请 [注册] 或 [登录]  | 返回主站

量化交易吧 /  数理科学 帖子:3364762 新帖:0

缠论K线图-笔-线段自动生成(17.6.22持续更新....)(迭代处理最后一段笔)让你更清晰看清买卖点

交易资深人士发表于:5 月 10 日 02:12回复(1)

缠论笔的基础就不再为大家阐述了
下面工具展示两种缠论K线 1、用普通K线展示 2、用处理了包含关系的缠论K线展示(强烈推荐)
!!!补充一句,特别感谢我们的缠论张老师 '杀手张一' 为我们提供迭代思路
!!! 硬货来了,2017-06-22 补充了线段,更方便大家查看

================需要修改的参数==============
股票代码, 生成k线的结束日期, 看几天/分钟前的k线,轴展示的时间距离,所看级别,是否是缠论K线

使用方法:1、克隆研究 2、复制my_chan.py 到研究下(创建文件my_chan.py 把代码拷入) 3、修改以上参数 并运行即可

!!!这里最后一段未完成的笔,迭代次级别,如果次级别形成线段,那么此级别就是一笔(对应此级别:日-30分钟-5分钟-1分钟)
这样能够更准确的找到最后一笔下或上,从而找买卖点
!!!这里 K线周期可以看 1d,30m,5m, 1m, 因为最后一笔只迭代了这几个时间段的笔,如果大家需要看其他时间的笔
请注释 代码 line 251:biIdx = con2Cxianduan(stock_code, k_data, chanK, frsBiType, biIdx, end_date, cur_ji) 即可

# 缠论K线图展示完整版

import my_chan as chan
import matplotlib as mat
import numpy as np
import datetime as dt
import matplotlib.pyplot as plt
# bokeh.plotting
import time

from bokeh.models import ColumnDataSource, Rect, HoverTool, Range1d, LinearAxis, WheelZoomTool, PanTool, ResetTool, ResizeTool, PreviewSaveTool
# ================需要修改的参数==============
stock_code = '002560.XSHE' # 股票代码
end_date = '2017-06-22 15:00:00' # 最后生成k线日期
stock_days = 6 # 看几天/分钟前的k线 
x_jizhun = 20 #x轴展示的时间距离  5:日,40:30分钟, 48: 5分钟
# stock_frequency = '5m' # 1d日线, 30m 30分钟, 5m 5分钟,1m 1分钟
stock_frequency = '1m' # 1d日线, 30m 30分钟, 5m 5分钟,1m 1分钟
chanK_flag = True# True 看缠论K线, False 看k线
# ============结束==================


initial_trend = "down"
cur_ji = 1 if stock_frequency=='1d' else \
2 if stock_frequency=='30m' else \
3 if stock_frequency=='5m' else 4


