Merge pull request #289 from quantopian/pandas_round

Better table output

.travis.yml

@@ -29,7 +29,7 @@ install:
   - pip install nose_parameterized contextlib2 logbook==0.10.1
   #- pip install --no-deps git+https://github.com/quantopian/zipline
   - if [ ${TRAVIS_PYTHON_VERSION:0:1} == "2" ]; then conda install --yes mock enum34; fi
-  - pip install --no-deps git+https://github.com/Theano/Theano.git@557fe6fab49262706241f0e774f3269a2ec0275b
+  - pip install --no-deps git+https://github.com/Theano/Theano.git@rel-0.8.1
   - pip install --no-deps git+https://github.com/pymc-devs/pymc3.git
   - python setup.py build_ext --inplace
 

pyfolio/bayesian.py

@@ -1,5 +1,5 @@
 #
-# Copyright 2015 Quantopian, Inc.
+# Copyright 2016 Quantopian, Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

pyfolio/interesting_periods.py

@@ -1,5 +1,5 @@
 #
-# Copyright 2015 Quantopian, Inc.
+# Copyright 2016 Quantopian, Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

pyfolio/plotting.py

@@ -1,5 +1,5 @@
 #
-# Copyright 2015 Quantopian, Inc.
+# Copyright 2016 Quantopian, Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -191,8 +191,7 @@ def plot_monthly_returns_heatmap(returns, ax=None, **kwargs):
 
     monthly_ret_table = timeseries.aggregate_returns(returns,
                                                      'monthly')
-    monthly_ret_table = monthly_ret_table.unstack()
-    monthly_ret_table = np.round(monthly_ret_table, 3)
+    monthly_ret_table = monthly_ret_table.unstack().round(3)
 
     sns.heatmap(
         monthly_ret_table.fillna(0) *
@@ -534,13 +533,13 @@ def show_perf_stats(returns, factor_returns, live_start_date=None,
         returns_backtest = returns[returns.index < live_start_date]
         returns_live = returns[returns.index > live_start_date]
 
-        perf_stats_live = np.round(perf_func(
+        perf_stats_live = perf_func(
             returns_live,
-            factor_returns=factor_returns).values, 2)
+            factor_returns=factor_returns)
 
-        perf_stats_all = np.round(perf_func(
+        perf_stats_all = perf_func(
             returns,
-            factor_returns=factor_returns).values, 2)
+            factor_returns=factor_returns)
 
         print('Out-of-Sample Months: ' +
               str(int(len(returns_live) / APPROX_BDAYS_PER_MONTH)))
@@ -550,18 +549,21 @@ def show_perf_stats(returns, factor_returns, live_start_date=None,
     print('Backtest Months: ' +
           str(int(len(returns_backtest) / APPROX_BDAYS_PER_MONTH)))
 
-    perf_stats = np.round(perf_func(
+    perf_stats = perf_func(
         returns_backtest,
-        factor_returns=factor_returns).values, 2)
+        factor_returns=factor_returns)
 
     if live_start_date is not None:
         perf_stats = pd.concat(OrderedDict([
             ('Backtest', perf_stats),
             ('Out of sample', perf_stats_live),
             ('All history', perf_stats_all),
         ]), axis=1)
+    else:
+        perf_stats = pd.DataFrame(perf_stats, columns=['Backtest'])
 
-    print(perf_stats)
+    utils.print_table(perf_stats, name='Performance statistics',
+                      fmt='{0:.2f}')
 
 
 def plot_rolling_returns(returns,
@@ -907,28 +909,23 @@ def show_and_plot_top_positions(returns, positions_alloc,
         positions_alloc)
 
     if show_and_plot == 1 or show_and_plot == 2:
-        print("\n")
-        print('Top 10 long positions of all time (and max%)')
-        print(pd.DataFrame(df_top_long).index.values)
-        print(np.round(pd.DataFrame(df_top_long)[0].values, 3))
-        print("\n")
-
-        print('Top 10 short positions of all time (and max%)')
-        print(pd.DataFrame(df_top_short).index.values)
-        print(np.round(pd.DataFrame(df_top_short)[0].values, 3))
-        print("\n")
-
-        print('Top 10 positions of all time (and max%)')
-        print(pd.DataFrame(df_top_abs).index.values)
-        print(np.round(pd.DataFrame(df_top_abs)[0].values, 3))
-        print("\n")
+        utils.print_table(pd.DataFrame(df_top_long * 100, columns=['max']),
+                          fmt='{0:.2f}%',
+                          name='Top 10 long positions of all time')
+
+        utils.print_table(pd.DataFrame(df_top_short * 100, columns=['max']),
+                          fmt='{0:.2f}%',
+                          name='Top 10 short positions of all time')
+
+        utils.print_table(pd.DataFrame(df_top_abs * 100, columns=['max']),
+                          fmt='{0:.2f}%',
+                          name='Top 10 positions of all time')
 
