Fixed issue with EventStudy

README.md

@@ -167,6 +167,9 @@ In finmarketpy/examples you will find several examples, including some simple tr
 
 # finmarketpy log
 
+* 21 Mar 2021
+ * Fixed issue with EventStudy
+ * Refactored TradingModel when constructing weights
 * 22 Feb 2021
  * Fix bug when plotting IR of strategies
 * 16 Feb 2021

finmarketpy/backtest/backtestrequest.py

@@ -33,6 +33,7 @@ def __init__(self):
 
  # output parameters for backtest (should we add returns statistics on legends, write CSVs with returns etc.)
  self.__plot_start = None
+ self.__plot_finish = None
  self.__calc_stats = True
  self.__write_csv = False
  self.__write_csv_pnl = False
@@ -43,6 +44,8 @@ def __init__(self):
 
  self.__tech_params = TechParams()
 
+ self.__portfolio_weight_construction = None
+
  # default parameters for portfolio level vol adjustment
  self.__portfolio_vol_adjust = False
  self.__portfolio_vol_period_shift = 0
@@ -101,6 +104,14 @@ def plot_start(self): return self.__plot_start
 
  @plot_start.setter
  def plot_start(self, plot_start): self.__plot_start = plot_start
+
+ @property
+ def plot_finish(self):
+ return self.__plot_finish
+
+ @plot_finish.setter
+ def plot_finish(self, plot_finish):
+ self.__plot_finish = plot_finish
 
  @property
  def calc_stats(self): return self.__calc_stats
@@ -139,8 +150,17 @@ def trading_field(self): return self.__trading_field
 
  @trading_field.setter
  def trading_field(self, trading_field): self.__trading_field = trading_field
+
+ @property
+ def portfolio_weight_construction(self):
+ return self.__portfolio_weight_construction
 
- ##### properties for portfolio level volatility adjustment
+ @portfolio_weight_construction.setter
+ def portfolio_weight_construction(self, portfolio_weight_construction):
+ self.__portfolio_weight_construction = portfolio_weight_construction
+
+ ##### Properties for portfolio level volatility adjustment
+
  @property
  def portfolio_vol_adjust(self): return self.__portfolio_vol_adjust
 

finmarketpy/economics/eventstudy.py

@@ -13,7 +13,7 @@
 #
 
 import math
-
+import numpy as np
 from pandas.tseries.offsets import CustomBusinessDay
 
 from findatapy.timeseries import Calculations, Timezone, Filter, Calendar
@@ -28,8 +28,7 @@ def __init__(self):
  return
 
  def get_economic_event_ret_over_custom_event_day(self, data_frame_in, event_dates, name, event, start, end,
- lagged=False,
- NYC_cutoff=10):
+ lagged=False, NYC_cutoff=10):
 
  filter = Filter()
  event_dates = filter.filter_time_series_by_date(start, end, event_dates)
@@ -46,7 +45,7 @@ def get_economic_event_ret_over_custom_event_day(self, data_frame_in, event_date
 
  event_dates = calendar.floor_date(event_dates)
 
- # realised is traditionally on later day eg. 3rd Jan realised ON is 2nd-3rd Jan realised
+ # Realised is traditionally on later day eg. 3rd Jan realised ON is 2nd-3rd Jan realised
  # so if Fed meeting is on 2nd Jan later, then we need realised labelled on 3rd (so minus a day)
  # implied_vol expires on next day eg. 3rd Jan implied_vol ON is 3rd-4th Jan implied_vol
 
@@ -62,9 +61,32 @@ def get_economic_event_ret_over_custom_event_day(self, data_frame_in, event_date
 