print '======笔形成最后一段未完成段判断是否是次级别的走势形成笔======='
def con2Cxianduan(stock, k_data, chanK, frsBiType, biIdx, end_date, cur_ji = 1):
    max_k_num = 4
    if cur_ji>=4 or len(biIdx)==0: return biIdx
    idx = biIdx[len(biIdx)-1]
    k_data_dts = list(k_data.index)
    st_data = chanK['enddate'][idx]
    if st_data not in k_data_dts: return biIdx
    # 重构次级别线段的点到本级别的chanK中
    def refactorXd(biIdx, xdIdxc, chanK, chanKc, cur_ji):
        new_biIdx = []
        biIdxB = biIdx[len(biIdx)-1] if len(biIdx)>0 else 0
        for xdIdxcn in xdIdxc:
            for chanKidx in range(len(chanK.index))[biIdxB:]:
                if judge_day_bao(chanK, chanKidx, chanKc, xdIdxcn, cur_ji):
                    new_biIdx.append(chanKidx)
                    break
        return new_biIdx
    # 判断次级别日期是否被包含
    def judge_day_bao(chanK, chanKidx, chanKc, xdIdxcn, cur_ji):
        _end_date = chanK['enddate'][chanKidx]+datetime.timedelta(hours=15) if cur_ji==1 else chanK['enddate'][chanKidx]
        _start_date = chanK.index[chanKidx] if chanKidx==0\
        else chanK['enddate'][chanKidx-1]+datetime.timedelta(minutes=1)
#         print '连接包含日期', chanK.index[chanKidx], chanKc.index[xdIdxcn], chanK['enddate'][chanKidx]
        return _start_date<=chanKc.index[xdIdxcn]<=_end_date
    # cur_ji = 1 #当前级别
    #符合k线根数大于4根 1日级别, 2 30分钟, 3 5分钟, 4 一分钟
    if len(k_data_dts) - k_data_dts.index(st_data)>4:
        frequency = '30m' if cur_ji+1==2 else '5m' if cur_ji+1==3 else '1m'
#         print "次级别为:%s"%frequency
#         print chanK
        k_data_c = get_price(stock, st_data, end_date, frequency=frequency)
        chanKc = chan.parse2ChanK(k_data_c, k_data_c.values)
        fenTypesc, fenIdxc = chan.parse2ChanFen(chanKc)
        if len(fenTypesc)==0: return biIdx
        biIdxc, frsBiTypec = chan.parse2ChanBi(fenTypesc, fenIdxc, chanKc)
        if len(biIdxc)==0: return biIdx
        xdIdxc, xdTypec = chan.parse2Xianduan(biIdxc, chanKc)
        biIdxc = chan.con2Cxianduan(stock, k_data_c, chanKc, frsBiTypec, biIdxc, end_date, cur_ji+1)
        if len(xdIdxc)==0: return biIdx
#         print cur_ji, '次级别线段', xdIdxc, xdTypec
        #连接线段位为上级别的bi
        lastBiType = frsBiType if len(biIdx)%2==0 else -frsBiType
        if len(biIdx)==0:
            return refactorXd(biIdx, xdIdxc, chanK, chanKc, cur_ji)
        lastbi = biIdx.pop()
        firstbic = xdIdxc.pop(0)
#         print '最后一笔下和连段上',lastBiType, xdTypec
        # 同向连接
        if lastBiType == xdTypec:
            print '同向连接'
            biIdx = biIdx + refactorXd(biIdx, xdIdxc, chanK, chanKc, cur_ji)
        # 逆向连接
        else:
#             print '开始逆向连接'
#             print chanKc.index[firstbic],len(chanK.index), biIdx[len(biIdx)-1]
#             print lastBiType, chanK['low'][lastbi], chanKc['low'][firstbic], chanK['high'][lastbi],chanKc['high'][firstbic]
            _mid = [lastbi] if (lastBiType == -1 and chanK['low'][lastbi]<=chanKc['low'][firstbic])\
            or (lastBiType == 1 and chanK['high'][lastbi]>=chanKc['high'][firstbic]) else\
            [chanKidx for chanKidx in range(len(chanK.index))[biIdx[len(biIdx)-1]:]\
            if judge_day_bao(chanK, chanKidx, chanKc, firstbic, cur_ji)]
#             print '中间点', _mid
#             print '日期比较', '次级别', frequency
#             for chanKidx in range(len(chanK.index))[biIdx[len(biIdx)-1]:]:
#                 print chanK.index[chanKidx], chanKc.index[firstbic],\
#                 chanK['enddate'][chanKidx], chanK.index[chanKidx]<=chanKc.index[firstbic]<=(chanK['enddate'][chanKidx]\
#                 if cur_ji>1 else chanK['enddate'][chanKidx]+datetime.timedelta(hours=15))
            biIdx = biIdx + [_mid[0]] + refactorXd(biIdx, xdIdxc, chanK, chanKc, cur_ji)
#     print '最后'
    return biIdx
'''
以下代码拷贝自https://www.joinquant.com/post/1756
感谢alpha-smart-dog
'''
# dt.datetime.strptime(end_date, "%Y-%m-%d %H:%M:%S")-dt.timedelta(days=5)
quotes = get_price(stock_code, datetime.datetime.strptime(end_date, "%Y-%m-%d %H:%M:%S")-datetime.timedelta(days=stock_days) , end_date,\
                   frequency=stock_frequency,skip_paused=False,fq='pre')
# print quotes
# 缠论k线
quotes = chan.parse2ChanK(quotes, quotes.values) if chanK_flag else quotes
# print quotes
quotes[quotes['volume']==0]=np.nan
quotes= quotes.dropna()
Close=quotes['close']
Open=quotes['open']
High=quotes['high']
Low=quotes['low']
T0 = quotes.index.values

length=len(Close)

# ht = HoverTool(tooltips=[
#             ("date", "@date"),
#             ("open", "@open"),
#             ("close", "@close"),
#             ("high", "@high"),
#             ("low", "@low"),
#             ("volume", "@volume"),
#             ("money", "@money"),])
# TOOLS = [ht, WheelZoomTool(dimensions=['width']),\
#          ResizeTool(), ResetTool(),\
#          PanTool(dimensions=['width']), PreviewSaveTool()]

fig = plt.figure(figsize=(16, 8))
ax1 = plt.subplot2grid((10,4),(0,0),rowspan=10,colspan=4)
#fig = plt.figure()
#ax1 = plt.axes([0,0,3,2])

X=np.array(range(0, length))
pad_nan=X+nan

    #计算上 下影线
max_clop=Close.copy()
max_clop[Close<Open]=Open[Close<Open]
min_clop=Close.copy()
min_clop[Close>Open]=Open[Close>Open]

    #上影线
line_up=np.array([High,max_clop,pad_nan])
line_up=np.ravel(line_up,'F')
    #下影线
line_down=np.array([Low,min_clop,pad_nan])
line_down=np.ravel(line_down,'F')

    #计算上下影线对应的X坐标
pad_nan=nan+X
pad_X=np.array([X,X,X])
pad_X=np.ravel(pad_X,'F')

