# 一、短期均營收(X月) > 長期均營收(Y月)*(1+Z%)
# 二、本月營收/上月營收 > A 
# 三、股池為上市，排除KY
# 四、選擇滿足條件下：peg較小的前B個，peg為 本益比/營業利益成長率
# 五、在每個月的最後營收公告截止日換股。
# 六、手續費為C折 ，各個股票張數金額不超過D萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=3
y=12
z=10
a=0.9
b=10
c=3
d=20
BM='0050'

adjclose =msmp.日_K.adj_Close
adjopen =msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open   = msmp.日_K.Open
rev=  msmp.月_營收.金額_千
peg =msmp.日_指標.本益比 / msmp.季財報_財務指標.營業利益成長率
list_STC = msmp.list['tse']
list_KY = msmp.list['KY']

#股數
qty = np.floor(d*10000/(Open*1000))*1000
con1= rev.rolling(x).mean()/rev.rolling(y).mean()>(1+0.01*z)
con2= rev/rev.shift()> a
# 隔天沒有鎖漲停
con3 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
position= con1 &con2 &con3 &(qty!=0)

position=position[position.columns[position.columns.isin(list_STC)]]
position=position[position.columns[~position.columns.isin(list_KY)]]
#換股
position=position*peg
position=position.apply(lambda row: row[row>0].nsmallest(b), axis=1).notna().ne(0)
#周期
setday=rev.index_str_to_date().index
position=position[position.index.isin(setday)]
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'],qty , position)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        fee = c*0.0001425,
        benchmark=benchmark,
        )

										
									

										
												
# 策略說明：基本面及技術面多因子選股策略，以下是策略的主要條件：
# 一、近一季月營收均值 > 近一年月營收均值*(1+X%)
# 二、這月營收 > 上月營收*(1+X%)
# 三、融資使用率 < Y%
# 四、entry_volatility < Z% ，entry_volatility = atr/adjclose
# 五、業外營收/營業收入 占比 < A%
# 六、選擇滿足條件下：peg較大的B個，peg =pe/營業利益成長率 。
# 七、在每個月的最後營收公告截止日換股。
# 八、手續費為C折 ，各個股票張數金額不超過D萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=10
y=80
z=10
a=50
b=10
c=3
d=20
BM='0050'

adjclose= msmp.日_K.adj_Close
adjopen= msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open = msmp.日_K.Open
list_STC = msmp.list['tse']
rev = msmp.月_營收.金額_千
融資使用率 =msmp.日_融資券.融資使用率

def calculate(window=14 ,high=None, low=None, close=None):
    high = msmp.日_K.adj_High if high is None else high
    low = msmp.日_K.adj_Low if low is None else low
    close = msmp.日_K.adj_Close if close is None else close    
    plus_dm = high.diff()
    minus_dm = low.diff()
    plus_dm = plus_dm.where((plus_dm > minus_dm) & (plus_dm > 0), 0)
    minus_dm = minus_dm.where((minus_dm > plus_dm) & (minus_dm > 0), 0)
    tr = np.maximum.reduce([high - low, abs(high - close.shift()), abs(low - close.shift())])
    tr = pd.DataFrame(tr,columns=adjclose.columns ,index = adjclose.index)
    atr = tr.rolling(window=window).mean()
    plus_di = 100 * (plus_dm.rolling(window=window).mean() / atr)
    minus_di = 100 * (minus_dm.rolling(window=window).mean() / atr)
    dx = 100 * abs(plus_di - minus_di) / (plus_di + minus_di)
    adx = dx.rolling(window=window).mean()
    plus_di.name = "DMI"  
    minus_di.name = "Minus DI"
    return atr ,adx ,plus_di ,minus_di

atr= calculate(10)[0]
entry_volatility = atr/adjclose
業外收支營收率= msmp.季財報_財務指標.業外收支營收率
pe= msmp.日_指標.本益比
營業利益成長率 = msmp.季財報_財務指標.營業利益成長率
peg =pe/營業利益成長率 
setday=rev.index_str_to_date().index

#股數
qty = np.floor(d*10000/(Open*1000))*1000
con1 = rev.rolling(3).mean()/rev.rolling(12).mean()> (1+0.01*x)
con2 = rev/rev.shift(1) > (1+0.01*x)
con3 = 融資使用率<= y
con4 = entry_volatility <= 0.01*z
con5 = 業外收支營收率< a
# 隔天沒有鎖漲停
con6 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
buy=con1&con2&con3&con4&con5&con6&(qty!=0)

buy=buy[buy.columns[buy.columns.isin(list_STC)]]
buy=(buy*peg).fillna(0).apply(lambda row: row[row > 0].nsmallest(b), axis=1).notna().ne(0)
buy=buy[buy.index.isin(setday)]
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'] ,qty , buy)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        fee = c*0.0001425,
        benchmark=benchmark,
        )

