start_date = '2016-01-01'
end_date = '2017-09-14'
instrument = D.instruments(start_date=start_date, end_date=end_date)

# 沪深300指数成分
df = D.history_data(instrument, start_date, end_date, ['in_csi300'])
instruments = list(set(df[df['in_csi300']==1]['instrument']))
print('沪深300指数成分股预览10只股票:', instruments[:10])

# （举例）获取国防军工行业股票列表
df = D.history_data(instrument, start_date, end_date, ['industry_sw_level1'])
instruments = list(set(df[df['industry_sw_level1'] == 650000]['instrument']))
D.history_data(instruments, '2017-07-27', '2017-07-27', ['company_name']).head() 

start_date='2017-05-23'
end_date = '2018-01-01'
df = D.history_data(D.instruments(start_date, end_date),start_date,end_date,fields=['in_csi300'])
# 获取沪深300的成分股股票列表  中证500：in_csi500，中证800：in_csi800  参考文档说明
instruments = list(set(df[df['in_csi300']==1]['instrument'])) 
print('特定时间段内沪深300指数成分股预览10只股票:',instruments[:10])

df = D.history_data(instrument, '2017-08-23', '2017-08-23', ['concept']).dropna()   # 'concept' 是股票的概念字段
df['is_ai'] = df['concept'].map(lambda x: '人工智能' in x)  # 以人工智能为例，找到相关概念股票
st = list(df[df['is_ai'] == True]['instrument'])
D.history_data(st,'2017-08-23','2017-08-23',['name', 'concept']).head()

df = D.history_data(instrument, start_date, end_date, ['list_date', 'list_board'])
from datetime import timedelta
df[(df['list_board']=='创业板')&(df['date']<=df['list_date']+timedelta(days=20))].head(10)

 # 获取市盈率、市净率、成交额数据
history_data = D.history_data(instrument, start_date=start_date, end_date= end_date, fields=[ 'pb_lf', 'pe_ttm','amount'])
result = history_data[(history_data['pb_lf'] < 1.5)
                                       & (history_data['pe_ttm'] < 15) 
                                       & (history_data['amount'] > 0) 
                                       & (history_data['pb_lf'] > 0)
                                       & (history_data['pe_ttm'] > 0)]
daily_buy_stock = result.groupby('date').apply(lambda df:list(df.sort_values(['pe_ttm','pb_lf']).instrument)[:30])
daily_buy_stock.head()

stock_num = 10
# 通过history_data接口获取历史数据
history_data = D.history_data(instrument, start_date, end_date, ['pb_lf', 'ps_ttm', 'market_cap', 'amount'])
# 通过feature接口获取财务数据
financial = D.features(instrument, start_date=start_date, end_date=end_date,
                        fields=['fs_operating_revenue_ttm_0','fs_current_assets_0','fs_non_current_assets_0']) 
# 总资产=流动资产+非流动资产
financial['total_assets'] = financial['fs_current_assets_0'] + financial['fs_non_current_assets_0']
# 资产周转率=营业收入/总资产
financial['asset_turnover'] = financial['fs_operating_revenue_ttm_0'] / financial['total_assets']
financial_data = financial[['date','instrument','asset_turnover']] 
# 两个DataFrame:历史数据、财务数据 合并
result = history_data.merge(financial_data, on=['date','instrument'], how='outer')  
# 按照选股法则选出股票
daily_buy_stock = result.groupby('date').apply(lambda df:list(df[(df['ps_ttm']>0.4) & (df['pb_lf']<8) & (df['asset_turnover']>0.4)
                                         & (df['amount']>100)].sort_values('market_cap')['instrument'])[:stock_num])
daily_buy_stock.head()

df= D.history_data(instrument, start_date, end_date, ['open', 'high', 'low', 'close'])

import talib as ta
from numpy import float as f

def seek_stocks(df):
    df['ma_7'] = ta.SMA(df.close.map(f).values, 7)
    df['ma_63'] = ta.SMA(df.close.map(f).values, 63)
    try:
        df['atr'] = ta.ATR(df.high.map(f).values, df.low.map(f).values, df.close.map(f).values, 14)
    except Exception as e:
        df['atr'] = np.nan
    df['upperline'] = df.close.rolling(14).mean() + df['atr'] 
    return df[(df['ma_7']>df['ma_63'])&(df['close']>=df['upperline'])].drop('instrument', axis=1)

result = df.groupby('instrument').apply(seek_stocks).reset_index()
daily_buy_stock = result.groupby('date').apply(lambda df:list(df['instrument']))
daily_buy_stock.head()

df =D.history_data(instrument, start_date, end_date, ['close', 'amount','high'])

def seek_stocks(df):
    df['highest_60'] = df['high'].rolling(60).max()  # 计算60天最高点
    df['ma3_cross_ma5'] = df['close'].rolling(3).mean() - df['close'].rolling(5).mean()  > 0  # 3日均线上穿5日均线
    df['ma5_cross_ma10'] = df['close'].rolling(5).mean() - df['close'].rolling(10).mean()  > 0  # 5日均线上穿10日均线
    df['amount_cond'] = df['amount'] / df['amount'].shift(1) - 1 >= 0.4   # 当日成交量比前一日成交量大40%
    prices = df['close'].map(np.float)  # 转化成float格式
    # macd:diff线 信号线:dea 柱状图：diff-dea
    macd, signal, hist = ta.MACD(np.array(prices), 12, 26, 9)  # 计算macd各个指标
    df['is_highest'] = df['close'] == df['highest_60']  # 该列是布尔型变量，表明是否是60日最高点
    df['hist_is_red'] =  hist > 0   # macd柱状图是否在红色区域
    return df
  
managed_df = df.groupby('instrument').apply(seek_stocks).reset_index()
result= managed_df[
    (managed_df['hist_is_red'])&   # macd在红色区域
    (managed_df['is_highest'])&   # 是60日最高点
    (managed_df['ma3_cross_ma5'])&  # 3日均线上穿5日均线
    (managed_df['ma5_cross_ma10'])&   # 5日均线上穿10日均线
    (managed_df['amount_cond'])]   # 满足成交量条件
    
# 整理出每日符合买入条件的列表               
daily_buy_stock = result.groupby('date').apply(lambda df:list(df.instrument)).reset_index().rename(columns={0:'stocks'})
daily_buy_stock.head()