    #画出实体部分,先画收盘价在上的部分
up_cl=Close.copy()
up_cl[Close<=Open]=nan
up_op=Open.copy()
up_op[Close<=Open]=nan

down_cl=Close.copy()
down_cl[Open<=Close]=nan
down_op=Open.copy()
down_op[Open<=Close]=nan

even=Close.copy()
even[Close!=Open]=nan

#画出收红的实体部分
pad_box_up=np.array([up_op,up_op,up_cl,up_cl,pad_nan])
pad_box_up=np.ravel(pad_box_up,'F')
pad_box_down=np.array([down_cl,down_cl,down_op,down_op,pad_nan])
pad_box_down=np.ravel(pad_box_down,'F')
pad_box_even=np.array([even,even,even,even,pad_nan])
pad_box_even=np.ravel(pad_box_even,'F')

#X的nan可以不用与y一一对应
X_left=X-0.25
X_right=X+0.25
box_X=np.array([X_left,X_right,X_right,X_left,pad_nan])
# print box_X
box_X=np.ravel(box_X,'F')
# print box_X
#Close_handle=plt.plot(pad_X,line_up,color='k') 

vertices_up=array([box_X,pad_box_up]).T
vertices_down=array([box_X,pad_box_down]).T
vertices_even=array([box_X,pad_box_even]).T

handle_box_up=mat.patches.Polygon(vertices_up,color='r',zorder=1)
handle_box_down=mat.patches.Polygon(vertices_down,color='g',zorder=1)
handle_box_even=mat.patches.Polygon(vertices_even,color='k',zorder=1)

ax1.add_patch(handle_box_up)
ax1.add_patch(handle_box_down)
ax1.add_patch(handle_box_even)

handle_line_up=mat.lines.Line2D(pad_X,line_up,color='k',linestyle='solid',zorder=0) 
handle_line_down=mat.lines.Line2D(pad_X,line_down,color='k',linestyle='solid',zorder=0) 

ax1.add_line(handle_line_up)
ax1.add_line(handle_line_down)

v=[0,length,Open.min()-0.5,Open.max()+0.5]
plt.axis(v)

T1 = T0[-len(T0):].astype(dt.date)/1000000000
# print T1
Ti=[]
for i in range(len(T0)/x_jizhun):
    a=i*x_jizhun
    d = dt.date.fromtimestamp(T1[a])
#     print d
    T2=d.strftime('$%Y-%m-%d$')
    Ti.append(T2)
    #print tab
d1= dt.date.fromtimestamp(T1[len(T0)-1])
d2=(d1+datetime.timedelta(days=1)).strftime('$%Y-%m-%d$')
Ti.append(d2)

ax1.set_xticks(np.linspace(-2,len(Close)+2,len(Ti))) 

ll=Low.min()*0.97
hh=High.max()*1.03
ax1.set_ylim(ll,hh) 

ax1.set_xticklabels(Ti)

plt.grid(True)
plt.setp(plt.gca().get_xticklabels(), rotation=45, horizontalalignment='right')

'''
以上代码拷贝自https://www.joinquant.com/post/1756
感谢alpha-smart-dog

K线图绘制完毕
'''
x_date_list = quotes.index.values.tolist()
# for x_date in x_date_list:
#     d = datetime.datetime.fromtimestamp(x_date/1000000000)
#     print d.strftime("%Y-%m-%d %H:%M:%S.%f")
# print x_date_list
k_data = quotes
k_values = k_data.values
# 缠论k线
chanK = quotes if chanK_flag else chan.parse2ChanK(k_data, k_values)
fenTypes, fenIdx = chan.parse2ChanFen(chanK)
# print "分型", fenTypes, fenIdx
biIdx, frsBiType = chan.parse2ChanBi(fenTypes, fenIdx, chanK)
# biIdx = con2Cxianduan(stock_code, k_data, chanK, frsBiType, biIdx, end_date, cur_ji)  
# print chanK
print "笔", biIdx, frsBiType
print '股票代码', get_security_info(stock_code).display_name
#  3.得到分笔结果,计算坐标显示
x_fenbi_seq = []
y_fenbi_seq = []
for i in range(len(biIdx)):
    if biIdx[i]:
        fenType = -frsBiType if i%2==0 else frsBiType
#         dt = chanK['enddate'][biIdx[i]]
          # 缠论k线
        dt = chanK.index[biIdx[i]] if chanK_flag else chanK['enddate'][biIdx[i]]
#         print k_data['high'][dt], k_data['low'][dt]
        time_long = long(time.mktime((dt+datetime.timedelta(hours=8)).timetuple())*1000000000)
#         print x_date_list.index(time_long) if time_long in x_date_list else 0
        if fenType == 1:
            plt.text(x_date_list.index(time_long), k_data['high'][dt], \
                     str(k_data['high'][dt]), ha='left', fontsize=12)
            x_fenbi_seq.append(x_date_list.index(time_long))
            y_fenbi_seq.append(k_data['high'][dt])
        if fenType == -1:
            plt.text(x_date_list.index(time_long), k_data['low'][dt], \
                     str(k_data['low'][dt]), va='bottom', fontsize=12)
            x_fenbi_seq.append(x_date_list.index(time_long))
            y_fenbi_seq.append(k_data['low'][dt])
#             bottom_time = None
#             for k_line_dto in m_line_dto.member_list[::-1]:
#                 if k_line_dto.low == m_line_dto.low:
#                     # get_price返回的日期,默认时间是08:00:00
#                     bottom_time = k_line_dto.begin_time.strftime('%Y-%m-%d') +' 08:00:00'
#                     break
#             x_fenbi_seq.append(x_date_list.index(long(time.mktime(datetime.strptime(bottom_time, "%Y-%m-%d %H:%M:%S").timetuple())*1000000000)))
#             y_fenbi_seq.append(m_line_dto.low)