										
									

										
												
# 策略說明：基本面及技術面多因子選股策略，以下是策略的主要條件：
# 一、今日股價創X日新高
# 二、近Y日發生股價創Y日新高
# 三、近Z日發生股價創Z日新高
# 四、近3月平均營收大於A月平均營收
# 五、近5日平均成交量大於近20日平均成交量
# 六、股池為上市 且 剔除KY
# 七、選擇滿足條件下：成交量較大的B個。
# 八、在每個月的最後營收公告截止日換股。
# 九、手續費為C折 ，各個股票張數金額不超過D萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=20
y=60
z=120
a=12
b=10
c=3
d=20

BM='0050'
adjclose= msmp.日_K.adj_Close
adjopen= msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open = msmp.日_K.Open
list_STC = msmp.list['tse']
list_KY  = msmp.list['KY']
rev = msmp.月_營收.金額_千
vol = msmp.日_K.Volume

#股數
qty = np.floor(d*10000/(Open*1000))*1000
# 近一日收盤價創近20日新高
con1 = adjclose == adjclose.rolling(x).max()
# 近一季度創最高點的股票
con2 = (adjclose==adjclose.rolling(y).max()).rolling(y).sum()>0
# 近半創最高點的股票
con3 = (adjclose==adjclose.rolling(z).max()).rolling(z).sum()>0
#近3月平均營收大於12月平均營收
con4 = rev.rolling(3).mean() > rev.rolling(a).mean()
# 近5日平均成交量大於近20日平均成交量
con5 = vol.rolling(5).mean() > vol.rolling(20).mean()
# 隔天沒有鎖漲停
con6 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
buy=con1&con2&con3&con4&con5&con6& (qty!=0)

#標的範圍
buy=buy[buy.columns[buy.columns.isin(list_STC)]]
buy=buy[buy.columns[~buy.columns.isin(list_KY)]]
#取10大
buy=buy*vol
buy=buy.apply(lambda row: row[row!=0].nlargest(b), axis=1).notna().ne(0)
# 回測周期
setday=rev.index_str_to_date().index
buy=buy[buy.index.isin(setday)]
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'],qty , buy)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        fee = c*0.0001425,
        benchmark=benchmark,
        )

										
									

										
												
# 一、近5日發生，還原收盤價創X日新高
# 二、還原收盤價低於中位數
# 三、收盤價 < Y
# 四、交易量的5日均 > Z
# 五、股池為上市、排除KY、排除金融股
# 六、選擇滿足條件下：收盤價較小的A個 
# 七、再每月第一個交易日換股
# 八、手續費為B折 ，各個股票張數金額不超過C萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=120
y=100
z=500
a=10
b=3
c=20

BM='0050'
adjclose = msmp.日_K.adj_Close
adjopen = msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open   =  msmp.日_K.Open
Close =  msmp.日_K.Close
vol = msmp.日_K.adj_Volume
list_STC = msmp.list['tse']
list_KY = msmp.list['KY']
listFinance = msmp.INFO[msmp.INFO.產業名稱.str.contains('金融', na=False)].股票代號.tolist()

#股數
qty = np.floor(c*10000/(Open*1000))*1000
con1= (adjclose == adjclose.rolling(x).max()).rolling(5).sum()>=1
con2= adjclose.apply(lambda row : row<row.quantile(0.5) ,axis=0)
con3 =Close <= y
con4= vol.rolling(5).mean()> z
# 隔天沒有鎖漲停
con5 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
buy=con1&con2&con3&con4&con5&(qty!=0)

buy=buy[buy.columns[buy.columns.isin(list_STC)]]    
buy=buy[buy.columns[~buy.columns.isin(list_KY)]]   
buy=buy[buy.columns[~buy.columns.isin(listFinance)]]   
#數量
buy=buy*Close
buy=buy.apply(lambda row: row[row > 0].nsmallest(a), axis=1).notna().ne(0)
setday=buy.groupby([buy.index.year, buy.index.month]).apply(lambda x: x.index[0]).tolist()
buy=buy[buy.index.isin(setday)]
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'],qty , buy)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        benchmark=benchmark,
        fee = b*0.0001425,
        )


										
									

