def GetTodayRealTimePrice( id, needNew):
if needNew:
g_pdReport = ts.get_day_all()
else:
if g_pdReport == None:
g_pdReport = ts.get_day_all()
return None
def GetTodayRealTimePrice( id, needNew = False):
global g_pdReport
if needNew:
g_pdReport = ts.get_day_all("2018-03-20")
else:
if g_pdReport.empty:
g_pdReport = ts.get_today_all()
try:
ret = g_pdReport.loc[g_pdReport['code'] == id]
except:
return None
return {"open":ret["open"].values[0],"high":ret["high"].values[0], "low":ret["low"].values[0], "price":ret["trade"].values[0], "volume":ret["volume"].values[0]*100, "preclose":ret["settlement"].values[0]}
sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options))
KTF.set_session(sess)
data_dir = './data/netease/hist_ma/'
code = 600082
#code = 600169
#code = 600815
#code = 600036
#code = 300104
#code = 600201
#code = '002608'
#stock_codes = [600082, 600169, 600036, 600201, 600400, 600448, 600536, 600339, 600103, 600166]
df = ts.get_day_all()
if df is None or df.empty:
print('failed to get codes')
exit(-1)
stock_codes = df['code'].tolist()
pick_index = -1
snapshot_dir = './snapshots_pick/pick_cnn_netease_all_clf_hkhsi'
if not os.path.exists(snapshot_dir):
print('snapshot dir not exists:%s' % snapshot_dir)
exit(-1)
timestr = str(datetime.now()).replace(' ', '@').replace(':', '_')
def update_db_day_all(date=''):
df = ts.get_day_all(date)
df.to_csv('%s/%s.csv' % (day_all_path, date), encoding='utf-8')
def update_all_today_online(self,
last_date=None,
code_set=None,
autype='D'):
'''
update all stock data today, using ts.get_day_all(date).
NOTE: 1. not hfq data, only use during 9:30am - 15:00pm for speed.
must use update_stock_data() to overwrite.
2. if one stock ting2 pai2 in last_date, this function will not conclude this stock.
3. the reference day to compute adj is *before* and *the nearest to* param "last_date"
details see the building of code_set
---
return None
'''
today_date = datetime.datetime.now().strftime('%Y-%m-%d')
today_all_df = ts.get_day_all(today_date)
lastday_date = last_date
if lastday_date > today_date:
raise OSError(
'"last_date" is later than today. please check again.')
if lastday_date is None:
lastday_date = today_date
lastday_date, lastday_all_df = self._get_nearest_trade_date_and_df_before(
lastday_date)
# 根据昨天的数据和本地csv数据，计算出后复权的权重比例
# 再计算今天对应的后复权股价，写进csv文件里
today_all_df = today_all_df[today_all_df['volume'] > 10]
lastday_all_df = lastday_all_df[lastday_all_df['volume'] > 10]
new_code_set = (set(today_all_df['code']) & set(lastday_all_df['code'])
) if code_set is None else set(code_set)
for code in new_code_set:
lastday_code_df = lastday_all_df[lastday_all_df['code'] == code]
old_df = self.get_stock_data(code, way='fs')
if lastday_date not in set(old_df['date']):
print(
'Warning: ', code,
' the last trade date in filesystem is not match with it in web.'
)
print(
' jump to next stock code. if all code makes warning, change the param "last_date".'