#  在原图基础上添加分笔蓝线
plt.plot(x_fenbi_seq,y_fenbi_seq)

#线段画到笔上
xdIdxs,xfenTypes = chan.parse2ChanXD(frsBiType,biIdx,chanK)
print '线段', xdIdxs,xfenTypes
x_xd_seq = []
y_xd_seq = []
for i in range(len(xdIdxs)):
    if xdIdxs[i]:
        fenType = xfenTypes[i]
#         dt = chanK['enddate'][biIdx[i]]
          # 缠论k线
        dt = chanK.index[xdIdxs[i]] if chanK_flag else chanK['enddate'][xdIdxs[i]]
#         print k_data['high'][dt], k_data['low'][dt]
        time_long = long(time.mktime((dt+datetime.timedelta(hours=8)).timetuple())*1000000000)
#         print x_date_list.index(time_long) if time_long in x_date_list else 0
        if fenType == 1:
            x_xd_seq.append(x_date_list.index(time_long))
            y_xd_seq.append(k_data['high'][dt])
        if fenType == -1:
            x_xd_seq.append(x_date_list.index(time_long))
            y_xd_seq.append(k_data['low'][dt])
#             bottom_time = None
#             for k_line_dto in m_line_dto.member_list[::-1]:
#                 if k_line_dto.low == m_line_dto.low:
#                     # get_price返回的日期,默认时间是08:00:00
#                     bottom_time = k_line_dto.begin_time.strftime('%Y-%m-%d') +' 08:00:00'
#                     break
#             x_fenbi_seq.append(x_date_list.index(long(time.mktime(datetime.strptime(bottom_time, "%Y-%m-%d %H:%M:%S").timetuple())*1000000000)))
#             y_fenbi_seq.append(m_line_dto.low)

#  在原图基础上添加分笔蓝线
plt.plot(x_xd_seq,y_xd_seq)


plt.show()


#  在原图基础上添加分笔蓝线
# plt.plot(x_fenbi_seq,y_fenbi_seq)

# plt.show()
======笔形成最后一段未完成段判断是否是次级别的走势形成笔=======
笔 [2, 10, 17, 48, 53, 64, 68, 72, 76, 88, 94, 103, 109, 127, 158, 178, 183, 194, 221, 229, 235, 239, 249, 254, 261, 265, 270, 281, 290] 1
股票代码 通达股份
线段 [2, 127, 221, 254, 290] [-1, 1, -1, 1, -1]
# coding=utf-8 
# 研究中创建my_chan.py 拷贝此代码即可
import pandas as pd
import numpy as np
import datetime

# 处理k线成缠论k线,临时函数
def parse2ChanKTMP(k_data, k_values, in_chan = False):
    baohan_columns = ['enddate'] + list(k_data.keys()) if not in_chan else list(k_data.keys())
#     print baohan_columns
    dfbao = pd.DataFrame(np.zeros(len(k_data.index)*len(baohan_columns)).reshape(len(k_data.index),len(baohan_columns)), index = k_data.index,columns=baohan_columns)\
    if not in_chan else k_data
#     print dfbao
    # stdt = k_data.index[0]
    # dfbao.set_value(stdt, baohan_columns,  [stdt] + list(k_data.values[0]))
    # 比较上一个包含的时间点
    dt_bf = 0
    def set_baohan_value(dt, value):
        dfbao.set_value(dt, baohan_columns,  value)
        return dt
    for i, dt in enumerate(k_data.index):
        if i==0:
            dt_bf = set_baohan_value(dt, [dt] + list(k_values[i])) if not in_chan else dt
            continue
        if in_chan and (k_data['enddate'][dt]==0 or i==0): continue
        # 非包含情况
        if (k_data['high'][dt]>dfbao['high'][dt_bf] and k_data['low'][dt]>dfbao['low'][dt_bf])\
        or (k_data['high'][dt]<dfbao['high'][dt_bf] and k_data['low'][dt]<dfbao['low'][dt_bf]):
            if in_chan:
                dt_bf = dt
                continue
            dt_bf = set_baohan_value(dt, [dt] + list(k_values[i]))
            continue
        # 包含情况
        enddate = dt if not in_chan else dfbao['enddate'][dt]
        dt_bf = set_baohan_value(dt_bf, \
                                 [enddate, dfbao['open'][dt_bf], k_data['close'][dt], \
                                 max(k_data['high'][dt], dfbao['high'][dt_bf]), \
                                 min(k_data['low'][dt], dfbao['low'][dt_bf]), \
                                 k_data['volume'][dt]+dfbao['volume'][dt_bf], \
                                 k_data['money'][dt]+dfbao['money'][dt_bf]])
        if in_chan:
            dfbao.set_value(dt, ['enddate'], 0)
    dfbao = dfbao[dfbao['enddate']!=0]
    return dfbao
# 处理k线成缠论k线
def parse2ChanK(k_data, k_values):  
    chanK = parse2ChanKTMP(k_data, k_values)
    while True:
        chanK1 = parse2ChanKTMP(chanK, chanK.values, True)
        if(len(chanK.index)==len(chanK1.index)):
            break
        chanK = chanK1
    return chanK
# chanK = parse2ChanK(k_data, k_values)
# print chanK
# 找顶底分型的idx
# 如果 连续顶顶,或底底: 顶:high最大的顶, 低:low最小的低
# 顶:1 低:-1
def parse2ChanFen(chanK):
    fenTypes = [] # 分型类型数组 1,-1构成
    fenIdx = [] # 分型对应缠论k的下标

