examples
=====================================================

trailing stop loss
"""""""""""""""""""

.. code-block:: python
   :linenos:

   for index, frame in tqdm(df.groupby('date'), 'Backtesting'):
      days_count = frame.index[0]
        frame.set_index('ticker', inplace=True)

      # Sell if stop loss is hit
      for ticker, info in portfolio.holding():
         if frame.loc[ticker, 'low'] <= info['cut_loss']:
             portfolio.sell(ticker, Reason.EXIT, index, info['cut_loss'], days_count)

      # Trailing stop loss
      for ticker, info in portfolio.holding():
         stop_percentage = 0.25
         trailing_stop_price = frame.loc[ticker, 'high'] * (1 - stop_percentage)  # Trail stop 25% below current high
         if trailing_stop_price > info['cut_loss']:
             portfolio.set_cut_loss_price(ticker, trailing_stop_price)

take profit
"""""""""""""

.. code-block:: python
   :linenos:

   # take profit can be set when initializing portfolio
   portfolio = Portfolio(df, EquityModel.CORE_EQUITY, PositionSizing.EQUAL_UNIT, initial_capital=100000, size=10, take_profit=0.34, cut_loss=0.1)

   for index, frame in tqdm(df.groupby('date'), "Backtesting"):
      days_count = frame.index[0]
      frame.set_index('ticker', inplace=True)

      for ticker, info in portfolio.holding():
         if frame.loc[ticker, 'low'] <= info['cut_loss']:
             portfolio.sell(ticker, Reason.CL, index, info['cut_loss'], days_count)

         elif frame.loc[ticker, 'sell_signal'] == 1:
             portfolio.sell(ticker, Reason.EXIT, index, frame.loc[ticker, 'open'], days_count)

         elif info['take_profit'] is not None: # examples for take profit
             if frame.loc[ticker, 'high'] >= info['take_profit']:
                 portfolio.sell(ticker, Reason.TP, index, info['take_profit'], days_count)

   # take profit can be overrode by calling portfolio.buy(..., tp_price= ?)

time stop
""""""""""

.. code-block:: python
   :linenos:

   df['sell_signal'] = np.where((df['buy_signal'].shift(10) > 0) & (df['ticker'] == df['ticker'].shift(10)),
                             df['date'].shift(10).dt.date, pd.to_datetime('2000-01-01'))

   for index, frame in tqdm(df.groupby('date'), "Backtesting"):  # loop dataframe by date
       days_count = frame.index[0]  # date
       frame.set_index('ticker', inplace=True)  # set ticker as index

       # Check portfolio and sell if stop loss is hit
       for ticker, info in portfolio.holding():
           if frame.loc[ticker, 'sell_signal'] == info['open_date']:  # 10 days sell signal
               portfolio.sell(ticker, Reason.EXIT, index, frame.loc[ticker, 'open'], days_count)

consider commissions and set brokerage fee in portfolio
""""""""""""""""""""""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   portfolio = Portfolio(df, EquityModel.CORE_EQUITY, PositionSizing.EQUAL_UNIT, initial_capital=100000, size=10,
                         take_profit=None, cut_loss=0.1, commission=True, brokerage_fee=0.08)

sell based on yesterday's condition
""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   yesterday_frame = pd.DataFrame()

   for index, frame in tqdm(df.groupby('date'), "Backtesting"):
       days_count = frame.index[0]
       frame.set_index('ticker', inplace=True)

       for ticker, info in portfolio.holding():
           # Sell at market open if previous bar close below SL
           if yesterday_frame.loc[ticker, 'close'] <= info['cut_loss']:
               portfolio.sell(ticker, Reason.EXIT, index, frame.loc[ticker, 'open'], days_count)

       yesterday_frame = frame.copy()  # Update yesterday frame

view all performances of strategy
""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   performance = Performance(portfolio, show=True)
   performance.objective_function()
   performance.performance_metric()
   performance.transactions_to_csv()
   performance.objective_to_csv('strategy 1', 'objective_function.csv')
   performance.performance_to_csv('performance 1', 'performance_metric.csv')
   performance.generate_all_files(strategy_name='strategy_name', code_path='file_name.py', version='v1',
                                  plot_image=True)

dataframe list for multi period backtesting
"""""""""""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   date1 = pd.to_datetime('2019-01-01')
   date2 = pd.to_datetime('2020-01-01')
   date3 = pd.to_datetime('2021-01-01')
   df_list = universe.df_multi_period(date1, date2, date3)

   for df in df_list:
   # do backtesting

stocks to buy tomorrow at market open
""""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   # Init df_today
   today = datetime.today().strftime('%Y-%m-%d')
   today = pd.to_datetime(today)
   mask = df.index.date == today
   df_today = df[mask]
   df_today.set_index('ticker', inplace=True)

   # Stocks to buy tomorrow
   buy = []
   df_today['buy_signal_today'] = np.where((df_today['turnover'] >= 1000000), df['turnover'].values, 0)
   df_today = df_today.loc[
                  (df_today['buy_signal_today'].values != 0) & (~df_today.index.isin(portfolio.list()))].sort_values(
       'buy_signal_today', ascending=False)[:portfolio.free()]
   for item in df_today.index.to_list():
       portfolio.buy(item, today, df_today.loc[item, 'close'], today)
       buy.append(item)
   print(f'Buy tomorrow at market open: {buy}')


backtest on different parameters
""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   def backtest(param1, param2):
       # do backtesting

       # record performance into csv
       performance = Performance(portfolio, show=False)
       performance.objective_function(need_print=False)
       performance.performance_metric(need_print=False)
       performance.objective_to_csv(f'{param1}_{param2}', 'objective_function.csv')
       performance.performance_to_csv(f'{param1}_{param2}', 'performance_metric.csv')


   if __name__ == '__main__':
       # Init indicator and signal

       # Backtest
       for i in param1_list:  # row
           for j in param2_list:  # col
               backtest(i, j)

plot heatmap for specific metric
""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   universe.plot_heatmap('objective_function.csv', 'Maximum Daily Drawdown %', param_row.tolist(),
                         param_col.tolist(), 'Param col', 'Param row', 15, 10, True,
                         'Drawdown Heatmap.png')

generate tick data
""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   from optimus.backtesting.tick_data import TickData
   tick_data = TickData()
   df = tick_data.df
   print(df)

resample a timeframe from tick data
""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   from optimus.backtesting.tick_data import TickData
   tick_data.resample('15min')

get dataframe of a timeframe
""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   from optimus.backtesting.tick_data import TickData
   timeframe_df = tick_data.get_timeframe('15min', '2021-07-12')
   print(timeframe_df)

get technical analysis of a timeframe
""""""""""""""""""""""""""""""""""""""""

.. code-block:: python
   :linenos:

   from optimus.backtesting.tick_data import TickData
   indicator = Indicator()
   indicator.sma([50])
   indicator.parabolic_sar()
   ta_df = tick_data.get_TA('5min', indicator)
   print(ta_df)