from jqdatasdk import * import pandas as pd import pymysql from sqlalchemy import create_engine import threading from datetime import datetime as dt from jqdatasdk.technical_analysis import * #原始版本 auth('18616891214', 'Ea?*7f68nD.dafcW34d!') # auth('18521506014', 'Abc123!@#') stocks = list(get_all_securities(['stock'], date=dt.today().strftime('%Y-%m-%d')).index) # stocks = stocks[0:200] def hlfx(stocks, engine_stock, engine_hlfx): for thd.stock in stocks: print(thd.stock) if ('stk%s_%s' % (thd.stock, fre)) in table_list: # 有历史数据 index_len = pd.read_sql_table('stk%s_%s' % (thd.stock, fre), con=engine_hlfx).iloc[-1, 0] startdate = pd.read_sql_table('stk%s_%s' % (thd.stock, fre), con=engine_hlfx).iloc[-1, 1] # thd.get_price = pd.read_sql_query( # 'select date,open,close,high,low,volume,money from `stk%s_%s`' % (thd.stock, fre), engine_stock) # thd.get_price = thd.get_price.loc[thd.get_price['date'] > startdate] thd.get_price = df.loc[thd.stock] thd.df_day = pd.read_sql_query( 'select date,open,close,high,low,volume,money,HL from `stk%s_%s`' % (thd.stock, fre), engine_hlfx) # 先处理去包含 for i in thd.get_price.index: # 不包含 if (thd.df_day.iloc[-1, 3] > thd.get_price.loc[i, 'high'] and thd.df_day.iloc[-1, 4] > thd.get_price.loc[i, 'low']) \ or (thd.df_day.iloc[-1, 3] < thd.get_price.loc[i, 'high'] and thd.df_day.iloc[-1, 4] < thd.get_price.loc[i, 'low']): thd.df_day = pd.concat([thd.df_day, thd.get_price.loc[[i]]], ignore_index=True) # 包含 else: # (new_df.iloc[-1,3]>=df_day.loc[i,'high'] and new_df.iloc[-1,4]<= df_day.loc[i,'low']): # 左高，下降 if thd.df_day.iloc[-2, 3] > thd.df_day.iloc[-1, 3]: thd.df_day.iloc[-1, 3] = min(thd.df_day.iloc[-1, 3], thd.get_price.loc[i, 'high']) thd.df_day.iloc[-1, 4] = min(thd.df_day.iloc[-1, 4], thd.get_price.loc[i, 'low']) else: # 右高，上升 thd.df_day.iloc[-1, 3] = max(thd.df_day.iloc[-1, 3], thd.get_price.loc[i, 'high']) thd.df_day.iloc[-1, 4] = max(thd.df_day.iloc[-1, 4], thd.get_price.loc[i, 'low']) # 寻找顶底分型 if len(thd.df_day.index) > 2: x = len(thd.df_day.index)-1 m = x - 1 # 底 if ((thd.df_day.loc[x, 'high'] > thd.df_day.loc[x - 1, 'high']) and ( thd.df_day.loc[x - 2, 'high'] > thd.df_day.loc[x - 1, 'high'])): # if ((stk.df_day.loc[i-2, 'date'] != stk.fxdf.iloc[-1,0]) and (stk.df_day.loc[i-3,'date'] != stk.fxdf.iloc[-1,0]) and (stk.df_day.loc[i-1,'date'] != stk.fxdf.iloc[-1,0])): # stk.fxdf = pd.concat([stk.fxdf, stk.df_day.iloc[[i]]], ignore_index=True) thd.df_day.loc[x, 'HL'] = 'L*' while m: if thd.df_day.loc[m, 'HL'] == 'H': if (x - m) > 3: thd.df_day.loc[x, 'HL'] = 'L' # 此处可以获得MACD指标 # pre-macd_dif, pre-macd_dea, pre-macd_macd = MACD(thd.stock,check_date=thd.df_day.loc[m, 'datetime'], SHORT = 12, LONG = 26, MID = 9) # pass print(thd.stock, '$$$$$$$', '\n', thd.df_day.loc[x, 'date'], '\n', '笔形成————买买买买买！！') results.append(thd.stock) print('222') # break elif (thd.df_day.loc[m, 'HL'] == 'L'): if thd.df_day.loc[x - 1, 'low'] < thd.df_day.loc[m - 1, 'low']: # 前一个为底，且中间存在不包含 or 更低的底 thd.df_day.loc[x, 'HL'] = 'L' x_macd_dif, x_macd_dea, x_macd_macd = MACD(thd.stock, check_date=thd.df_day.loc[x, 