Removed .ix

README.md

@@ -105,10 +105,13 @@ In findatapy/examples you will find several demos
 
 # Release Notes
 
+* 0.1.13 - findatapy (24 Aug 2020)
 * 0.1.12 - findatapy (06 May 2020)
 
 # Coding log
 
+* 24 Aug 2020
+    * Removed .ix references (to work with newer Pandas)
 * 06 May 2020
     * Amended function to remove points outside FX hours to exclude 1 Jan every year
     * RetStats can now resample time series (removed kurtosis)

findatapy/market/market.py

@@ -598,23 +598,23 @@ def extract_vol_surface_for_date(self, df, cross, date_index):
         df_surf = pandas.DataFrame(index=strikes, columns=tenor)
 
         for ten in tenor:
-            df_surf.ix["10DP", ten] = df.ix[date_index, cross + "V" + ten + ".close"] \
-                                      - (df.ix[date_index, cross + "10R" + ten + ".close"] / 2.0) \
-                                      + (df.ix[date_index, cross + "10B" + ten + ".close"])
+            df_surf[ten]["10DP"] = df[cross + "V" + ten + ".close"][date_index] \
+                                      - (df[cross + "10R" + ten + ".close"][date_index] / 2.0) \
+                                      + (df[cross + "10B" + ten + ".close"][date_index])
 
-            df_surf.ix["10DC", ten] = df.ix[date_index, cross + "V" + ten + ".close"] \
-                                      + (df.ix[date_index, cross + "10R" + ten + ".close"] / 2.0) \
-                                      + (df.ix[date_index, cross + "10B" + ten + ".close"])
+            df_surf[ten]["10DC"] = df[cross + "V" + ten + ".close"][date_index] \
+                                      + (df[cross + "10R" + ten + ".close"][date_index] / 2.0) \
+                                      + (df[cross + "10B" + ten + ".close"][date_index])
 
-            df_surf.ix["25DP", ten] = df.ix[date_index, cross + "V" + ten + ".close"] \
-                                      - (df.ix[date_index, cross + "25R" + ten + ".close"] / 2.0) \
-                                      + (df.ix[date_index, cross + "25B" + ten + ".close"])
+            df_surf[ten]["25DP"] = df[cross + "V" + ten + ".close"][date_index] \
+                                      - (df[cross + "25R" + ten + ".close"][date_index] / 2.0) \
+                                      + (df[cross + "25B" + ten + ".close"][date_index])
 
-            df_surf.ix["25DC", ten] = df.ix[date_index, cross + "V" + ten + ".close"] \
-                                      + (df.ix[date_index, cross + "25R" + ten + ".close"] / 2.0) \
-                                      + (df.ix[date_index, cross + "25B" + ten + ".close"])
+            df_surf[ten]["25DC"] = df[cross + "V" + ten + ".close"][date_index] \
+                                      + (df[cross + "25R" + ten + ".close"][date_index] / 2.0) \
+                                      + (df[cross + "25B" + ten + ".close"][date_index])
 
-            df_surf.ix["ATM", ten] = df.ix[date_index, cross + "V" + ten + ".close"]
+            df_surf[ten]["ATM"] = df[cross + "V" + ten + ".close"][date_index]
 
         return df_surf
 

findatapy/timeseries/calculations.py

@@ -152,10 +152,10 @@ def calculate_individual_trade_gains(self, signal_data_frame, strategy_returns_d
             # TODO experiment with quicker ways of writing below?
             # for val in col.index:
             # trade_returns.set_value(val, col_name, col[val])
-            # trade_returns.ix[val, col_name] = col[val]
+            # trade_returns[col_name][val] = col[val]
 
             date_indices = trade_returns.index.searchsorted(col.index)
-            trade_returns.ix[date_indices, col_name] = col
+            trade_returns[col_name][date_indices] = col
 
         return trade_returns
 
@@ -1175,6 +1175,14 @@ def average_by_hour_min_of_day_pretty_output(self, data_frame):
 