										
												# 策略說明：基本面多因子選股策略，以下是策略的主要條件：
# 一、市值 < X億
# 二、自由現金流 > 0
# 三、股東權益報酬率 > 0
# 四、營業利益成長率 > 0
# 五、交易量 > Z張
# 六、選擇滿足條件下：以A為周期的rsv，選擇rsv較大的B個。
# 七、在每個月的最後營收公告截止日換股。
# 八、手續費為C折 ，各個股票張數金額不超過D萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=50
z=300
a=60
b=10
c=3
d=20
BM='0050'

adjclose= msmp.日_K.adj_Close
adjopen= msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open = msmp.日_K.Open
list_STC = msmp.list['tse']
list_KY  = msmp.list['KY']
自由現金流 = msmp.季財報_現金流表_千.自由現金流
股東權益報酬率 = msmp.季財報_財務指標.ROE_稅後權益報酬率
營業利益成長率 = msmp.季財報_財務指標.營業利益成長率

#股數
qty = np.floor(d*10000/(Open*1000))*1000
#市值<p
市值 = msmp.日_K.市值_百萬/100  
con1 = 市值 < x
#自由現金流>0
con2 = 自由現金流 > 0
#股東權益報酬率 > 0
con3 = 股東權益報酬率 > 0
#營業利益成長率 > 0
con4 = 營業利益成長率 > 0
#交易量 > p
vol = msmp.日_K.adj_Volume
con5 =vol > z
# 隔天沒有鎖漲停
con6 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
buy=con1&con2&con3&con4&con5&con6&(qty!=0)

buy=buy[buy.columns[buy.columns.isin(list_STC)]]
buy=buy[buy.columns[~buy.columns.isin(list_KY)]]
rsv = (adjclose  - adjclose .rolling(a).min()) / (adjclose .rolling(a).max() - adjclose .rolling(a).min())
buy=buy*rsv
buy=buy.apply(lambda row: row[row > 0].nlargest(b), axis=1).notna().ne(0)
rev = msmp.月_營收.金額_千
setday=rev.index_str_to_date().index
buy=buy[buy.index.isin(setday)]
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'],qty , buy)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        fee = c*0.0001425,
        benchmark=benchmark,
        )

										
									

										
												
# 策略說明：基本面數據計分，以下是策略的主要條件：
# 一、ROE 排名劃分0-X
# 二、營業毛利率 排名劃分0-Y
# 三、稅前淨利成長率 排名劃分0-Z
# 四、應收帳款週轉率 排名劃分0-A
# 五、負債比率 排名劃分0-B
# 六、將分數加總 ，選出>20
# 七、股池為上市、排除KY
# 八、選擇滿足條件下：還原收盤價較小的10個 
# 九、再每周第一個交易日換股
# 十、手續費為C折 ，各個股票張數金額不超過D萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=10
y=10
z=10
a=10
b=10
c=3
d=20

BM='0050'
adjclose = msmp.日_K.adj_Close
adjopen =msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open   = msmp.日_K.Open
list_STC = msmp.list['tse']
list_KY = msmp.list['KY']
營業收入= msmp.季財報_損益表_千.營業收入
總應收帳款= msmp.季財報_資產負債表_千.總應收帳款
應收帳款週轉率 = 營業收入/總應收帳款.rolling(2).mean()

#股數
qty = np.floor(d*10000/(Open*1000))*1000
roe_score= msmp.季財報_財務指標.ROE_稅後權益報酬率.rank(axis=1, pct=True, ascending=True).mul(x).apply(np.ceil)
營業毛利率_score =  msmp.季財報_財務指標.營業毛利率.rank(axis=1, pct=True, ascending=True).mul(y).apply(np.ceil)
稅前淨利年增率_score =  msmp.季財報_財務指標.稅前淨利成長率.rank(axis=1, pct=True, ascending=True).mul(z).apply(np.ceil)
負債比率_score=  msmp.季財報_財務指標.負債比率.rank(axis=1, pct=True, ascending=False).mul(a).apply(np.ceil)
應收帳款週轉率_score =應收帳款週轉率.rank(axis=1, pct=True, ascending=True).mul(b).apply(np.ceil)
all_score = roe_score.fillna(0) + 營業毛利率_score.fillna(0)+ 稅前淨利年增率_score.fillna(0) + 應收帳款週轉率_score.fillna(0) + 負債比率_score.fillna(0)
buy=all_score>=20
# 隔天沒有鎖漲停
sig = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
buy= buy &sig &(qty!=0)

buy=buy[buy.columns[buy.columns.isin(list_STC)]]
buy=buy[buy.columns[~buy.columns.isin(list_KY)]]
buy=buy.index_str_to_date().reindex(adjclose.index).fillna(method='ffill')
buy=buy*adjclose
buy=buy.astype(float).apply(lambda row: row[row > 0].nsmallest(10), axis=1).notna().ne(0)

setday=buy.groupby([buy.index.year, buy.index.isocalendar().week]).apply(lambda x: x.index[0]).tolist()
buy=buy[buy.index.isin(setday)]
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'],qty , buy)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        benchmark=benchmark,
        fee = c*0.0001425,
        )

