def cal_reverse_5days(self):
"""
每天开盘前计算 前5个交易日 各个基金的股票成交情况
不能自反向 ：若每只基金自身 前五天有卖出（买入）情况某只股票的情况 那么基金当天不能买入（卖出）
"""
day_period = 5
end_date = Date().get_trade_date_offset(self.today, -1)
beg_date = Date().get_trade_date_offset(end_date, -(day_period - 1))
trade_series = Date().get_trade_date_series(beg_date, end_date)
manager_info = MfcData().get_mfc_manage_info()
for i_col in range(len(manager_info.columns)):
person_name = manager_info.columns[i_col]
manager_fund = manager_info[person_name]
manager_fund = manager_fund.dropna()
fund_list = list(manager_fund.values)
for i_fund in range(len(fund_list)):
fund_name = fund_list[i_fund]
print(" Cal Fund Reverse 5 days ", fund_name, person_name)
result = pd.DataFrame([])
for i_date in range(len(trade_series)):
cur_date = trade_series[i_date]
data = MfcData().get_trade_statement(cur_date)
data = data.dropna(subset=['基金名称'])
data = data[['基金名称', '证券代码', '委托方向', '成交数量', '资产类别']]
data.columns = [
'FundName', 'StockCode', 'Direction', 'TradeNumber',
'Type'
]
data = data[data.FundName == fund_name]
data = data[data.Type == '股票资产']
data.StockCode = data.StockCode.map(
CodeFormat().stock_code_add_postfix)
data.Direction = data.Direction.map(lambda x: 2
if x == '卖出' else 1)
result_date = pd.DataFrame(data.Direction.values,
index=data.StockCode.values,
columns=[cur_date])
try:
result = pd.concat([result, result_date], axis=1)
except Exception as e:
pass
result = result.fillna(0)
result.index.name = 'CODE'
result = result.astype(np.int)
my_result = []
for i_row in range(len(result)):
vals_list = list(result.ix[i_row, trade_series])
vals_set = list(set(result.ix[i_row, trade_series]))
if 2 in vals_set:
append_row = [result.index[i_row], 2]
append_row.extend(vals_list)
my_result.append(append_row)
if 1 in vals_set:
append_row = [result.index[i_row], 1]
append_row.extend(vals_list)
my_result.append(append_row)
col = ['CODE', 'FLAG']
col.extend(trade_series)
my_result = pd.DataFrame(my_result, columns=col)
out_sub_path = os.path.join(self.data_path, person_name,
self.today, 'reverse_5days')
if not os.path.exists(out_sub_path):
os.makedirs(out_sub_path)
out_file = os.path.join(out_sub_path, fund_name + '.csv')
my_result.to_csv(out_file, index=None)
print(fund_name, len(result), len(my_result))
def cal_reverse_5days(today, project_path, out_path):
# 参数
###########################################################################
person_list = ['liuxin', 'yangchao', 'caolongjie', 'liuyang']
day_period = 5
# 日期参数
###########################################################################
before_trade_data = Date().get_trade_date_offset(today, -1)
trade_series = Date().get_trade_date_series(end_date=before_trade_data)
trade_series = pd.DataFrame(trade_series, index=trade_series)
trade_series = list(trade_series.index[-day_period:])
# 基金经理管理的基金
###########################################################################
fund = pd.read_excel(project_path + 'Manage_Fund_Name.xlsx',
encoding='gbk')
result = pd.DataFrame([])
# 开始计算每只基金的5日五日反向
#######################################################################################################
for i_col in range(len(person_list)):
person_name = person_list[i_col]
fund_val = fund.ix[:, person_name]
fund_val = fund_val.dropna()
fund_list = list(fund_val.values)
for i_fund in range(len(fund_list)):
fund_name = fund_list[i_fund]
print(" Cal Fund Reverse 5 days ", fund_name)
for i_date in range(len(trade_series)):
cur_date = trade_series[i_date]
data = MfcData().get_trade_statement(cur_date)
data = data.dropna(subset=['基金名称'])
data = data[['基金名称', '证券代码', '委托方向', '成交数量', '资产类别']]
data.columns = [
'FundName', 'StockCode', 'Direction', 'TradeNumber', 'Type'
]
data = data[data.FundName == fund_name]
data = data[data.Type == '股票资产']
data.StockCode = data.StockCode.map(stock_code_add_postfix)
data.Direction = data.Direction.map(lambda x: 2
if x == '卖出' else 1)
result_date = pd.DataFrame(data.Direction.values,
index=data.StockCode.values,
columns=[cur_date])
if i_date == 0:
result = result_date
else:
try:
result = pd.concat([result, result_date], axis=1)
except:
pass
result = result.fillna(0)
result.index.name = 'CODE'
result = result.astype(np.int)
my_result = []
for i_row in range(len(result)):
vals_list = list(result.ix[i_row, trade_series])
vals_set = list(set(result.ix[i_row, trade_series]))
if 2 in vals_set:
append_row = [result.index[i_row], 2]
append_row.extend(vals_list)
my_result.append(append_row)
if 1 in vals_set:
append_row = [result.index[i_row], 1]
append_row.extend(vals_list)
my_result.append(append_row)
col = ['CODE', 'FLAG']
col.extend(trade_series)
my_result = pd.DataFrame(my_result, columns=col)
out_sub_path = os.path.join(out_path, person_name, today,
'reverse_5days')
if not os.path.exists(out_sub_path):
os.makedirs(out_sub_path)
out_file = os.path.join(out_sub_path, fund_name + '.csv')
my_result.to_csv(out_file, index=None)
print(fund_name, len(result), len(my_result))