib(self, kwdict): ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/btalib/meta/lines.py in _lfilter(self, alpha, beta) 341 return scipy.signal.lfilter([alpha], [1.0, -beta], x) 342 --> 343 return self._apply(_sp_lfilter) # trigger __getattr__ for _apply 344 345 def _mean(self): # meant for ewm with dynamic alpha ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/btalib/meta/lines.py in call_op(*args, **kwargs) 377 sargs.append(arg) 378 --> 379 result[self._minidx:] = r = op(*sargs, **kwargs) # run/store 380 result = result.astype(r.dtype, copy=False) 381 return self._line._clone(result, period=self._minperiod) ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/pandas/core/window/rolling.py in _apply(self, func, name, numba_cache_key, **kwargs) 467 return result 468 --> 469 return self._apply_blockwise(homogeneous_func, name) 470 471 def aggregate(self, func, *args, **kwargs): ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/pandas/core/window/rolling.py in _apply_blockwise(self, homogeneous_func, name) 382 """ 383 if self._selected_obj.ndim == 1: --> 384 return self._apply_series(homogeneous_func, name) 385 386 obj = self._create_data(self._selected_obj) ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/pandas/core/window/rolling.py in _apply_series(self, homogeneous_func, name) 371 raise DataError("No numeric types to aggregate") from err 372 --> 373 result = homogeneous_func(values) 374 return obj._constructor(result, index=obj.index, name=obj.name) 375 ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/pandas/core/window/rolling.py in homogeneous_func(values) 459 result = np.apply_along_axis(calc, self.axis, values) 460 else: --> 461 result = calc(values) 462 result = np.asarray(result) 463 ~/Documents/external_projects/trading/venv/lib/python3.8/site-packages/pandas/core/window/rolling.py in calc(x) 453 closed=self.closed, 454 ) --> 455 return func(x, start, end, min_periods) 456 457 with np.errstate(all="ignore"): TypeError: _sp_lfilter() takes 1 positional argument but 4 were given ------ Python version 3.8.2 Package versions: bta-lib 1.0.0 numpy 1.19.5 pandas 1.2.1 scipy 1.6.0
drhighliner commented 3 years ago

It has to do with the pandas version. You need to downgrade to pandas 1.1.5 as bta-lib hasn't been updated to the 1.2.x versions yet. (maybe we can have an update here?) This should have been added to install requirements for bta-lib anyway, but that's another topic.

Let me know, if this works for you.