    # 添加分型数据
    # 过滤连续同分型
    def appendFen(ft, fidx):
        if len(fenIdx)==0:
            fenTypes.append(ft)
            fenIdx.append(fidx)
            return
        fenType_bf = fenTypes[len(fenTypes)-1]
        if fenType_bf == ft:
             fenType_bf, fenIdx_bf = fenTypes.pop(), fenIdx.pop()
             fidx = fenIdx_bf if (ft==1 and chanK['high'][fenIdx_bf]>chanK['high'][fidx])\
             or (ft==-1 and chanK['low'][fenIdx_bf]<chanK['low'][fidx]) else fidx
        fenTypes.append(ft)
        fenIdx.append(fidx)
    for i, dt in enumerate(chanK.index):
        if i==0 or i==len(chanK.index)-1:continue
        # 顶分型
        if chanK['high'][i+1]<chanK['high'][i]>chanK['high'][i-1]:
            appendFen(1, i)
        # 底分型
        if chanK['low'][i+1]>chanK['low'][i]<chanK['low'][i-1]:
            appendFen(-1, i)
    return fenTypes, fenIdx
# fenTypes, fenIdx = parse2ChanFen(chanK)

# print fenTypes, fenIdx

# 分型构成笔
# 构成笔条件,1、顶低分型间隔了n个chanK线, 2、中间不会出现比第一个分型结构更高(顶)或更低(底)的分型,否则线段破坏,连接上一笔
def parse2ChanBi(fenTypes, fenIdx, chanK):
    biIdx = []  # 笔对应的缠论k线 idx
    frsBiType = 0 # 起始笔的走势,biIdx 奇数下标就是相反走势 1、向上,-1向下
    least_khl_num = 3 # 分笔间隔的最小 chanK 数量 中间排除顶低的chanK
    toConBfIdx = 0 # 连接到上一笔末尾的 分型idx
    # 判断笔破坏
    def judgeBiBreak(idxb, idxa):
        fenType = fenTypes[idxb]
        fenType1 = -fenType
#         print '分型破坏前', fenType, fenIdx[idxb], fenIdx[idxa]
        _break = False
        _breaki_k = 0
        _breakj_k = 0
        for k in range(idxb, idxa)[2::2]:
            # 当前i分型破坏
            if judgeBreak(fenType, fenIdx[idxb], fenIdx[k]):
#                 print '首分型破坏里', fenType, fenIdx[idxb], fenIdx[k], k
                _break, _breaki_k = True, k
                break
        for k in range(idxb, idxa)[1::2]:
            # 末尾j分型破坏
            if judgeBreak(fenType1, fenIdx[idxa], fenIdx[k]):
#                 print '末尾分型破坏里', fenType1, fenIdx[k], fenIdx[idxa],
                _break, _breakj_k = True, k
                break
#         print '破坏结果', _break, fenIdx[_breaki_k], fenIdx[_breakj_k]
        return _break, _breaki_k, _breakj_k
    # 分型破坏
    def judgeBreak(fenType, bf, af):
        return (fenType==-1 and chanK['low'][af]<chanK['low'][bf])\
            or (fenType==1 and chanK['high'][af]>chanK['high'][bf])
    def reAssignBi(biIdx, breakBi, breakBj, i, j):
        toConBfIdx = i+1 if len(biIdx)==0 else breakBi
        if breakBi>0 and len(biIdx)>0:
            fb_ = biIdx.pop()
#                         print '首分型破坏, 旧分型:%d, 新的分型:%d'%(fb_,fenIdx[breakBi])
            biIdx.append(fenIdx[breakBi])
            if 0<breakBj<breakBi and judgeBreak(fenTypes[i-1], biIdx[len(biIdx)-2], fenIdx[breakBj]):
                fa_=biIdx.pop()
                fb_=biIdx.pop()
#                             print '尾分型破坏并连接上一点, 移除分型:%d,旧分型:%d, 新的分型:%d'%(fa_, fb_,fenIdx[breakBj])
                biIdx.append(fenIdx[breakBj])
                toConBfIdx = breakBj
            return toConBfIdx,  -1 # -1:break 1:continue 0:不执行
        if breakBj>0: 
#                         breakBj = breakBj1 if breakBj==-1 or judgeBreak(fenTypes[i-1], fenIdx[breakBj], fenIdx[breakBj1]) else breakBj
            if j+2>=len(fenIdx) and len(biIdx)>1 and\
            judgeBreak(fenTypes[i-1], biIdx[len(biIdx)-2], fenIdx[breakBj]):
                fa_=biIdx.pop()
                fb_=biIdx.pop()
#                             print '尾分型破坏并连接上一点, 移除分型:%d,旧分型:%d, 新的分型:%d'%(fa_, fb_,fenIdx[breakBj])
                biIdx.append(fenIdx[breakBj])
                toConBfIdx = breakBj
                return toConBfIdx,  -1 # -1:break 1:continue 0:不执行
            return toConBfIdx,  1 # -1:break 1:continue 0:不执行
        return toConBfIdx,  0 # -1:break 1:continue 0:不执行
    for i, kidx in enumerate(fenIdx):
#         print '生成的笔', biIdx
#         print toConBfIdx, 'the i is ', i
        if i<toConBfIdx or i==len(fenIdx)-1 : continue
        # 后面没有符合条件的笔
        if len(biIdx)>1 and toConBfIdx==0:break
        toConBfIdx = 0
        for j in range(len(fenIdx))[i+1::2]:
#             print '差是', fenIdx[j]-kidx
            if (fenIdx[j]-kidx)>least_khl_num:
#                 print 'append', i, j, fenIdx[i], fenIdx[j]
                # breakType True 同分型, False 末尾分型
                flag, breakBi, breakBj = judgeBiBreak(i, j)
#                 print flag, breakBi 
                if flag: 
                    toConBfIdx, _bcn = reAssignBi(biIdx, breakBi, breakBj, i, j)
                    if _bcn==-1:break
                    if _bcn==1:continue
                if len(biIdx)==0: 
                    biIdx.append(kidx)
                    frsBiType = -fenTypes[i]       
                    biIdx.append(fenIdx[j])
                    toConBfIdx, _bcn = reAssignBi(biIdx, breakBi, breakBj, i, j)
                    if _bcn==-1 or _bcn==1:
                        biIdx = []
                        toConBfIdx = i+1
                        break  
                    toConBfIdx = j
                    break
                biIdx.append(fenIdx[j])
#                 print biIdx
                toConBfIdx = j
#                 print toConBfIdx
                break
        