										
									

										
												
# 策略說明：基本面及技術面多因子選股策略，以下是策略的主要條件：
# 一、近5日内，有發生過還原收盤價創X日新高
# 二、近Y日内高低點振幅在30%以上
# 三、收盤價低於quantile的80%*1.9
# 四、還原收盤價< Z元
# 五、5日均交易量 > A張
# 六、波動率 < B
# 七、營收短期動能大於長期動能
# 八、股池為非上市、非KY、非金融股  
# 九、選擇滿足條件下：價格較低的10檔
# 十、在每個月的最後營收公告截止日換股
# 十一、手續費為C折 ，各個股票張數金額不超過D萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=102
y=75
z=100
a=500
b=12
c=3
d=20
BM='0050'

Close= msmp.日_K.Close
adjclose= msmp.日_K.adj_Close
adjopen= msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
vol=msmp.日_K.Volume
Wadjclose = msmp.週_K.adj_Close

Open   = msmp.日_K.Open
list_STC = msmp.list['tse']
list_KY = msmp.list['KY']
info= msmp.INFO
list_金融= info[info.產業名稱.str.contains('金融', na=False)].股票代號
rev= msmp.月_營收.金額_千

bullish_candle = adjclose.diff() >=0
#如果漲
bullish_volatility = abs(adjclose.shift() - adjopen) + abs(adjopen- adjlow) + abs(adjlow - adjhigh) + abs(adjhigh - adjclose)
#如果跌
bearish_volatility = abs(adjclose.shift() - adjopen) + abs(adjopen - adjhigh) + abs(adjhigh -adjlow) + abs(adjlow - adjclose)
#創表格
candle_volatility = adjclose[adjclose<0]
candle_volatility[bullish_candle] = bullish_volatility
candle_volatility[~bullish_candle] = bearish_volatility

if Wadjclose.index[-1].dayofweek==4:
    setday = Wadjclose.index
else:
    setday = Wadjclose.index[:-1]

#股數
qty = np.floor(d*10000/(Open*1000))*1000
# 收盤價近5日至少有1日創收盤價近100日創新
con1 = (adjclose == adjclose.rolling(x).max()).rolling(5).sum()>=1
# 近60日股價高低區間在30%內
con2 = adjlow.rolling(y).min()*1.3 < adjhigh.rolling(y).max()
# 收盤價低於整體市場分級的90%
標準價= Wadjclose.rolling(1000,min_periods=1).quantile(0.8)
con3= adjclose <標準價*1.9
# 收盤價低於100元
con4 = adjclose <= z
# 5日均大於500張
con5 = vol.rolling(5).mean() > a
# k線波動率濾網
volatility = candle_volatility.rolling(20).mean() / adjclose.rolling(20).mean() * 100
con6 = volatility <= b
# 營收短期動能大於長期動能
con7 = rev.rolling(2).mean() > rev.rolling(12).mean()
# 隔天沒有鎖漲停
con8 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
buy = con1&con2&con3&con4&con5&con6&con7&con8&(qty!=0)


#清單
buy=buy[buy.columns[~buy.columns.isin(list_STC)]]
buy=buy[buy.columns[~buy.columns.isin(list_KY)]]
buy=buy[buy.columns[~buy.columns.isin(list_金融)]]
#換股
buy=buy*adjclose
buy=buy.apply(lambda row: row[row > 0].nsmallest(10), axis=1).notna().ne(0)
#轉換回測格式
buy=buy[buy.index.isin(setday)]
dfX = msmp_to_dic(['qty','sig'],qty , buy)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        benchmark=benchmark,
        fee = c*0.0001425,
        )

										
									

										
												
# 一、連續Y年，填權息天數 <=X天
# 二、現金殖利率 > Z
# 三、股池為上市 但剔除ETF，換股日期為最後財報公告日
# 四、滿足以上條件，選擇現金殖利率較小的A檔股票
# 五、手續費為 B 折 ,股票最大開倉金額不超過 C 萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=30
y=3
z=5.5
a=10
b=3
c=20
BM='0050'