  return data_frame_events
 
+ def get_daily_moves_over_custom_event(self, data_frame_rets, ef_time_frame, vol=False,
+ day_start=20, days=20, day_offset=0, create_index=False,
+ resample=False, cumsum=True, adj_cumsum_zero_point=False,
+ adj_zero_point=2):
+
+ return self.get_intraday_moves_over_custom_event(data_frame_rets, ef_time_frame, vol=vol,
+ minute_start=day_start, mins=days, min_offset=day_offset,
+ create_index=create_index, resample=resample, freq='days',
+ cumsum=cumsum, adj_cumsum_zero_point=adj_cumsum_zero_point,
+ adj_zero_point=adj_zero_point)
+
+ def get_weekly_moves_over_custom_event(self, data_frame_rets, ef_time_frame, vol=False,
+ day_start=20, days=20, day_offset=0, create_index=False,
+ resample=False, cumsum=True, adj_cumsum_zero_point=False,
+ adj_zero_point=2):
+
+ return self.get_intraday_moves_over_custom_event(data_frame_rets, ef_time_frame, vol=vol,
+ minute_start=day_start, mins=days, min_offset=day_offset,
+ create_index=create_index, resample=resample, freq='weeks',
+ cumsum=cumsum, adj_cumsum_zero_point=adj_cumsum_zero_point,
+ adj_zero_point=adj_zero_point)
+
  def get_intraday_moves_over_custom_event(self, data_frame_rets, ef_time_frame, vol=False,
  minute_start=5, mins=3 * 60, min_offset=0, create_index=False,
- resample=False, freq='minutes', cumsum=True):
+ resample=False, freq='minutes', cumsum=True, adj_cumsum_zero_point=False,
+ adj_zero_point=2):
 
  filter = Filter()
 
@@ -78,45 +100,75 @@ def get_intraday_moves_over_custom_event(self, data_frame_rets, ef_time_frame, v
  ann_factor = 252 * 1440
  elif freq == 'days':
  ef_time = ef_time_frame.index.normalize()
- ef_time_start = ef_time - timedelta(days=minute_start)
- ef_time_end = ef_time + timedelta(days=mins)
+ ef_time_start = ef_time - pandas.tseries.offsets.BusinessDay() * minute_start
+ ef_time_end = ef_time + pandas.tseries.offsets.BusinessDay() * mins
  ann_factor = 252
+ elif freq == 'weeks':
+ ef_time = ef_time_frame.index.normalize()
+ ef_time_start = ef_time - pandas.tseries.offsets.Week() * minute_start
+ ef_time_end = ef_time + pandas.tseries.offsets.Week() * mins
+ ann_factor = 52
 
- ords = range(-minute_start + min_offset, mins + min_offset)
+ ords = list(range(-minute_start + min_offset, mins + min_offset))
+ lst_ords = list(ords)
 
- # all data needs to be equally spaced
+ # All data needs to be equally spaced
  if resample:
  # make sure time series is properly sampled at 1 min intervals
- data_frame_rets = data_frame_rets.resample('1min')
- data_frame_rets = data_frame_rets.fillna(value=0)
- data_frame_rets = filter.remove_out_FX_out_of_hours(data_frame_rets)
-
- data_frame_rets['Ind'] = numpy.nan
+ if freq == 'minutes':
+ data_frame_rets = data_frame_rets.resample('1min').last()
+ data_frame_rets = data_frame_rets.fillna(value=0)
+ data_frame_rets = filter.remove_out_FX_out_of_hours(data_frame_rets)
+ elif freq == 'daily':
+ data_frame_rets = data_frame_rets.resample('B').last()
+ data_frame_rets = data_frame_rets.fillna(value=0)
+ elif freq == 'daily':
+ data_frame_rets = data_frame_rets.resample('W').last()
+ data_frame_rets = data_frame_rets.fillna(value=0)
+
+ # data_frame_rets['Ind'] = numpy.nan
 
  start_index = data_frame_rets.index.searchsorted(ef_time_start)
  finish_index = data_frame_rets.index.searchsorted(ef_time_end)
 
