From 71f779f1ba5cd10072b58ab1470d1ae109b5dfbb Mon Sep 17 00:00:00 2001 From: tsunglin Date: Sun, 31 Oct 2021 01:10:15 +0800 Subject: [PATCH] first commit --- main.py | 111 +++++++++++++++ plot.py | 419 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 530 insertions(+) create mode 100644 main.py create mode 100644 plot.py diff --git a/main.py b/main.py new file mode 100644 index 0000000..6e2067d --- /dev/null +++ b/main.py @@ -0,0 +1,111 @@ + +from plot import Plot + +class My_Strategy(Plot): + + def __init__(self,api): + super().__init__(api) + + def KDJMA(self, time_period = 10,oversell = 20,overbuy = 80, ma1 = 5,ma2 = 20,trading_type='standard'): + + #------------------------------------------------- + # calculate k, d , j + #------------------------------------------------- + data = self.kbars_df + ini_k = 50 + ini_d = 50 + k=[] + d=[] + rsv = (data['Close'].rolling(time_period).apply(lambda x:x[-1]) - data["Low"].rolling(time_period).min() ) / ( data["High"].rolling(time_period).max() - data["Low"].rolling(time_period).min() ) *100 + rsv = rsv.dropna() + + for _ in rsv: + ini_k = 2/3 * ini_k + 1/3 * _ + k.append(ini_k) + + for _ in k: + ini_d = 2/3 * ini_d + 1/3 * _ + d.append(ini_d) + k = pd.Series(k,index = rsv.index) + d = pd.Series(d,index = rsv.index) + j = 3 * k - 2 * d + ma1 = data['Close'].rolling(ma1).mean() + ma2 = data['Close'].rolling(ma2).mean() + #------------------------------------------------- + # make strategy signals + #------------------------------------------------- + + init_buy_sig = ( j>oversell ) & ( j.shift()ma2) + init_sell_sig = ( joverbuy ) & (ma1ma2 ) & ( ma1.shift()ma2 ) + indicators = [('main',{"ma1":ma1,"ma2":ma2})] # [('sub', {indicators1}), ('main'{indicators2})......] + + self.init_buy_sig = init_buy_sig + self.init_sell_sig = init_sell_sig + self.indicators = indicators + self.trading_type = trading_type + + def KDJ(self, time_period = 10,oversell = 20,overbuy = 80,trading_type='standard'): + + #------------------------------------------------- + # calculate k, d , j + #------------------------------------------------- + data = self.kbars_df + ini_k = 50 + ini_d = 50 + k=[] + d=[] + rsv = (data['Close'].rolling(time_period).apply(lambda x:x[-1]) - data["Low"].rolling(time_period).min() ) / ( data["High"].rolling(time_period).max() - data["Low"].rolling(time_period).min() ) *100 + rsv = rsv.dropna() + + for _ in rsv: + ini_k = 2/3 * ini_k + 1/3 * _ + k.append(ini_k) + + for _ in k: + ini_d = 2/3 * ini_d + 1/3 * _ + d.append(ini_d) + k = pd.Series(k,index = rsv.index) + d = pd.Series(d,index = rsv.index) + j = 3 * k - 2 * d + + #------------------------------------------------- + # make strategy signals + #------------------------------------------------- + + init_buy_sig = ( j>oversell ) & ( j.shift()overbuy ) + + # 模擬真實交易狀況，在訊號出現的下一個交易機會買入 + init_buy_sig = init_buy_sig.shift().dropna() + init_sell_sig = init_sell_sig.shift().dropna() + + indicators = [('sub',{"K":k,"D":d,"J":j})] # [('sub', {indicators1}), ('main'{indicators2})......] + + self.init_buy_sig = init_buy_sig + self.init_sell_sig = init_sell_sig + self.indicators = indicators + self.trading_type = trading_type + + diff --git a/plot.py b/plot.py new file mode 100644 index 0000000..90a59f8 --- /dev/null +++ b/plot.py @@ -0,0 +1,419 @@ +import numpy as np +import pandas as pd +import matplotlib.pyplot as plt +import plotly.graph_objects as go +import shioaji as sj +from tabulate import tabulate +import yfinance as yf +import warnings +warnings.simplefilter(action='ignore', category=FutureWarning) +warnings.filterwarnings("ignore") + +api = sj.Shioaji() +api.login("id_num", "password") +api.activate_ca( +ca_path="C:\\Users\\tsunglin\\Desktop\\SinoPac.pfx", +ca_passwd="id_num", +person_id="id_num") + +class