    return biIdx, frsBiType
#最终线段生成
#1、遍历相对高低点,判断线段破坏,
#  破坏以后,如果总长度是0,i可以后移,否则重构之前的线段
#  重构规则,找破坏点相对高点/低点,如果存在线段高点/低点>/<破坏高点/低点, 则连接此线段
#2、形成线段中间至少有两点j-i>2
def parse2ChanXD(frsBiType,biIdx,chanK):
    lenBiIdx = len(biIdx)
    xdIdx = []
    xfenTypes = []
    if lenBiIdx==0 : return xdIdx,xfenTypes
    afIdx = 0
    #重构线段    
    def refactorXd(txIdx, nxIdx, chanK):
        xdIdxn = xdIdx
        xfenTypesn = xfenTypes
        if len(xdIdxn)==0 or txIdx==-1 or nxIdx==-1: return 0,xdIdxn,xfenTypesn
        for m in range(-len(xdIdxn)+1, 1)[1::2]:
            k = -m
            #满足逆向不破坏
#             print '开始逆向破坏:逆向分型,逆向点,线段点', xfenTypesn[k], nxIdx, xdIdxn[k]
            if (xfenTypesn[k]==-1 and chanK['low'][xdIdxn[k]]<chanK['low'][nxIdx])\
            or (xfenTypesn[k]==1 and chanK['high'][xdIdxn[k]]>chanK['high'][nxIdx]):
                for n in range(-len(xdIdxn)+1, m): 
                    xfenTypesn.pop()
                    xdIdxn.pop()
                xdIdxn.append(txIdx)
                xfenTypesn.append(-xfenTypesn[len(xfenTypes)-1])
                return biIdx.index(txIdx), xdIdxn, xfenTypesn
        xdIdxn = []
        xfenTypesn = []
#         print '逆向破坏', nxIdx, xdIdxn, xfenTypesn
        return biIdx.index(nxIdx), xdIdxn, xfenTypesn
    # 判断线段破坏
    def judgeBreak(fenType, afPrice, idx, chanK):
        return (fenType==-1 and afPrice<chanK['low'][idx]) \
    or (fenType==1 and afPrice>chanK['high'][idx])
    for i, idx in enumerate(biIdx):
        if afIdx < 0 :break # 线段破坏以后没有合适线段
        fenType = 1 if (frsBiType==1 and i%2==1) or (frsBiType==-1 and i%2==0) \
        else -1
#         print '开始判断%d,分型类型%d'%(idx, fenType)
        #符合要求的连段
        if i<afIdx:continue              
        #找同向相对高低点
        afPrice = 0 if fenType==-1 else 10000
        tongxiang_price_ = 10000 - afPrice
        nixiang_idx, tongxiang_idx = -1, -1
        i_continued = False
        for j in range(i+1, lenBiIdx)[0::2]:
            #同向相对高低点
            if (fenType==-1 and tongxiang_price_>chanK['low'][biIdx[j-1]]) \
            or (fenType==1 and tongxiang_price_<chanK['high'][biIdx[j-1]]):
                tongxiang_price_ = chanK['high'][biIdx[j-1]] if fenType==1 else chanK['low'][biIdx[j-1]]
                tongxiang_idx = biIdx[j-1]
            #线段破坏
#             print '线段破坏前', idx, tongxiang_idx
            #同向破坏
            if judgeBreak(fenType, tongxiang_price_, idx, chanK) and idx!=tongxiang_idx:
#                 print '同向已经破坏'
                afIdx, xdIdx, xfenTypes = refactorXd(tongxiang_idx, nixiang_idx, chanK)
                i_continued = True
                break     
#             print '符合要求前', biIdx[i], biIdx[j], afPrice, chanK['high'][biIdx[j]], chanK['low'][biIdx[j]],fenType
            if (fenType==-1 and chanK['high'][biIdx[j]]>afPrice) or (fenType==1 and chanK['low'][biIdx[j]]<afPrice):
                afPrice = chanK['high'][biIdx[j]] if fenType==-1 else chanK['low'][biIdx[j]]
                nixiang_idx = biIdx[j]