'date'], SHORT=12, LONG=26, MID=9, unit=fre) m_macd_dif, m_macd_dea, m_macd_macd = MACD(thd.stock, check_date=thd.df_day.loc[m, 'date'], SHORT=12, LONG=26, MID=9, unit=fre) if x_macd_dif[thd.stock] > m_macd_dif[thd.stock]: # pass # print(thd.df_day.loc[m, 'date'], thd.df_day.loc[m, 'low'], # m_macd_dif[thd.stock]) # print(thd.df_day.loc[x, 'date'], thd.df_day.loc[x, 'low'], # x_macd_dif[thd.stock]) print(thd.stock, '$$$$$$$', '\n', thd.df_day.loc[x, 'date'], 'MACD背驰————买吗？！') results.append(thd.stock) print('333') break else: # 底更低但没有背驰 break m = m - 1 if m == 0: # 第一个底 thd.df_day.loc[x, 'HL'] = 'L' results.append(thd.stock) print('444') # 顶 elif ((thd.df_day.loc[x, 'high'] < thd.df_day.loc[x - 1, 'high']) and ( thd.df_day.loc[x - 2, 'high'] < thd.df_day.loc[x - 1, 'high'])): # if ((stk.df_day.loc[i-2, 'date'] != stk.fxdf.iloc[-1,0]) and (stk.df_day.loc[i-3,'date'] != stk.fxdf.iloc[-1,0]) and (stk.df_day.loc[i-1,'date'] != stk.fxdf.iloc[-1,0])): # stk.fxdf = pd.concat([stk.fxdf, stk.df_day.iloc[[i]]], ignore_index=True) thd.df_day.loc[x, 'HL'] = 'H*' while m: if thd.df_day.loc[m, 'HL'] == 'L': if x - m > 3: thd.df_day.loc[x, 'HL'] = 'H' print(thd.stock, '!!!!!!!', '\n', thd.df_day.loc[x, 'date'], '笔形成————卖卖卖卖卖卖卖!') # pass results_short.append(thd.stock) if thd.stock in results: results.remove(thd.stock) # break elif (thd.df_day.loc[m, 'HL'] == 'H'): if thd.df_day.loc[x - 1, 'high'] > thd.df_day.loc[m - 1, 'high']: # 前一个为顶，且中间存在不包含 or 更高的顶 thd.df_day.loc[x, 'HL'] = 'H' x_macd_dif, x_macd_dea, x_macd_macd = MACD(thd.stock, check_date=thd.df_day.loc[x, 'date'], SHORT=12, LONG=26, MID=9, unit=fre) m_macd_dif, m_macd_dea, m_macd_macd = MACD(thd.stock, check_date=thd.df_day.loc[m, 'date'], SHORT=12, LONG=26, MID=9, unit=fre) if x_macd_dif[thd.stock] < m_macd_dif[thd.stock]: # pass print(thd.stock, '/\/\/\/\/\/\/', '顶背离了！！！！') results_short.append(thd.stock) if thd.stock in results: results.remove(thd.stock) break break m = m - 1 if m == 0: thd.df_day.loc[x, 'HL'] = 'H' results_short.append(thd.stock) if thd.stock in results: results.remove(thd.stock) else: thd.df_day.loc[x, 'HL'] = '-' # 更新数据库 # 可以使用normalize_code(code) 方法 改变代码格式 # thd.df_day[index_len + 1:].to_sql('stk%s_%s' % (thd.stock, fre), con=engine_hlfx, index=True, if_exists='append') else: # 没有历史数据表 thd.df_day = pd.DataFrame(columns=('date', 'open', 'close', 'high', 'low', 'volume', 'money', 'HL')) thd.get_price = pd.read_sql_query( 'select date,open,close,high,low,volume,money from `stk%s_%s`' % (thd.stock, fre), engine_stock) # 先处理去包含 for i in thd.get_price.index: if i == 0 or i == 1: thd.df_day = pd.concat([thd.df_day, thd.get_price.iloc[[i]]], ignore_index=True) # 不包含 elif (thd.df_day.iloc[-1, 3] > thd.get_price.loc[i, 'high'] and thd.df_day.iloc[-1, 4] > thd.get_price.loc[i, 'low']) \ or (thd.df_day.iloc[-1, 3] < thd.get_price.loc[i, 'high'] and thd.df_day.iloc[-1, 4] < thd.get_price.loc[i, 'low']): thd.df_day = pd.concat([thd.df_day, thd.get_price.loc[[i]]], ignore_index=True) # 包含 else: # 左高，下降 if thd.df_day.iloc[-2, 3] > thd.df_day.iloc[-1, 3]: thd.df_day.iloc[-1, 