)
else:
old_last_series = old_df[old_df['date'] == lastday_date]
adj = self._comput_adj_with(old_last_series, lastday_code_df)
new_df = today_all_df[today_all_df['code'] == code]
new_df['date'] = today_date
new_df = new_df[[
'date', 'open', 'price', 'high', 'low', 'volume', 'amount',
'turnover'
]]
new_df.columns = [
'date', 'open', 'close', 'high', 'low', 'volume', 'amount',
'tor'
]
new_df['open'] = round(new_df['open'] * adj, 2)
new_df['close'] = round(new_df['close'] * adj, 2)
new_df['high'] = round(new_df['high'] * adj, 2)
new_df['low'] = round(new_df['low'] * adj, 2)
old_df = old_df[old_df['date'] != today_date]
df = pd.concat([old_df, new_df])
df.to_csv(FS_PATH_OL + '/' + code + '.csv', index=False)
val = int(val)
if val > 0:
color = 'red'
elif val < 0:
color = 'green'
else:
color = 'black'
#color = 'red' if val > 0 elif < 0 'green' else 'black'
return 'color: %s' % color
if __name__ == "__main__":
#file = sys.argv[1]
#stock_list = get_stock_list(file)
df = ts.get_day_all()
my_chioce = df[(df.pe <= 30) & (df.interval3 <= -15) & (df.open >= 50 ) & (df.pe != 0) & (df.price < df.avgprice)]
stock_list = [ code[0] for code in my_chioce[['code']].values ]
stock_basics=get_stock_basics()
num4days = 90
cycle_time = 30
day_list = [i for i in range(cycle_time)]
#stock_code = sys.argv[1]
now = datetime.date.today()
end_day = get_end_day(now)
start_day = get_start_day(now,num4days)
cpus = multiprocessing.cpu_count()
import sys
import hdfs
from hdfs.client import Client
import tushare as ts
import pandas
import csv
test_trade_data = ts.get_day_all('2019-03-29')
client = Client("http://172.23.253.80:50070/",
root="/",
timeout=10000,
session=False)
def upload(hdfs_path='', local_file_path='', cleanup=True):
return client.upload(hdfs_path, local_file_path, cleanup)
def delete_hdfs_file(hdfs_path):
return client.delete(hdfs_path)
def read_csv_from_hdfs(hdfs_path):
with client.read(hdfs_path, encoding='utf-8') as reader:
csv = pandas.read_csv(reader)
return csv
if __name__ == '__main__':
def update_all(self,
start_date,
end_date=None,
code_set=None,
processors=10):
if end_date is None:
end_date = self.utils.get_logdate('today')
self.utils.logger.info("开始日期 %s, 结束日期 %s" % (start_date, end_date))
# 每天的code基本面数据，写入基本面表
date_list = self.utils.get_date_list(start_date, end_date)
for date in date_list:
try:
self.utils.logger.info('日期 %s' % date)
cur_code_df = ts.get_day_all(date)
if cur_code_df is not None and len(cur_code_df) > 0:
self.utils.update_daily_df(date, cur_code_df, 'basic',
'dayall')
except HTTPError as err:
if err.msg == 'Not Found':
self.utils.logger.warning(
'date: %s, ts.get_day_all return no data.' % date)
else:
raise err
self.utils.logger.info("基本面数据写入完成...")
# 更新天级k线hfq数据，写入周级表
def process_one_code_d(code, start_date, end_date):
utils = Utils()
utils.logger.info("处理%s日k线数据..." % code)
df = ts.get_k_data(code,
start_date,
end_date,
autype='hfq',
ktype='D')
# 计算复权因子，写入基本面表
records = utils.find_code_date('basic', 'dayall', code, start_date,
end_date)
if len(df) <= 0:
return
df['tor'] = df['date'].apply(lambda x: records[x]['turnover']
if x in records.keys() else None)
utils.update_kdata_df(code, df, 'kdata', 'D')
for date in df.date:
if date not in records.keys():
utils.logger.warning(
"股票代码%s 在%s 没有基本面数据，且存在日k线数据，待人工确认..." % (code, date))
return
nfq_open = records[date]['open']
hfq_open = df[df.date == date].open.values[0]
aufactor = hfq_open / nfq_open
docs.append({'code': code, 'date': date, 'aufactor': aufactor})
utils.update_doc(docs, 'basic', 'dayall')
if code_set is None:
code_set = self.utils.get_all_code_list(start_date, end_date)
docs = []
pool = ThreadPool(processors)
for code in code_set:
pool.apply_async(process_one_code_d,
args=(code, start_date, end_date))
pool.close()
pool.join()
self.utils.logger.info("日k线数据写入完成，后复权因子写入基本面表完成...")