                #线段不符合要求
#                 print '符合要求的i,j', biIdx[i], biIdx[j]
                if j-i<=2:continue    
                #逆向破坏
#                 if judgeBreak(-fenType, nixiang_idx, idx, chanK) and idx!=nixiang_idx:
# #                     print '逆向已经破坏'
#                     afIdx, xdIdx, xfenTypes = refactorXd(tongxiang_idx, nixiang_idx, chanK)
#                     break 
                if len(xdIdx)==0:
#                     print '线段长度为0破坏', fenType, idx, judgeBreak(fenType, tongxiang_price_, idx, chanK)
                    if judgeBreak(fenType, tongxiang_price_, idx, chanK):
                        i_continued = True
                        break
                    xfenTypes.append(fenType)
                    xdIdx.append(idx)
                    xdIdx.append(biIdx[j])
                    xfenTypes.append(-fenType)
                else:
                    #不用同向线段连接
#                     fenTypeb = xfenTypes[len(xfenTypes)-1]
#                     xdIdxb = xdIdx.pop()
#                     if fenTypeb == -fenType:
#                         xdIdx.append(biIdx[j])
#                     else:
#                         xdIdx.append(xdIdxb)
#                         xfenTypes.append(-fenTypeb)
#                         xdIdx.append(biIdx[j])
                    xfenTypes.append(-xfenTypes[len(xfenTypes)-1])
                    xdIdx.append(biIdx[j])
                afIdx = j
                i_continued = True
                break
            else:continue
        if not i_continued and len(xdIdx)>0 : 
            #都不符合要求时,最后重构最小线段
            last_idx = xdIdx.pop()
            last_type = xfenTypes[len(xfenTypes)-1]
            for j in range(biIdx.index(last_idx), len(biIdx))[2::2]:
                if judgeBreak(last_type, chanK['low'][biIdx[j]] if last_type==-1 else chanK['high'][biIdx[j]], last_idx, chanK):
                    last_idx = biIdx[j]
            xdIdx.append(last_idx)
            break
    return xdIdx,xfenTypes
# biIdx, frsBiType = parse2ChanBi(fenTypes, fenIdx, chanK)
# print biIdx,frsBiType
#简单线段形成,主要用于判断是否是大级别的笔
#区间找高低点,判断是否符合 高低点中包含>2个笔点就阔以
def parse2Xianduan(biIdx, chanK):
    xdIdx = []
    if len(biIdx)==0: return xdIdx, 0
    def appendXd(lowIdx, highIdx, xdType):
        if len(xdIdx)==0:
            if xdType == 1:
                xdIdx.append(lowIdx)
            else:
                xdIdx.append(highIdx) 
        if (xdType == 1 and len(xdIdx)%2==1) or (xdType == -1 and len(xdIdx)%2==0):
            xdIdx.append(highIdx)
        else:
            xdIdx.append(lowIdx)
    def genXianduan(biIdx, chanK, xdType=0):
        highMax, lowMin = 0, 10000
        highIdx, lowIdx = -1,-1
        lenXd = len(xdIdx)
        for idx in  biIdx:
            if chanK['high'][idx]>highMax:
                highMax = chanK['high'][idx] 
                highIdx = idx
            if chanK['low'][idx]<lowMin:
                lowMin = chanK['low'][idx] 
                lowIdx = idx
                