- # not all observation windows will be same length (eg. last one?)
+ # Not all observation windows will be same length (eg. last one?)
+
+ # # fill the indices which represent minutes
+ # # TODO vectorise this!
+ # for i in range(0, len(ef_time_frame.index)):
+ # try:
+ # data_frame_rets['Ind'][start_index[i]:finish_index[i]] = ords
+ # except:
+ # data_frame_rets['Ind'][start_index[i]:finish_index[i]] = ords[0:(finish_index[i] - start_index[i])]
+ #
+ # data_frame_rets['Rel'] = numpy.nan
+ #
+ # # Set the release dates
+ # data_frame_rets['Rel'][start_index] = ef_time # set entry points
+ # data_frame_rets['Rel'][finish_index + 1] = numpy.zeros(len(start_index)) # set exit points
+ # data_frame_rets['Rel'] = data_frame_rets['Rel'].fillna(method='pad') # fill down signals
+ #
+ # data_frame_rets = data_frame_rets[pandas.notnull(data_frame_rets['Ind'])] # get rid of other
+ #
+ # data_frame = data_frame_rets.pivot(index='Ind',
+ # columns='Rel', values=data_frame_rets.columns[0])
+
+ data_frame = pandas.DataFrame(index=ords, columns=ef_time_frame.index)
 
- # fill the indices which represent minutes
- # TODO vectorise this!
  for i in range(0, len(ef_time_frame.index)):
- try:
- data_frame_rets['Ind'][start_index[i]:finish_index[i]] = ords
- except:
- data_frame_rets['Ind'][start_index[i]:finish_index[i]] = ords[0:(finish_index[i] - start_index[i])]
+ #try:
+ vals = data_frame_rets.iloc[start_index[i]:finish_index[i]].values
 
- data_frame_rets['Rel'] = numpy.nan
+  print(vals.size)
 
- # Set the release dates
- data_frame_rets['Rel'][start_index] = ef_time # set entry points
- data_frame_rets['Rel'][finish_index + 1] = numpy.zeros(len(start_index)) # set exit points
- data_frame_rets['Rel'] = data_frame_rets['Rel'].fillna(method='pad') # fill down signals
+ if len(vals) < len(lst_ords):
+ extend = np.zeros((len(lst_ords) - len(vals), 1)) * np.nan
+ vals = np.append(vals, extend)
 
- data_frame_rets = data_frame_rets[pandas.notnull(data_frame_rets['Ind'])] # get rid of other
+  data_frame[ef_time_frame.index[i]] = vals
 
- data_frame = data_frame_rets.pivot(index='Ind',
-  columns='Rel', values=data_frame_rets.columns[0])
+  #except:
+ # data_frame_rets['Ind'][start_index[i]:finish_index[i]] = ords[0:(finish_index[i] - start_index[i])]
 
  data_frame.index.names = [None]
 
@@ -126,11 +178,19 @@ def get_intraday_moves_over_custom_event(self, data_frame_rets, ef_time_frame, v
  data_frame = calculations.create_mult_index(data_frame)
  else:
  if vol is True:
- # annualise (if vol)
+ # Annualise (if vol)
  data_frame = data_frame.rolling(center=False, window=5).std() * math.sqrt(ann_factor)
  elif cumsum:
+
  data_frame = data_frame.cumsum()
 
+ # Adjust DataFrame so zero point shows zero returns
+ if adj_cumsum_zero_point:
+ ind = abs(minute_start) - adj_zero_point
+
+ for i, c in enumerate(data_frame.columns):
+ data_frame[c] = data_frame[c] - data_frame[c].values[ind]
+
  return data_frame
 
  def get_surprise_against_intraday_moves_over_custom_event(
@@ -426,7 +486,7 @@ def get_daily_moves_over_event(self):
  # TODO
  pass
 
- # return only US events etc. by dates
+ # Return only US events etc. by dates
  def get_intraday_moves_over_event(self, data_frame_rets, cross, event_fx, event_name, start, end, vol, mins=3 * 60,
  min_offset=0, create_index=False, resample=False, freq='minutes'):
 

finmarketpy/economics/techindicator.py

@@ -42,7 +42,6 @@ class TechIndicator(object):
  """
 
  def __init__(self):
- self.logger = LoggerManager().getLogger(__name__)
  self._techind = None
  self._signal = None
 
@@ -52,6 +51,9 @@ def create_tech_ind(
  name,
  tech_params,
  data_frame_non_nan_early=None):
+
+ logger = LoggerManager().getLogger(__name__)
+
  self._signal = None
  self._techind = None
 
