Develop 026 - work in progress

Examples/Basic/tutorial1.py

@@ -51,6 +51,12 @@ def main():
     ldf_data = c_dataobj.get_data(ldt_timestamps, ls_symbols, ls_keys)
     d_data = dict(zip(ls_keys, ldf_data))
 
+    # Filling the data for NAN
+    for s_key in ls_keys:
+        d_data[s_key] = d_data[s_key].fillna(method='ffill')
+        d_data[s_key] = d_data[s_key].fillna(method='bfill')
+        d_data[s_key] = d_data[s_key].fillna(1.0)
+
     # Getting the numpy ndarray of close prices.
     na_price = d_data['close'].values
 

Examples/Basic/tutorial3.py

@@ -77,6 +77,7 @@ def main():
     # Filling the data.
     df_rets = df_rets.fillna(method='ffill')
     df_rets = df_rets.fillna(method='bfill')
+    df_rets = df_rets.fillna(1.0)
 
     # Numpy matrix of filled data values
     na_rets = df_rets.values

Examples/Basic/tutorial5.py

@@ -46,6 +46,9 @@ def main():
 
     # Reading just the close prices
     df_close = c_dataobj.get_data(ldt_timestamps, ls_symbols, "close")
+    df_close = df_close.fillna(method='ffill')
+    df_close = df_close.fillna(method='bfill')
+    df_close = df_close.fillna(1.0)
 
     # Creating the allocation dataframe
     # We offset the time for the simulator to have atleast one

Examples/Validation.py

@@ -151,7 +151,7 @@
 ls_symbols = ['MSFT', 'GOOG']
 
 # Creating an object of the dataaccess class with Yahoo as the source.
-c_dataobj = da.DataAccess('Yahoo')
+c_dataobj = da.DataAccess('Yahoo', verbose=True)
 # Reading adjusted_close prices
 df_close = c_dataobj.get_data(ldt_timestamps, ls_symbols, "close")
 print df_close

MANIFEST.in

@@ -14,4 +14,5 @@ include QSTK/QSData/Yahoo/*.csv'
 include QSTK/QSData/Yahoo/Lists/*.txt
 include QSTK/qstkutil/NYSE_dates.txt
 include QSTK/qstktest/*.csv
-include QSTK/qstkstudy/sp500.txt
+include QSTK/qstkstudy/sp500.txt
+include QSTK/__init__.py

QSTK/QSData/Yahoo/Lists/sp5002008.txt

@@ -91,7 +91,6 @@ CBS
 CCE
 CCL
 CCMO
-CCTYQ
 CEG
 CELG
 CFC+A
@@ -218,7 +217,6 @@ HNZ
 HOG
 HON
 HOT
-HPC
 HPQ
 HRB
 HSP

QSTK/qstkfeat/features.py

@@ -341,7 +341,8 @@ def featVolumeDelta( dData, lLookback=30, b_human=False ):
 
 def featAroon( dData, bDown=False, lLookback=25, b_human=False ):
     '''
-    @summary: Calculate Aroon - indicator indicating days since a 25-day high/low, weighted between 0 and 100
+    @summary: Calculate Aroon - indicator indicating days since a 25-day 
+              high/low, weighted between 0 and 100
     @param dData: Dictionary of data to use
     @param bDown: If false, calculates aroonUp (high), else aroonDown (lows)
     @param lLookback: Days to lookback to calculate high/low from
@@ -351,19 +352,21 @@ def featAroon( dData, bDown=False, lLookback=25, b_human=False ):
 
     dfPrice = dData['close']
 
-    #''' Feature DataFrame will be 1:1, we can use the price as a template '''
-    dfRet = pand.DataFrame( index=dfPrice.index, columns=dfPrice.columns, 
+    #Feature DataFrame will be 1:1, we can use the price as a template
+    dfRet = pd.DataFrame( index=dfPrice.index, columns=dfPrice.columns, 
                             data=np.zeros(dfPrice.shape) )
 
-    #''' Loop through time '''
+    #Loop through time
     for i in range(dfPrice.shape[0]):
-        if( (i-lLookback)+1 < 0 ):
+        if( (i-lLookback) < 0 ):
             dfRet.ix[i,:] = np.NAN
         else:
             if bDown:
-                dfRet.ix[i,:] = dfPrice.values[i+1:(i-lLookback)+1:-1,:].argmin(axis=0)
+                dfRet.ix[i,:] = dfPrice.values[i:(i-lLookback):-1,:].argmin(
+                                axis=0)
             else:
-                dfRet.ix[i,:] = dfPrice.values[i+1:(i-lLookback)+1:-1,:].argmax(axis=0)
+                dfRet.ix[i,:] = dfPrice.values[i:(i-lLookback):-1,:].argmax(
+                                axis=0)
 
     dfRet = ((lLookback - 1.) - dfRet) / (lLookback - 1.) * 100.
 

QSTK/qstkstudy/EventProfiler.py

@@ -20,12 +20,15 @@
 import QSTK.qstkutil.qsdateutil as du
 
 
-def eventprofiler(df_events, d_data, i_lookback=20, i_lookforward=20,
+def eventprofiler(df_events_arg, d_data, i_lookback=20, i_lookforward=20,
                 s_filename='study', b_market_neutral=True, b_errorbars=True,
                 s_market_sym='SPY'):
     ''' Event Profiler for an event matix'''
     df_close = d_data['close'].copy()
     df_rets = df_close.copy()
+
+    # Do not modify the original event dataframe.
+    df_events = df_events_arg.copy()
     tsu.returnize0(df_rets.values)
 
     if b_market_neutral == True:
@@ -40,7 +43,8 @@ def eventprofiler(df_events, d_data, i_lookback=20, i_lookforward=20,
     df_events.values[-i_lookforward:, :] = np.NaN
 
     # Number of events
-    i_no_events = int(np.nansum(df_events.values))
+    i_no_events = int(np.logical_not(np.isnan(df_events.values)).sum())
+    assert i_no_events > 0, "Zero events in the event matrix"
     na_event_rets = "False"
 
     # Looking for the events and pushing them to a matrix
@@ -53,6 +57,9 @@ def eventprofiler(df_events, d_data, i_lookback=20, i_lookforward=20,
                 else:
                     na_event_rets = np.vstack((na_event_rets, na_ret))
 
+    if len(na_event_rets.shape) == 1:
+        na_event_rets = np.expand_dims(na_event_rets, axis=0)
+
     # Computing daily rets and retuns
     na_event_rets = np.cumprod(na_event_rets + 1, axis=1)
     na_event_rets = (na_event_rets.T / na_event_rets[:, i_lookback]).T