         return data_frame
 
+    def average_by_hour_min_sec_of_day_pretty_output(self, data_frame):
+        data_frame = data_frame. \
+            groupby([data_frame.index.hour.rename('hour'), data_frame.index.minute.rename('minute'), data_frame.index.minute.rename('second')]).mean()
+
+        data_frame.index = data_frame.index.map(lambda t: datetime.time(*t))
+
+        return data_frame
+
     def all_by_hour_min_of_day_pretty_output(self, data_frame):
 
         df_new = []
@@ -1193,7 +1201,7 @@ def average_by_year_hour_min_of_day_pretty_output(self, data_frame):
         # time_of_day = []
         #
         # for year in years:
-        #     temp = data_frame.ix[data_frame.index.year == year]
+        #     temp = data_frame[data_frame.index.year == year]
         #     time_of_day.append(temp.groupby(temp.index.time).mean())
         #
         # data_frame = pandas.concat(time_of_day, axis=1, keys = years)

findatapy/timeseries/dataquality.py

@@ -122,8 +122,8 @@ def percentage_nan_between_start_finish_dates(self, df, df_properties, asset_fie
         c_new = [x.split(".")[0] for x in df.columns]
 
         index = df_dates.index.searchsorted(c_new)
-        start_date = df_dates.ix[index, start_date_field]
-        finish_date = df_dates.ix[index, finish_date_field]
+        start_date = df_dates[start_date_field][index]
+        finish_date = df_dates[finish_date_field][index]
 
         for i in range(0, len(df.columns)):
             df_sub = df[df.columns[i]]

findatapy/timeseries/filter.py

@@ -146,7 +146,7 @@ def filter_time_series_by_holidays(self, data_frame, cal = 'FX', holidays_list =
 
         # optimal case for weekdays: remove Saturday and Sunday
         if (cal == 'WEEKDAY'):
-            return data_frame.ix[data_frame.index.dayofweek <= 4]
+            return data_frame[data_frame.index.dayofweek <= 4]
 
         # select only those holidays in the sample
         holidays_start = self.get_holidays(data_frame.index[0], data_frame.index[-1], cal, holidays_list = holidays_list)
@@ -179,7 +179,7 @@ def filter_time_series_by_holidays(self, data_frame, cal = 'FX', holidays_list =
         # indices = list(range(counter, len(floored_dates)))
         # indices_to_keep = indices_to_keep + indices
         #
-        # data_frame_filtered = data_frame.ix[indices_to_keep]
+        # data_frame_filtered = data_frame[indices_to_keep]
 
         if data_frame.index.tz is None:
             holidays_start = holidays_start.tz_localize(None)
@@ -191,8 +191,8 @@ def filter_time_series_by_holidays(self, data_frame, cal = 'FX', holidays_list =
         for i in range(0, len(holidays_start)):
 
 
-            data_frame_temp = data_frame_left.ix[data_frame_left.index < holidays_start[i]]
-            data_frame_left = data_frame_left.ix[data_frame_left.index >= holidays_end[i]]
+            data_frame_temp = data_frame_left[data_frame_left.index < holidays_start[i]]
+            data_frame_left = data_frame_left[data_frame_left.index >= holidays_end[i]]
 
             data_frame_filtered.append(data_frame_temp)
 
@@ -385,7 +385,7 @@ def filter_time_series_aux(self, start_date, finish_date, data_frame, offset):
         #     if (0 <= start_index + offset < len(data_frame.index)):
         #         start_index = start_index + offset
         #
-        #         # data_frame = data_frame.ix[start_date < data_frame.index]
+        #         # data_frame = data_frame[start_date < data_frame.index]
         #
         # if(finish_date is not None):
         #     finish_index = data_frame.index.searchsorted(finish_date)

findatapy_examples/fxspotdata_example.py

@@ -29,7 +29,7 @@
     # run_example = 5 - download second FX data from Bloomberg
     # run_example = 6 - download free tick data from FXCM example (compare with DukasCopy)
 
-    run_example = 6
+    run_example = 1
 
     if run_example == 1 or run_example == 0:
 

setup.py

@@ -5,7 +5,7 @@
 tickers, using configuration files. There is also functionality which is particularly useful for those downloading FX market data."""
 
 setup(name='findatapy',
-      version='0.1.12',
+      version='0.1.13',
       description='Market data library',
       author='Saeed Amen',
       author_email='[email protected]',