@@ -66,7 +68,7 @@ def create_tech_ind(
  if name == "SMA":
 
  if (data_frame_non_nan_early is not None):
- # calculate the lagged sum of the n-1 point
+ # Calculate the lagged sum of the n-1 point
  if pd.__version__ < '0.17':
  rolling_sum = pd.rolling_sum(
  data_frame.shift(1).rolling,
@@ -120,7 +122,7 @@ def create_tech_ind(
  x + " EMA" for x in data_frame.columns.values]
 
  elif name == "ROC":
-
+ # Rate of change
  if (data_frame_non_nan_early is not None):
  self._techind = data_frame_early / \
  data_frame.shift(tech_params.roc_period) - 1
@@ -139,6 +141,7 @@ def create_tech_ind(
  x + " ROC" for x in data_frame.columns.values]
 
  elif name == "polarity":
+ # If data is positive or negative
  self._techind = data_frame
 
  narray = np.where(self._techind > 0, 1, -1)
@@ -151,6 +154,7 @@ def create_tech_ind(
  x + " Polarity" for x in data_frame.columns.values]
 
  elif name == "SMA2":
+ # Double moving average
  sma = data_frame.rolling(
  window=tech_params.sma_period,
  center=False).mean()
@@ -171,6 +175,8 @@ def create_tech_ind(
  self._techind = pd.concat([sma, sma2], axis=1)
 
  elif name in ['RSI']:
+ # Relative Strength Index
+
  # delta = data_frame.diff()
  #
  # dUp, dDown = delta.copy(), delta.copy()
@@ -239,7 +245,7 @@ def create_tech_ind(
  x + " RSI Signal" for x in data_frame.columns.values]
 
  elif name in ["BB"]:
- # calcuate Bollinger bands
+ # Calcuate Bollinger bands
  mid = data_frame.rolling(
  center=False, window=tech_params.bb_period).mean()
  mid.columns = [x + " BB Mid" for x in data_frame.columns.values]
@@ -257,7 +263,7 @@ def create_tech_ind(
  index=mid.index,
  columns=data_frame.columns)
 
- # calculate signals
+ # Calculate signals (buy/sell)
  signal = data_frame.copy()
 
  buys = signal > upper
@@ -281,7 +287,7 @@ def create_tech_ind(
 
  self._techind = pd.concat([lower, mid, upper], axis=1)
  elif name == "long-only":
- # have +1 signals only
+ # Have +1 signals only
  self._techind = data_frame # the technical indicator is just "prices"
 
  narray = np.ones((len(data_frame.index), len(data_frame.columns)))
@@ -293,22 +299,35 @@ def create_tech_ind(
  self._techind.columns = [
  x + " Long Only" for x in data_frame.columns.values]
 
+ elif name == "short-only":
+ # Have -1 signals only
+ self._techind = data_frame # the technical indicator is just "prices"
+
+ narray = np.ones((len(data_frame.index), len(data_frame.columns)))
+
+ self._signal = pd.DataFrame(index=data_frame.index, data=narray)
+ self._signal.columns = [
+ x + " Short Only Signal" for x in data_frame.columns.values]
+
+ self._techind.columns = [
+ x + " Short Only" for x in data_frame.columns.values]
+
  elif name == "ATR":
- # get all the asset names (assume we have names 'close', 'low', 'high' in the Data)
+ # Get all the asset names (assume we have names 'close', 'low', 'high' in the Data)
  # keep ordering of assets
  asset_name = list(OrderedDict.fromkeys(
  [x.split('.')[0] for x in data_frame.columns]))
 
  df = []
 
- # can improve the performance of this if vectorise more!
+ # Can improve the performance of this if vectorise more!
  for a in asset_name:
 
  close = [a + '.close']
  low = [a + '.low']
  high = [a + '.high']
 
- # if we don't fill NaNs, we need to remove those rows and then
+ # If we don't fill NaNs, we need to remove those rows and then
  # calculate the ATR
  if not(tech_params.fillna):
  data_frame_short = data_frame[[close[0], low[0], high[0]]]
@@ -425,7 +444,7 @@ def get_signal(self):
 ##########################################################################
 
 
-class TechParams:
+class TechParams(object):
  """Holds parameters for calculation of technical indicators.
 
  """