         _, _, df_top_abs_all = pos.get_top_long_short_abs(
             positions_alloc, top=9999)
-        print('All positions ever held')
-        print(pd.DataFrame(df_top_abs_all).index.values)
-        print(np.round(pd.DataFrame(df_top_abs_all)[0].values, 3))
-        print("\n")
+        utils.print_table(pd.DataFrame(df_top_abs_all * 100, columns=['max']),
+                          fmt='{0:.2f}%',
+                          name='All positions ever held')
 
     if show_and_plot == 0 or show_and_plot == 2:
 
@@ -1392,11 +1389,9 @@ def show_worst_drawdown_periods(returns, top=5):
 
     """
 
-    print('\nWorst Drawdown Periods')
     drawdown_df = timeseries.gen_drawdown_table(returns, top=top)
-    drawdown_df['net drawdown in %'] = list(
-        map(utils.round_two_dec_places, drawdown_df['net drawdown in %']))
-    print(drawdown_df.sort('net drawdown in %', ascending=False))
+    utils.print_table(drawdown_df.sort('net drawdown in %', ascending=False),
+                      name='Worst Drawdown Periods', fmt='{0:.2f}')
 
 
 def plot_monthly_returns_timeseries(returns, ax=None, **kwargs):
@@ -1516,8 +1511,10 @@ def show_profit_attribution(round_trips):
     pct_profit_attribution = round_trips.groupby(
         'symbol')['pnl'].sum() / total_pnl
 
-    print('\nProfitability (PnL / PnL total) per name:')
-    print(pct_profit_attribution.sort(inplace=False, ascending=False))
+    utils.print_table(pct_profit_attribution.sort(inplace=False,
+                                                  ascending=False),
+                      name='Profitability (PnL / PnL total) per name',
+                      fmt='{0:.2f}%')
 
 
 def plot_prob_profit_trade(round_trips, ax=None):

pyfolio/pos.py

@@ -1,5 +1,5 @@
 #
-# Copyright 2015 Quantopian, Inc.
+# Copyright 2016 Quantopian, Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

pyfolio/round_trips.py

@@ -1,5 +1,5 @@
 #
-# Copyright 2015 Quantopian, Inc.
+# Copyright 2016 Quantopian, Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

pyfolio/tears.py

@@ -1,5 +1,5 @@
 #
-# Copyright 2015 Quantopian, Inc.
+# Copyright 2016 Quantopian, Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -628,9 +628,11 @@ def create_interesting_times_tear_sheet(
                       'interesting times.', UserWarning)
         return
 
-    print('\nStress Events')
-    print(np.round(pd.DataFrame(rets_interesting).describe().transpose().loc[
-          :, ['mean', 'min', 'max']], 3))
+    utils.print_table(pd.DataFrame(rets_interesting)
+                      .describe().transpose()
+                      .loc[:, ['mean', 'min', 'max']] * 100,
+                      name='Stress Events',
+                      fmt='{0:.2f}%')
 
     if benchmark_rets is None:
         benchmark_rets = utils.get_symbol_rets('SPY')

pyfolio/timeseries.py

@@ -1,5 +1,5 @@
 #
-# Copyright 2015 Quantopian, Inc.
+# Copyright 2016 Quantopian, Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

pyfolio/txn.py

@@ -1,5 +1,5 @@
 #
-# Copyright 2015 Quantopian, Inc.
+# Copyright 2016 Quantopian, Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

pyfolio/utils.py

@@ -1,5 +1,5 @@
 #
-# Copyright 2015 Quantopian, Inc.
+# Copyright 2016 Quantopian, Inc.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -470,3 +470,42 @@ def get_symbol_rets(symbol, start=None, end=None):
     return SETTINGS['returns_func'](symbol,
                                     start=start,
                                     end=end)
+
+
+def print_table(table, name=None, fmt=None):
+    """Pretty print a pandas DataFrame.
+
+    Uses HTML output if running inside Jupyter Notebook, otherwise
+    formatted text output.
+
+    Parameters
+    ----------
+    table : pandas.Series or pandas.DataFrame
+        Table to pretty-print.
+    name : str, optional
+        Table name to display in upper left corner.
+    fmt : str, optional
+        Formatter to use for displaying table elements.
+        E.g. '{0:.2f}%' for displaying 100 as '100.00%'.
+        Restores original setting after displaying.
+
+    """
+    if isinstance(table, pd.Series):
+        table = pd.DataFrame(table)
+
+    if fmt is not None:
+        prev_option = pd.get_option('display.float_format')
+        pd.set_option('display.float_format', lambda x: fmt.format(x))
+
+    if name is not None:
+        table.columns.name = name
+
+    try:
+        get_ipython()  # noqa
+        from IPython.display import display, HTML
+        display(HTML(table.to_html()))
+    except:
+        print(table)
+
+    if fmt is not None:
+        pd.set_option('display.float_format', prev_option)