Plot(): + def __init__(self,api=api): + self.api = api + + def input_stock(self,kbars_df,stock_id,freq='B'): + + if freq =='M': + label = 'left' + else: + label = 'right' + + data = { + "Open":kbars_df["Open"].resample(freq,closed='right',label=label).first(), + "High":kbars_df["High"].resample(freq,closed='right',label=label).max(), + "Low":kbars_df["Low"].resample(freq,closed='right',label=label).min(), + "Close":kbars_df["Close"].resample(freq,closed='right',label=label).last(), + "Volume":kbars_df["Volume"].resample(freq,closed='right',label=label).sum() + } + kbars_df = pd.DataFrame(data,index=data['Open'].index).dropna() + self.freq = freq + self.stock_id = stock_id + self.kbars_df = kbars_df + + def get_stock(self,stock_id,start,end,freq='B'): + ''' + freq = Day -> 'B', + Minutes -> '10T', + Week -> 'W', + Month -> 'M', set label = 'left' + ''' + + if freq =='M': + label = 'left' + else: + label = 'right' + + contract = self.api.Contracts.Stocks[stock_id] + kbars = self.api.kbars(contract, start=start, end=end) + kbars_df = pd.DataFrame({**kbars}) + kbars_df.ts = pd.to_datetime(kbars_df.ts) + kbars_df = kbars_df.set_index('ts') + data = { + "Open":kbars_df["Open"].resample(freq,closed='right',label=label).first(), + "High":kbars_df["High"].resample(freq,closed='right',label=label).max(), + "Low":kbars_df["Low"].resample(freq,closed='right',label=label).min(), + "Close":kbars_df["Close"].resample(freq,closed='right',label=label).last(), + "Volume":kbars_df["Volume"].resample(freq,closed='right',label=label).sum() + } + kbars_df = pd.DataFrame(data,index=data['Open'].index).dropna() + kbars_df.index.name="Date" + + self.freq = freq + self.stock_id = stock_id + self.kbars_df = kbars_df + + def __strategy_type(self): + data = self.kbars_df + trading_type = self.trading_type + init_buy_sig = self.init_buy_sig + init_sell_sig = self.init_sell_sig + + if trading_type == 'standard': + + buy_sig=init_buy_sig.copy() + buy_sig[init_buy_sig]=1 + buy_sig[init_sell_sig]=-1 + strategy_sig = buy_sig.replace(0,np.nan).ffill() + + strategy_buy = strategy_sig[strategy_sig.diff()==2] + strategy_sell = strategy_sig[strategy_sig.diff()==-2] + + strategy_buy_mark = data[data.index.isin(strategy_buy.index)] + strategy_buy_mark['Signal'] = 1 + strategy_sell_mark = data[data.index.isin(strategy_sell.index)] + strategy_sell_mark['Signal'] = 0 + + if trading_type =='customize': + pass + + return strategy_sell_mark, strategy_buy_mark + + def run(self,enter_comm=0.001425*0.28,exit_comm=0.001425*0.28+0.003,buy_at_kbar="Close", sell_at_kbar="Close"): + + sell_mark, buy_mark = self.__strategy_type() + self.buy_mark = buy_mark + self.sell_mark = sell_mark + self.buy_at_kbar = buy_at_kbar + self.sell_at_kbar = sell_at_kbar + + df = pd.concat([buy_mark, sell_mark],join='outer',sort=False).sort_index() + df['Holding periods'] = df.index.to_series().diff().shift(-1) + df['Profit'] = (df[sell_at_kbar].shift(-1)*(1-exit_comm) - df[buy_at_kbar]*(1+enter_comm)) / (df[buy_at_kbar]*(1+enter_comm)) + df['Absolute value'] = df['Profit'].abs() + df['Gain / Loss'] = df['Profit']>0 + df['Gain / Loss'] = df['Gain / Loss'].map({True:'$Gain$',False:'Loss'}) + + profit = df[df['Signal'] == 1][['Profit','Holding periods','Absolute value','Gain / Loss']] + profit = profit.dropna() + profit['Cumulative ret'] = (profit['Profit']+1).cumprod() + profit['Maximum draw drown'] = ((profit['Cumulative ret'] - profit['Cumulative ret'].cummax())/profit['Cumulative ret'].cummax()) + + + average = (profit['Profit'].mean()) * 100 + std = profit['Profit'].std() + win_rate = ((profit['Profit']>0).sum() / len(profit))*100 + average_period = profit['Holding periods'].mean() + num_trade = len(profit) + data = {"Average return(%)":average,"Strategy std":std,"Number of trades":num_trade,"Winning rate(%)":win_rate,'Average holding periods':average_period} + strategy_info = pd.DataFrame(data,index=['Strategy info']) + + profit[['Profit','Absolute value','Cumulative ret','Maximum draw drown']] = (profit[['Profit','Absolute value','Cumulative ret','Maximum draw drown']]).apply(lambda x:round(x,4)) + strategy_info[['Average return(%)','Strategy std','Winning rate(%)']] = (strategy_info[['Average return(%)','Strategy std','Winning rate(%)']]).apply(lambda x:round(x,4)) + print(tabulate(profit,headers=profit.columns, tablefmt='pretty')) + print(tabulate(strategy_info,headers=strategy_info.columns, tablefmt='pretty')) + + self.trading_detail = df + self.profit_detail = profit + self.strategy_info = strategy_info + + def plot(self): + + df = self.kbars_df + freq = self.freq + stock_id = self.stock_id + indicator_set = self.indicators + buy_mark = self.buy_mark + sell_mark = self.sell_mark + profit_detail = self.profit_detail + buy_at_kbar = self.buy_at_kbar + sell_at_kbar = self.buy_at_kbar + #---------------------------------------------------------------------------------- + # create plot + #---------------------------------------------------------------------------------- + # candle stick, buy and sell signals (y6) + #---------------------------------------------------------------------------------- + fig = go.Figure() + + fig.add_trace( + go.Candlestick(x=df.index, + open=df['Open'], + high=df['High'], + low=df['Low'], + close=df ['Close'], + name="Stock Prices", + increasing={'line': {'color': '#DC3912'}}, + decreasing={'line': {'color': '#222A2A'}}, + yaxis='y6' )) + + fig.add_trace( + go.Scatter( + x=buy_mark.index, + y=buy_mark[buy_at_kbar], + name="Buy Signal", + marker=dict(color="#0DF9FF", size=15), + mode="markers", + marker_symbol="triangle-up", + yaxis='y6')) + + fig.add_trace( + go.Scatter( + x=sell_mark.index, + y=sell_mark[sell_at_kbar], + name="Sell Signal", + marker=dict(color="#778AAE", size=15), + mode="markers", + marker_symbol="triangle-down", + yaxis='y6')) + #---------------------------------------------------------------------------------- + # technical indicator (y5 or y6) + #---------------------------------------------------------------------------------- + + if indicator_set: + for indicators in indicator_set: + + if indicators[0]=="main": + + for indicator_name, indicator in indicators[1].items(): + fig.add_trace( + go.Scatter( + x=indicator.index, + y=indicator, + name=indicator_name, + yaxis='y6')) + + if indicators[0]=="sub": + + for indicator_name, indicator in indicators[1].items(): + fig.add_trace( + go.Scatter( + x=indicator.index, + y=indicator, + name=indicator_name, + yaxis='y5')) + + #---------------------------------------------------------------------------------- + # volume (y4) + #---------------------------------------------------------------------------------- + fig.add_trace( + go.Bar( + x=df.index, + y=df['Volume'], + yaxis='y4', + name='Volume', + marker_color='#795548')) + + #---------------------------------------------------------------------------------- + # individual profit plot (y3) + #---------------------------------------------------------------------------------- + + + + fig.add_trace( + go.Scatter( + x=profit_detail.index, + y=profit_detail.Profit*100, + marker_size = (profit_detail["Absolute value"])*100, + marker_color = profit_detail["Gain / Loss"].map({'$Gain$':'#D62728',"Loss":'#2CA02C'}), + + marker_symbol='circle', + mode='markers', + yaxis='y3', + name = 'Gain/Loss for each