填權息天數 = msmp.季股利資訊.填權息天數.index_str_to_date()[:-1]  #年資料
adjclose = msmp.日_K.adj_Close
adjhigh =msmp.日_K.adj_High
adjlow  = msmp.日_K.adj_Low
現金股利 = msmp.季股利資訊.現金股利_季.rolling(4,min_periods=1).sum()
現金殖利率 = 100*現金股利.fillna(0)/adjclose
現金殖利率  = 現金殖利率.reindex(adjclose.index, method='ffill')
Open= msmp.日_K.Open
info =msmp.INFO
pool =info[(~info.股票代號.str.startswith('00')) &(info.上市上櫃.isin([1]))].股票代號.tolist()
index = msmp.INDEX['Q']

con1 = ((填權息天數 <=x)&(填權息天數 >0)).rolling(y).sum()==y
con1 = con1.reindex(adjclose.index, method='ffill')
con2 = 現金殖利率 >z
con3 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
qty = np.floor(c * 10000 / (Open * 1000)) * 1000
long =con1 &con2 &con3 &(qty!=0) 

long = long*現金殖利率
long = long.loc[long.index.isin(index) ,long.columns.isin(pool) ]
pos = long.apply(lambda row : row[row>0].nsmallest(a) ,axis=1).notna().ne(0)
dic = msmp_to_dic(['qty', 'long' ], qty, pos )

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp = api.backtest(dic ,
        symbol='id',date='date',qty='qty', sig_value='adjclose',
        open_value='adjopen', open_shift=True,
        close_value='adjopen', close_shift=True,
        longsig='long',
        benchmark=benchmark,    fee=0.0001425 * b)


















										
									

										
												
# 策略說明：透過均線指標和量能分析，找尋有機會成長的股票。以下是策略的主要條件：
# 一、近一日漲幅 > X%。
# 二、收盤價突破Y日均綫。
# 三、成交量 > Z日均量 *A倍 。
# 四、收盤價 < B元。
# 五、股池為上市 且 剔除KY
# 六、止盈為45%止損為15%，以發生信號的隔天開盤價，開倉及平倉。
# 七、手續費為C折 ，各個股票張數金額不超過D萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=5
y=60
z=5
a=4
b=20
c=3
d=20
BM='0050'

Close=msmp.日_K.Close
adjclose= msmp.日_K.adj_Close
adjopen= msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open = msmp.日_K.Open
list_STC = msmp.list['tse']
list_KY  = msmp.list['KY']

#股數
qty = np.floor(d*10000/(Open*1000))*1000
#近一日漲幅 > p%         
con1= adjclose>adjclose.shift(1)*(1+0.01*x)
#收盤價突破p日均綫
p = adjclose>adjclose.rolling(y).mean()
con2= p & (p.shift(1)==False)
#交易量 > period均 *n
Volume = msmp.日_K.adj_Volume    
con3= Volume>Volume.rolling(z).mean().shift(1)*(a)
#加入低價條件
con4= Close <b
# 隔天沒有鎖漲停
con5 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
position=con1&con2&con3&con4&con5&(qty!=0)

#清單
position=position[position.columns[position.columns.isin(list_STC)]]
position=position[position.columns[~position.columns.isin(list_KY)]]
#出場
position=position.mask(position==False)
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'],qty ,position)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        earn=45,loss=15,
        fee = c*0.0001425,
        benchmark=benchmark,
        )

										
									

										
												
# 策略說明：價量分析及營收增長選股策略，以下是策略的主要條件：
# 一、「平均營收」定義：以近X個月的營收計算均值。
# 二、平均營收創近Y個月營收新高。
# 三、近Z日內至少有2日,股價創200日新高
# 四、近A日平均交易量 > B張
# 五、股池為上市 且 剔除KY
# 六、選擇滿足條件下：月營收年增率較大的10個。
# 七、在每個月的最後營收公告截止日換股。
# 八、手續費為C折 ，各個股票張數金額不超過D萬


import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=2
y=6
z=7
a=5
b=2000
c=3
d=20

BM='0050'
adjclose= msmp.日_K.adj_Close
adjopen= msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open = msmp.日_K.Open
list_STC = msmp.list['tse']
list_KY  = msmp.list['KY']