3] = min(thd.df_day.iloc[-1, 3], thd.get_price.loc[i, 'high']) thd.df_day.iloc[-1, 4] = min(thd.df_day.iloc[-1, 4], thd.get_price.loc[i, 'low']) else: # 右高，上升 thd.df_day.iloc[-1, 3] = max(thd.df_day.iloc[-1, 3], thd.get_price.loc[i, 'high']) thd.df_day.iloc[-1, 4] = max(thd.df_day.iloc[-1, 4], thd.get_price.loc[i, 'low']) if len(thd.df_day.index) > 2: # 寻找顶底分型 x = len(thd.df_day.index)-1 m = x - 1 # 底 if ((thd.df_day.loc[x, 'high'] > thd.df_day.loc[x - 1, 'high']) and ( thd.df_day.loc[x - 2, 'high'] > thd.df_day.loc[x - 1, 'high'])): # if ((stk.df_day.loc[i-2, 'date'] != stk.fxdf.iloc[-1,0]) and (stk.df_day.loc[i-3,'date'] != stk.fxdf.iloc[-1,0]) and (stk.df_day.loc[i-1,'date'] != stk.fxdf.iloc[-1,0])): # stk.fxdf = pd.concat([stk.fxdf, stk.df_day.iloc[[i]]], ignore_index=True) thd.df_day.loc[x, 'HL'] = 'L*' while m: if thd.df_day.loc[m, 'HL'] == 'H': if (x - m) > 3: thd.df_day.loc[x, 'HL'] = 'L' # 此处可以获得MACD指标 # pre-macd_dif, pre-macd_dea, pre-macd_macd = MACD(thd.stock,check_date=thd.df_day.loc[m, 'datetime'], SHORT = 12, LONG = 26, MID = 9) # pass print(thd.stock, '$$$$$$$', '\n', thd.df_day.loc[x, 'date'], '买买买买买！！') results.append(thd.stock) # break elif (thd.df_day.loc[m, 'HL'] == 'L'): if thd.df_day.loc[x - 1, 'low'] < thd.df_day.loc[m - 1, 'low']: # 前一个为底，且中间存在不包含 or 更低的底 thd.df_day.loc[x, 'HL'] = 'L' x_macd_dif, x_macd_dea, x_macd_macd = MACD(thd.stock, check_date=thd.df_day.loc[x, 'date'], SHORT=12, LONG=26, MID=9, unit=fre) m_macd_dif, m_macd_dea, m_macd_macd = MACD(thd.stock, check_date=thd.df_day.loc[m, 'date'], SHORT=12, LONG=26, MID=9, unit=fre) if x_macd_dif[thd.stock] > m_macd_dif[thd.stock]: # pass # print(thd.df_day.loc[m, 'date'], thd.df_day.loc[m, 'low'], # m_macd_dif[thd.stock]) # print(thd.df_day.loc[x, 'date'], thd.df_day.loc[x, 'low'], # x_macd_dif[thd.stock]) print(thd.stock, '$$$$$$$', '\n', thd.df_day.loc[x, 'date'], 'MACD背驰————买吗？！') results.append(thd.stock) break else: break m = m - 1 if m == 0: thd.df_day.loc[x, 'HL'] = 'L' results.append(thd.stock) # 顶 elif ((thd.df_day.loc[x, 'high'] < thd.df_day.loc[x - 1, 'high']) and ( thd.df_day.loc[x - 2, 'high'] < thd.df_day.loc[x - 1, 'high'])): # if ((stk.df_day.loc[i-2, 'date'] != stk.fxdf.iloc[-1,0]) and (stk.df_day.loc[i-3,'date'] != stk.fxdf.iloc[-1,0]) and (stk.df_day.loc[i-1,'date'] != stk.fxdf.iloc[-1,0])): # stk.fxdf = pd.concat([stk.fxdf, stk.df_day.iloc[[i]]], ignore_index=True) thd.df_day.loc[x, 'HL'] = 'H*' while m: if thd.df_day.loc[m, 'HL'] == 'L': if x - m > 3: thd.df_day.loc[x, 'HL'] = 'H' print(thd.stock, '!!!!!!!', '\n', '卖卖卖卖卖卖卖!') # pass results.remove(thd.stock) # break elif (thd.df_day.loc[m, 'HL'] == 'H'): if thd.df_day.loc[x - 1, 'high'] > thd.df_day.loc[m - 1, 'high']: # 前一个为顶，且中间存在不包含 or 更高的顶 thd.df_day.loc[x, 'HL'] = 'H' x_macd_dif, x_macd_dea, x_macd_macd = MACD(thd.stock, check_date=thd.df_day.loc[x, 'date'], SHORT=12, LONG=26, MID=9, unit=fre) m_macd_dif, m_macd_dea, m_macd_macd = MACD(thd.stock, check_date=thd.df_day.loc[m, 'date'], SHORT=12, LONG=26, MID=9, unit=fre) if x_macd_dif[thd.stock] < m_macd_dif[ thd.stock]: # pass print(thd.stock, '/\/\/\/\/\/\/', '顶背离了！！！！') results.remove(thd.stock) break break m = m - 1 if m == 0: thd.df_day.loc[x, 'HL'] = 'H' results.remove(thd.stock) else: thd.df_day.loc[x, 'HL'] = '-' # print(thd.df_day[-20:]) # 更新数据库 # thd.df_day.to_sql('stk%s_%s' % (thd.stock, fre), con=engine_hlfx, index=True, if_exists='append') start = dt.now() while True: now_date = dt.now() date_morning_begin = now_date.replace(hour=9, minute=25, second=0) date_morning_end = now_date.replace(hour=11, minute=31, second=0) date_afternooe_begin = now_date.replace(hour=13, minute=0, second=0) date_afternooe_end = now_date.replace(hour=15, minute=0, second=0) # print(now_date,date_morning_begin,date_morning_end,date_afternooe_begin,date_afternooe_end) # if date_morning_begin < now_date < date_afternooe_end: if True: for fre in ['1d']: start = dt.now() print(fre) # 连接数据库 db = pymysql.connect(host='localhost', user='root', port=3307, password='r6kEwqWU9!v3', database='hlfx') cursor = db.cursor() cursor.execute("show tables like '%%%s%%' " % fre) table_list = [tuple[0] for tuple in cursor.fetchall()] print('取得 table_list %s' % fre) db_pool = pymysql.connect(host='localhost', user='root', port=3307, password='r6kEwqWU9!v3', database='hlfx_pool') cursor_pool = db_pool.cursor() stk = locals() thd = threading.local() # 进程准备 step = 400 thread_list = [] engine_stock = [] engine_hlfx = [] times_engine = 0 engine_hlfx_pool = create_engine('mysql+pymysql://root:r6kEwqWU9!v3@localhost:3307/hlfx_pool?charset=utf8') # 获得hlfx_pool池子 # results = pd.read_sql_query( # 'select value from `%s`' % fre, engine_hlfx_pool) # for i in range(0, len(results)): # print(len(results.iloc[i, 0].split(","))) results = pd.read_sql_query( 'select value from `%s`' % fre, engine_hlfx_pool).iloc[-1, 0].split(",") results_short = [] print('数据库读取', len(results)) df = get_bars(stocks, count=5, unit=fre, fields=['date', 'open', 'close', 'high', 'low', 'volume', 'money'], include_now=True, df=True) print(df, type(df)) print(df.loc['603566.XSHG']) print(dt.now(), 'get_bars 成功') exit() for i in range(0, len(stocks), step): engine_stock.append(create_engine('mysql+pymysql://root:r6kEwqWU9!v3@localhost:3307/stocks?charset=utf8')) engine_hlfx.append(create_engine('mysql+pymysql://root:r6kEwqWU9!v3@localhost:3307/hlfx?charset=utf8')) thread = threading.Thread(target=hlfx, args=(stocks[i:i + step], engine_stock[times_engine], engine_hlfx[times_engine])) times_engine = times_engine + 1 thread.start() thread_list.append(thread) for thread in thread_list: thread.join() db.close() time = dt.now().strftime('%Y-%m-%d %H:%M:%S') results_list =','.join(set(results)) print(set(results)) sql = "INSERT INTO %s (date,value) VALUES('%s','%s')" % (fre, dt.now().strftime('%Y-%m-%d %H:%M:%S'), results_list) cursor_pool.execute(sql) db_pool.commit() print(fre, '\n', '做多:', len(set(results)), set(results)) print('做空', len(set(results_short)), set(results_short)) end= dt.now() print('总时长：', (end - start).seconds) elif now_date>date_afternooe_end: pass # print("HLFX_收盘了",now_date) # break