# 更新周级k线hfq数据，写入周级表
def process_one_code_w(code, start_date, end_date):
utils = Utils()
utils.logger.info("处理%s周k线数据..." % code)
df = ts.get_k_data(code,
start_date,
end_date,
autype='hfq',
ktype='W')
if len(df) <= 0:
return
records = utils.find_code_date('basic', 'dayall', code, start_date,
end_date)
daily_date_list = sorted(records.keys())
weekly_date_set = set(df['date'])
if len(weekly_date_set - set(daily_date_list)) > 0:
utils.logger.warning(
'在股票代码%s中，以下日期存在周k线数据但不存在基本面对应数据，待人工确认...%s' %
(code, weekly_date_set - set(daily_date_list)))
tor_dic = {}
cur_tor = 0
for date in daily_date_list:
cur_tor += records[date]['turnover']
if date in weekly_date_set:
tor_dic[date] = cur_tor
cur_tor = 0
df['tor'] = df['date'].apply(lambda x: tor_dic[x]
if x in tor_dic.keys() else 0)
df = df[[
'date', 'open', 'close', 'high', 'low', 'volume', 'tor', 'code'
]]
utils.update_kdata_df(code, df, 'kdata', 'W')
if code_set is None:
code_set = self.utils.get_all_code_list(start_date, end_date)
pool = ThreadPool(processors)
for code in code_set:
pool.apply_async(process_one_code_w,
args=(code, start_date, end_date))
pool.close()
pool.join()
self.utils.logger.info("周k线数据写入完成...")
# 更新分钟k线数据 （可能是多个分钟级k线list）
def process_one_code_Mn(code, start_date, end_date, ktype):
utils = Utils()
utils.logger.info("处理%s %s分钟k线数据..." % (code, ktype))
df = ts.get_k_data(code,
start_date,
end_date,
autype='hfq',
ktype=ktype)
if len(df) == 0:
return
df = df[(df['date'] >= start_date)
& (df['date'] <= (end_date + '-'))]
if len(df) == 0:
return
# 读每天的复权因子字段，使用复权因子更新价格
records = utils.find_code_date('basic', 'dayall', code, start_date,
end_date)
#TODO: 如果没有，应该保持和前一天一样的aufactor
cur_date_set = set(df['date'].apply(lambda x: x[0:10]))
basic_date_set = set(records.keys())
if len(cur_date_set - basic_date_set) > 0:
utils.logger.warning(
'在股票代码%s中，以下日期存在k线数据但不存在基本面对应数据，待人工确认... %s' %
(code, cur_date_set - basic_date_set))
df['aufactor'] = df['date'].apply(lambda x: records[x[0:10]][
'aufactor'] if x[0:10] in records.keys() else 1)
for col in ('open', 'close', 'high', 'low'):
df[col] = round(df[col] * df['aufactor'], 3)
# 使用全天交易量和换手率更新换手率
for _, r in records.items():
if r['volume'] > 0:
r['tor_p_v'] = r['turnover'] / r['volume']
else:
r['tor_p_v'] = int(ktype) / 240.0
df['tor'] = df.apply(
lambda x: records[x['date'][0:10]]['tor_p_v'] * x['volume']
if x['date'][0:10] in records.keys() else 0,
axis=1)
df = df[[
'date', 'open', 'close', 'high', 'low', 'volume', 'tor', 'code'
]]
utils.update_kdata_df(code, df, 'kdata', 'M' + ktype)
if code_set is None:
code_set = self.utils.get_all_code_list(start_date, end_date)
for ktype in ('60', ):
pool = ThreadPool(processors)
for code in code_set:
pool.apply_async(process_one_code_Mn,
args=(code, start_date, end_date, ktype))
pool.close()
pool.join()
self.utils.logger.info("%s分钟k线数据写入完成..." % ktype)
print("请先调用getCodeHist()方法获得code历史价格")
sys.exit()