#股數
qty = np.floor(d*10000/(Open*1000))*1000
#近a月平均營收創b月來新高
rev = msmp.月_營收.金額_千
rev_ma= rev.rolling(x).mean()
con1 = rev_ma == rev_ma.rolling(y, min_periods=1).max()
#近n日內至少有2日,股價創200日新高
Close = msmp.日_K.Close
Closeweek=msmp.週_K.Close
N=int(np.ceil(200/7))
con2 = (Close>=Closeweek.rolling(N).max()).rolling(z).sum()>=2
#period日成交均量大於n張
vol = msmp.日_K.adj_Volume
con3 = vol.rolling(a).mean() > b
# 隔天沒有鎖漲停
con4 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
buy = con1&con2&con3&con4& (qty!=0)

#標的範圍
buy=buy[buy.columns[buy.columns.isin(list_STC)]]
buy=buy[buy.columns[~buy.columns.isin(list_KY)]]
# 符合選股條件的名單中，再選出單月營收年增率前10強，並在營收公告截止日換股。
rev_yoy_growth = msmp.月_營收.YoY
buy=buy*rev_yoy_growth
buy=buy.apply(lambda row: row.nlargest(10), axis=1).notna().ne(0)
setday=rev_yoy_growth.index_str_to_date().index
buy=buy[buy.index.isin(setday)]
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'],qty , buy)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        fee = c*0.0001425,
        benchmark=benchmark,
        )

										
									

										
												
# 一、以參數X Y Z參數得到macd
# 二、MACD>前日 ,且往上突破 信號線 
# 三、股價淨值比（PB）排名位於前 A 名
# 四、以上為進場條件，股池為上市剔除KY
# 五、出場條件：MACD 向下且死叉
# 六、手續費為 B 折 ,支股票最大開倉金額不超過 C 萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=12
y=26
z=9
a=200
b=3
c=20

BM='0050'
Open = msmp.日_K.Open
adjclose = msmp.日_K.adj_Close
adjhigh = msmp.日_K.adj_High
adjlow = msmp.日_K.adj_Low
ls_stc = msmp.list['tse']  
ls_KY = msmp.list['KY']  
pool = sorted([item for item in (ls_stc) if item not in set(ls_KY)])
index = adjclose.index.to_series()
index = index.groupby(index.dt.to_period('W')).first().tolist()

qty = np.floor(c * 10000 / (Open * 1000)) * 1000
ema12 = adjclose.ewm(span=x).mean()  # 短期EMA
ema26 = adjclose.ewm(span=y).mean()  # 長期EMA
macd = ema12 - ema26
signal_line = macd.ewm(span=z).mean()  # 計算信號線
con1 = macd>signal_line
con2 = macd.shift(1)<signal_line.shift(1)
con3 = macd.diff()>0  
con4 = msmp.日_指標.股價淨值比.rank(axis=1 ,ascending=True) <a
con5 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
long = con1 &con2 &con3 &con4 &con5 &(qty!=0)
long = long.mask(long==False)
long = long.loc[long.index.isin(index) , long.columns.isin(pool)]

out1 = macd<signal_line
out2 = macd.shift(1)>signal_line.shift(1)
out3 = macd.diff()<0
longout = out1 &out2 &out3 
longout = longout.loc[longout.index.isin(index)]

dic = msmp_to_dic(['qty', 'long' ,'longout'], qty, long ,longout)
benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp = api.backtest(dic ,symbol='id',date='date',qty='qty',    
    sig_value='adjclose',
    open_value='adjopen', open_shift=True,
    close_value='adjopen', close_shift=True,
    longsig='long',longout='longout',
    benchmark=benchmark,    fee=0.0001425 * b)






										
									

										
												
# 策略說明：基本面及技術面多因子選股策略，以下是策略的主要條件：
# 一、股票還原收盤價 < X元。
# 二、股票還原收盤價 > Y日均
# 三、選擇滿足條件下： 股價净值比較小的Z個。   
# 四、股池為上市 且 剔除KY           
# 五、每個月第一天輪換一次
# 六、手續費為A折 ，各個股票張數金額不超過B萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=50
y=60
z=10
a=3
b=20
BM='0050'

Close= msmp.日_K.Close
adjclose= msmp.日_K.adj_Close
adjopen= msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open   = msmp.日_K.Open
list_STC = msmp.list['tse']
list_KY = msmp.list['KY']
股價淨值比= msmp.日_指標.股價淨值比

#股數
qty = np.floor(b*10000/(Open*1000))*1000
con1= (adjclose < x)
con2= adjclose>adjclose.rolling(y).mean()
# 隔天沒有鎖漲停
con3 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
buy=con1&con2&con3&(qty!=0)