trade(%)')) + + + + #---------------------------------------------------------------------------------- + # cumulative profit plot (y2) + #---------------------------------------------------------------------------------- + fig.add_trace( + go.Scatter( + x=profit_detail.index, + y=profit_detail['Cumulative ret']*100, + name='Cumulative return(%)', + line_color='#3366CC', + yaxis='y2')) + #---------------------------------------------------------------------------------- + # maximum drawdown (y) + #---------------------------------------------------------------------------------- + + fig.add_trace( + go.Scatter( + x=profit_detail.index, + y=profit_detail['Maximum draw drown']*100, + name='Maximum draw down(%)', + line_color='#DC3912', + yaxis='y')) + + + #---------------------------------------------------------------------------------- + # setup layout + #---------------------------------------------------------------------------------- + # y axis + #---------------------------------------------------------------------------------- + fig.update_layout( + xaxis=dict( + autorange=True, + rangeslider=dict( + autorange=True, + ), + type="date" + ), + yaxis=dict( + anchor="x", + autorange=True, + domain=[0, 0.1], + linecolor="#DC3912", + mirror=True, + showline=True, + side="right", + tickfont={"color": "#DC3912"}, + tickmode="auto", + ticks="", + titlefont={"color": "#DC3912"}, + type="linear", + zeroline=False, + ), + + yaxis2=dict( + anchor="x", + autorange=True, + domain=[0.11, 0.2], + linecolor="#3366CC", + mirror=True, + showline=True, + side="right", + tickfont={"color": "#3366CC"}, + tickmode="auto", + ticks="", + titlefont={"color": "#3366CC"}, + type="linear", + zeroline=False, + ), + + yaxis3=dict( + anchor="x", + autorange=True, + domain=[0.21, 0.35], + linecolor="#222A2A", + mirror=True, + showline=True, + side="right", + tickfont={"color": "#222A2A"}, + tickmode="auto", + ticks="", + titlefont={"color": "#222A2A"}, + type="linear", + zeroline=True, + ), + + yaxis4=dict( + anchor="x", + autorange=True, + domain=[0.36, 0.45], + linecolor="#795548", + mirror=True, + showline=True, + side="right", + tickfont={"color": "#795548"}, + tickmode="auto", + ticks="", + titlefont={"color": "#795548"}, + type="linear", + zeroline=False, + + ), + + + yaxis5=dict( + anchor="x", + autorange=True, + domain=[0.46, 0.55], + linecolor="#673ab7", + mirror=True, + showline=True, + side="right", + tickfont={"color": "#673ab7"}, + tickmode="auto", + ticks="", + titlefont={"color": "#673ab7"}, + type="linear", + zeroline=False, + + ), + + + + yaxis6=dict( + anchor="x", + autorange=True, + domain=[0.56, 1], + linecolor="#E91E63", + mirror=True, + showline=True, + side="right", + tickfont={"color": "#E91E63"}, + tickmode="auto", + ticks="", + titlefont={"color": "#E91E63"}, + type="linear", + zeroline=False, + + )) + + #---------------------------------------------------------------------------------- + # final settings + #---------------------------------------------------------------------------------- + + fig.update_traces(decreasing_line_width=2,increasing_line_width=2,whiskerwidth=0.5, selector=dict(type='candlestick')) + + if freq=='B': + fig.update_xaxes(rangebreaks=[ + dict(bounds=['sat', 'mon'])]) + + if freq[-1]=='T': + fig.update_xaxes(rangebreaks=[dict(bounds=[14, 9], pattern="hour"), + dict(bounds=['sat', 'mon'])]) + + + + fig.update_layout( + title = f'STOCK ID : {stock_id}, Data frequency:{freq}', + dragmode="zoom", + hovermode="x", + legend=dict(orientation="h", + yanchor="top", + xanchor="center", + y=-0.3, + x=0.5, + traceorder="reversed"), + height=1000, + width=1000, + template="plotly_white", + margin=dict( + l = 70, + r = 70, + t=100, + b=50 + ), + ) + + fig.show() \ No newline at end of file