        # 构成简易线段
#         print biIdx, xdIdx, lowIdx, highIdx
        xdDiff = biIdx.index(lowIdx)-biIdx.index(highIdx)
        if abs(xdDiff)>2:
            if lenXd==0:
                xdType = 1 if xdDiff<0 else -1
            appendXd(lowIdx, highIdx, xdType)
#             print lowIdx, highIdx, xdIdx
            genXianduan(biIdx[biIdx.index(xdIdx[len(xdIdx)-1]):], chanK, xdType)
        return xdType
    xdType = genXianduan(biIdx, chanK)
    return xdIdx,xdType

# 笔形成最后一段未完成段判断是否是次级别的走势形成笔
def con2Cxianduan(stock, k_data, chanK, frsBiType, biIdx, end_date, cur_ji = 1):
    max_k_num = 4
    if cur_ji>=4 or len(biIdx)==0: return biIdx
    idx = biIdx[len(biIdx)-1]
    k_data_dts = list(k_data.index)
    st_data = chanK['enddate'][idx]
    if st_data not in k_data_dts: return biIdx
    # 重构次级别线段的点到本级别的chanK中
    def refactorXd(biIdx, xdIdxc, chanK, chanKc, cur_ji):
        new_biIdx = []
        biIdxB = biIdx[len(biIdx)-1] if len(biIdx)>0 else 0
        for xdIdxcn in xdIdxc:
            for chanKidx in range(len(chanK.index))[biIdxB:]:
                if judge_day_bao(chanK, chanKidx, chanKc, xdIdxcn, cur_ji):
                    new_biIdx.append(chanKidx)
                    break
        return new_biIdx
    # 判断次级别日期是否被包含
    def judge_day_bao(chanK, chanKidx, chanKc, xdIdxcn, cur_ji):
        _end_date = chanK['enddate'][chanKidx]+datetime.timedelta(hours=15) if cur_ji==1 else chanK['enddate'][chanKidx]
        _start_date = chanK.index[chanKidx] if chanKidx==0\
        else chanK['enddate'][chanKidx-1]+datetime.timedelta(minutes=1)
        return _start_date<=chanKc.index[xdIdxcn]<=_end_date
    # cur_ji = 1 #当前级别
    #符合k线根数大于4根 1日级别, 2 30分钟, 3 5分钟, 4 一分钟
    if len(k_data_dts) - k_data_dts.index(st_data)>4:
        frequency = '30m' if cur_ji+1==2 else '5m' if cur_ji+1==3 else '1m'
        k_data_c = get_price(stock, st_data, end_date, frequency=frequency)
        chanKc = parse2ChanK(k_data_c, k_data_c.values)
        fenTypesc, fenIdxc = parse2ChanFen(chanKc)
        if len(fenTypesc)==0: return biIdx
        biIdxc, frsBiTypec = parse2ChanBi(fenTypesc, fenIdxc, chanKc)
        if len(biIdxc)==0: return biIdx
        xdIdxc, xdTypec = parse2Xianduan(biIdxc, chanKc)
        biIdxc = con2Cxianduan(stock, k_data_c, chanKc, frsBiTypec, biIdxc, end_date, cur_ji+1)
        if len(xdIdxc)==0: return biIdx
        # 连接线段位为上级别的bi
        lastBiType = frsBiType if len(biIdx)%2==0 else -frsBiType
        if len(biIdx)==0:
            return refactorXd(biIdx, xdIdxc, chanK, chanKc, cur_ji)
        lastbi = biIdx.pop()
        firstbic = xdIdxc.pop(0)
        # 同向连接
        if lastBiType == xdTypec:
            biIdx = biIdx + refactorXd(biIdx, xdIdxc, chanK, chanKc, cur_ji)
        # 逆向连接
        else:
#             print '开始逆向连接'
            _mid = [lastbi] if (lastBiType == -1 and chanK['low'][lastbi]<=chanKc['low'][firstbic])\
            or (lastBiType == 1 and chanK['high'][lastbi]>=chanKc['high'][firstbic]) else\
            [chanKidx for chanKidx in range(len(chanK.index))[biIdx[len(biIdx)-1]:]\
            if judge_day_bao(chanK, chanKidx, chanKc, firstbic, cur_ji)]
            biIdx = biIdx + [_mid[0]] + refactorXd(biIdx, xdIdxc, chanK, chanKc, cur_ji)
    return biIdx

全部回复

0/140

达人推荐

量化课程

    移动端课程