#標的範圍
buy=buy[buy.columns[buy.columns.isin(list_STC)]]
buy=buy[buy.columns[~buy.columns.isin(list_KY)]]
#換股(權重加成)
buy= buy*股價淨值比 *adjclose
buy= buy.apply(lambda row: row[row > 0].nsmallest(z), axis=1).notna().ne(0)
#出場
setday=buy.groupby([buy.index.year, buy.index.month]).apply(lambda x: x.index[0]).tolist()
buy=buy[buy.index.isin(setday)]
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'],qty , buy)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        fee = a*0.0001425,
        benchmark=benchmark,
        )

										
									

										
												
def max_df(*dfs):
    arr = np.stack([df.values for df in dfs])
    result = np.max(arr, axis=0)
    result_df = pd.DataFrame(result, index=dfs[0].index, columns=dfs[0].columns)
    return msmp.DataFrame(result_df)

def min_df(*dfs):
    arr = np.stack([df.values for df in dfs])
    result = np.min(arr, axis=0)
    result_df = pd.DataFrame(result, index=dfs[0].index, columns=dfs[0].columns)
    return msmp.DataFrame(result_df)

# 一、股池為所有 ETF
# 二、前一日的五線最大值（含收盤價、10MA、20MA、50MA、60MA） < 五線最小值的 X% 範圍內
# 三、當日收盤價向上突破昨日五線最大值
# 四、淨利/淨值的排名小於等於 Y
# 五、以上為入場條件，出場條件為: 收盤價已連續 Z日未創Z日新高
# 六、手續費為 B 折 ,股票最大開倉金額不超過 A 萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=4
y=50
z=7
a=20
b=3
BM='0050'

info = msmp.INFO
pool = info[info.ETF==1].股票代號.tolist()

Open = msmp.日_K.Open
data = msmp.日_指標.淨利 /msmp.日_指標.淨值
adjclose = msmp.日_K.adj_Close
adjhigh = msmp.日_K.adj_High
adjlow = msmp.日_K.adj_Low
ma1 = adjclose.rolling(10).mean()
ma2 = adjclose.rolling(20).mean()
ma3 = adjclose.rolling(50).mean()
ma4 = adjclose.rolling(60).mean()

df_max = max_df(adjclose ,ma1 ,ma2, ma3 ,ma4)
df_min = min_df(adjclose ,ma1 ,ma2, ma3 ,ma4)

qty = np.floor(a * 10000 / (Open * 1000)) * 1000
con1 = df_max.shift(1) < df_min.shift(1)*(1+0.01*x)
con2 = adjclose >df_max.shift(1)
con3 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
con4 = data.loc[:,data.columns.isin(pool)].rank(axis=1) <= y
pos = con1 &con2&con3&con4&(qty!=0)
pos = pos.mask(pos==False)
pos = pos.loc[: , pos.columns.isin(pool)]

out = (adjclose > adjclose.shift(1).rolling(z).max()).rolling(z).sum()==0

dic = msmp_to_dic(['qty', 'long' ,'longout'], qty, pos ,out)
benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp = api.backtest(dic ,
    symbol='id',date='date',qty='qty', sig_value='adjclose',
    open_value='adjopen', open_shift=True,
    close_value='adjopen', close_shift=True,
    longsig='long', longout='longout',
    benchmark=benchmark,    fee=0.0001425 * b)


										
									

										
												
# 策略說明：基本面及技術面多因子選股策略，以下是策略的主要條件：
# 一、股價創X日新高。
# 二、排除月營收連3月衰退10%
# 三、排除近一年月營收增長50%發生8次以上。
# 四、近月營收脫離近年谷底，達連續Y次。
# 五、連續3月近月營收MOM > Z%
# 六、10日均交易量 > A張
# 七、股池為上市
# 八、選擇滿足條件下：10日均交易量較大的B個。
# 九、在每個月的最後營收公告截止日換股。
# 十、手續費為C折 ，各個股票張數金額不超過D萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=60
y=3
z=-40
a=200
b=10
c=3
d=20
BM='0050'

adjclose =msmp.日_K.adj_Close
adjopen =msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open   = msmp.日_K.Open
vol = msmp.日_K.adj_Volume
vol_ma= vol.rolling(10).mean()
rev = msmp.月_營收.金額_千
rev_month_growth = msmp.月_營收.MoM
rev_year_growth = msmp.月_營收.YoY
list_STC = msmp.list['tse']