RSI6 = talib.RSI(self.codeHistPrice["close"], timeperiod=6)
RSI12 = talib.RSI(self.codeHistPrice["close"], timeperiod=12)
RSI24 = talib.RSI(self.codeHistPrice["close"], timeperiod=24)
# def toCsv(self):
if __name__ == "__main__":
# 时间为收盘时间
Stime = "2020-08-13" #选择上一个交易日，因为使用实时预测是将今天的实时价格进行拼接计算的，如果收盘了就用收盘价格进行拼接了
startClock = time.time()
nowClock = time.time()
today_all = ts.get_day_all()
choosePara = ChoosePara(endTime=Stime, allDataFrame=today_all)
fBuy1 = open("../BuyMACD接近交叉.txt", 'w')
fBuy2 = open("../BuyMACD梯度为0.txt", 'w')
jd = tqdm.tqdm(total=len(today_all["code"]))
print(len(today_all["code"]), len(today_all["name"]))
for index in range(len(today_all["code"])):
code, name = today_all["code"][index], today_all["name"][index]
if ((name[:2] != "ST") & (name[:3] != "*ST")):
# if choosePara.getSingleCodeRealTimeByAPI(code):
# if choosePara.getSingleCodeHistByAPI(code):
if choosePara.getSingleCodeSQLByAPI(code):
flag = choosePara.BSdic[choosePara.getCodeMACD()]
#!/bin/python
# -*- coding:utf-8 -*-
import time
import price as prc
import tushare as ts
import storage as sv
all_stocks = ts.get_day_all()
print("get data done")
sv.saveData(all_stocks)
print(all_stocks)
# all_stocks=pd.read_json(sv.readData(), convert_axes=False, convert_dates=False)
today = time.strftime('%Y-%m-%d', time.localtime(time.time()))
start = time.time()
for index in all_stocks.index:
stock_code = all_stocks.code[index]
if all_stocks.abvalues[index] > 50:
continue
if not prc.istradeable(all_stocks.high[index], all_stocks.low[index]):
continue
if all_stocks.volratio[index] < 1.5:
continue
if not all_stocks.preprice[index] < all_stocks.price[
index] < all_stocks.preprice[index] * 1.05:
continue
# -*- coding:utf-8 -*-
__author__ = 'Administrator'
#coding: utf-8
def run():
print("hello")
if __name__ == "__main__":
run()
import tushare as ts
hq = ts.get_day_all()
shares = ts.get_stock_basics()
hq = hq.set_index('code')
basics = shares[['resevedPerShare', 'esp', 'timeToMarket']]
df = hq.merge(basics, left_index=True, right_index=True)
print df
#开发目标
# 1.取每天股票数据，按规则取前30条
# 2.市值递增排序（动态、静态）
# 3.市盈率递增排序
# 4.股价递增排序
# 5.近五个交易日涨幅排序
import tushare as ts
from pandas import DataFrame
#获取当日行情数据
df = ts.get_day_all('2019-04-23')[[
'code', 'name', 'pe', 'floats', 'fvalues', 'industry', 'change', 'price',
'p_change'
]]
#对股票进行条件排序
dfSort = df.sort_values(by=['pe', 'fvalues', 'price'],
ascending=[True, True, True])
#取前n行数据，并重置索引,删除原索引
dfData = dfSort[1:100].reset_index(drop=True)
lst = [] #新股列表
listBanstock = [] #st板块列表
listCreate = [] #创业板
#遍历dfsort中数据，获取市值为0的新股数据，并对st与*st股的索引，并对相应数据进行打标签
for index, row in dfData.iterrows():
if row["fvalues"] == 0:
lst.append(index)
elif row["name"].find('ST') >= 0:
listBanstock.append(index)
import tushare as ts
import pandas as pd
import time
import time, datetime
import calendar
a=ts.get_day_all('2017-10-20')
b=ts.get_day_all('2018-07-06')
print(a.columns)
# print(b)