#股數
qty = np.floor(d*10000/(Open*1000))*1000
# 股價創年新高
con1 = (adjclose == adjclose.rolling(x, min_periods=20).max())
# 排除月營收連3月衰退10%以上&排除近一年漲太多
con2 = ~((rev_year_growth < -10).rolling(3).sum()==3)
con3 = ~((rev_year_growth> 50).rolling(12).sum()>=8)
# 確認營收底部，近月營收脫離近年谷底(連續3月的"單月營收近12月最小值/近月營收" < 0.8)
con4 = ((rev.rolling(12).min()/rev) < 0.8).rolling(y).sum()==y
# 單月營收月增率連續3月大於-40%
con5 = (rev_month_growth > z).rolling(3).sum()==3
# 流動性條件
con6 = vol_ma > a
# 隔天沒有鎖漲停
con7 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
buy = con1 & con2  & con3 & con4 & con5 & con6 & con7 &(qty!=0)

buy=buy[buy.columns[buy.columns.isin(list_STC)]]
#換股
buy = vol_ma*buy
buy = buy.apply(lambda row: row[row > 0].nsmallest(b), axis=1).notna().ne(0)
#周期
setday=rev.index_str_to_date().index
buy=buy[buy.index.isin(setday)]
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'],qty , buy)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        fee = c*0.0001425,
        benchmark=benchmark,
        )

										
									

										
												
# 本策略專門追蹤市場短期下跌趨勢，由以下四種條件組合。
# 一、X億 <市值 < Y億。
# 二、大盤在Z日均綫之上。
# 三、A日區間内平台整理，今日突破。
# 四、股池為上市 且 剔除KY
# 五、止盈為40%止損為15%，以發生信號的隔天開盤價，開倉及平倉。
# 六、手續費為B折 ，各個股票張數金額不超過C萬

import numpy as np
import pandas as pd
import kgisuperpy as kgi

person_id = 'A123456789'
person_pwd = '0000'
#simulation可輸入參數True：連線至測試環境；Fales：連線至正式環境
simulation = True

api = kgi.login(person_id ,person_pwd ,simulation)

msmp = api.MSMP
msmp_to_dic = api.MSMP.msmp_to_dic


x=5
y=20
z=20
a=20
b=3
c=20
BM='0050'

adjclose= msmp.日_K.adj_Close
adjopen= msmp.日_K.adj_Open
adjhigh= msmp.日_K.adj_High
adjlow= msmp.日_K.adj_Low
Open = msmp.日_K.Open
list_STC = msmp.list['tse']
list_KY  = msmp.list['KY']

def period_nhigh(data,period=7,n=1): #期間第n大
        return data.rolling(period).quantile((period-n)/(period-1))

#股數
qty = np.floor(c*10000/(Open*1000))*1000
#市值 <> n億
Marketvalue= msmp.日_K.市值_百萬
con1= Marketvalue >(x*100)
#市值 <> n億
con2= Marketvalue <(y*100)
#大盤在period日均綫之上
stc = msmp.TSE.adj_Close
con3= stc>stc.rolling(z).mean()
# p區間内震蕩,平台整理后突破
high4= period_nhigh(adjhigh,20,4)
low4=  period_nhigh(adjlow,20,4)
p1= adjhigh.rolling(a).max().shift(1)<adjlow.rolling(20).min().shift(1)*(1.1)
p2= adjhigh.rolling(a).max().shift(1)<high4.shift(1)*(1+0.05)
p3= adjlow.rolling(a).min().shift(1)*(1+0.05)>low4.shift(1)
p4= adjclose>adjhigh.rolling(10).max().shift(1)
con4= p1&p2&p3&p4
# 隔天沒有鎖漲停
con5 = ~((adjlow.shift(-1)>adjclose*1.0947)&(adjhigh.shift(-1)==adjlow.shift(-1)))
position=con1&con2&con3&con4&con5&(qty!=0)

#清單
position=position[position.columns[position.columns.isin(list_STC)]]
position=position[position.columns[~position.columns.isin(list_KY)]]
#出場
position=position.mask(position==False)
#轉換回測格式
dfX = msmp_to_dic(['qty','sig'],qty , position)

benchmark= api.Data.get('還原日K(個股、ETF、大盤)-單股單期' ,BM).set_index('日期')['收盤價']
rp=api.backtest(data=dfX,
        symbol='id',
        date='date',
        sig_value='adjclose',
        open_value='adjopen',  open_shift=True,   
        close_value='adjopen', close_shift=True,  
        qty='qty',
        longsig='sig',
        earn=40,loss=15,
        fee = b*0.0001425,
        benchmark=benchmark,
        )