def update_factor_in_category(con, category, factors, cur_date,
code_date_series):
# 解析
basic_data_info, factor_list = resolve_factors(factors)
print(basic_data_info)
# 获取该因子类中因子字段及它们的排列情况
fields = si.get_table_fields(con, category)
print(fields)
codes = si.get_all_codes(con)
data_list = [] # 存储所有因子数据
for code in codes:
# 根据收集的信息从数据库中去获取基础数据
for table in basic_data_info:
dict = {}
for field in basic_data_info[table]:
dict[field] = si.get_data_series(con, table, field, code,
code_date_series[code])
# 更新基础数据
basic_data_info[table] = dict
code_list = [code, cur_date]
row = {}
# 计算因子,通过反射找到方法并传入参数计算
# 注意之后要进行数据对齐，为了避免不同时间窗口下计算的因子数据结果长度不同，只取最后一天数据
for factor in factor_list:
func = getattr(module_dict[category], factor["name"])
param_list = [] # 存放参数列表
# 获取参数
for table in factor["params0"]:
for field in factor["params0"][table]:
param_list.append(basic_data_info[table][field])
for val in factor["params1"]:
param_list.append(val)
ret_data = func(param_list) # 调用函数计算因子
if len(ret_data) > 0:
data = ret_data[len(ret_data) - 1] # 只取最后一天（当日的数据）
row[factor["field"]] = float("%.5f" % data)
if row: # 字典对象不为空
# 重新排序使得结果与表中字段顺序一致
for field in fields:
code_list.append(row[field])
print(code_list)
data_list.append(code_list) # 添加该股票下的因子数据
si.insert_factor_data(con, category, data_list) # 批量更新当天的因子数据
def add_factor(con, category, factor, code_date_series):
# 解析因子
basic_data_info, factor_dict, factor_filed = resolve_factor(factor)
print(basic_data_info)
# 增加字段
si.add_factor_column(con, category, factor_filed)
codes = si.get_all_codes(con)
values = []
for code in codes:
# 根据收集的信息从数据库中去获取基础数据
for table in basic_data_info:
dict = {}
for field in basic_data_info[table]:
dict[field] = si.get_data_series(con, table, field, code,
code_date_series[code])
# 更新基础数据
basic_data_info[table] = dict
func = getattr(module_dict[category],
factor_dict["name"]) # 通过反射拿到相应的函数对象
param_list = []
# 向参数列表中计算因子要用到的基本数据
for table in factor_dict["params0"]:
for field in factor_dict["params0"][table]:
param_list.append(basic_data_info[table][field])
# 向参数列表中增加常数参数
for val in factor_dict["params1"]:
param_list.append(val)
ret_data = func(param_list) # 计算因子
values += si.get_column_tuple_list(code, code_date_series[code],
ret_data) # 增加该股票下的因子数据
# 更新数据
si.update_factor_column(con, category, factor_filed, values)
def update_emotion_factor(code_date_serise):
con = pymysql.connect(host="localhost",
user="******",
passwd="mysql",
db="stockvision")
table_name = "emotion_factor"
pool = si.get_all_codes(con)
logvol = Factor("logvol", [])
logvol_10 = Factor("logvol_10", [])
logvol_15 = Factor("logvol_15", [])
logvol_20 = Factor("logvol_20", [])
vol_change_rate = Factor("vol_change_rate", [])
vol_change_rate_5 = Factor("vol_change_rate_5", [])
vol_change_rate_10 = Factor("vol_change_rate_10", [])
vol_change_rate_20 = Factor("vol_change_rate_20", [])
vol_relative_ratio_0 = Factor("vol_relative_ratio_0", [])
vol_relative_ratio_1 = Factor("vol_relative_ratio_1", [])
ar = Factor("ar", [])
ar_14 = Factor("ar_14", [])
br = Factor("br", [])
br_14 = Factor("br_14", [])
obv_0 = Factor("obv_0", [])
obv_1 = Factor("obv_1", [])
vma = Factor("vma", [])
pvt_0 = Factor("pvt_0", [])
pvt_1 = Factor("pvt_1", [])
mfi = Factor("mfi", [])
# 更新因子数据
for code in pool:
# 获取基础数据
vol_series = si.get_data_series(con, "daily", "vol", code,
code_date_serise[code])
open_series = si.get_data_series(con, "daily", "open", code,
code_date_serise[code])
high_series = si.get_data_series(con, "daily", "high", code,
code_date_serise[code])
low_series = si.get_data_series(con, "daily", "high", code,
code_date_serise[code])
close_series = si.get_data_series(con, "daily", "close", code,
code_date_serise[code])
# 避免时间序列不够长无法进行计算/避免输入空值
if len(vol_series) > 0:
vol_change_rate_list = cal_vol_change_rate_n(vol_series, 1)
vol_change_rate.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
vol_change_rate_list))
obv_0_list = cal_obv(close_series, vol_series, 0)
obv_0.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
obv_0_list))
obv_1_list = cal_obv(close_series, vol_series, 1)
obv_1.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
obv_1_list))
pvt_0_list = cal_pvt(close_series, vol_series, 0)
pvt_0.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
pvt_0_list))
pvt_1_list = cal_pvt(close_series, vol_series, 1)
pvt_1.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
pvt_1_list))
vma_list = cal_vma_n(vol_series, 6)
vma.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
vma_list))
if len(vol_series) >= 5:
# 得到因子
logvol_list = cal_logvol_n(vol_series, 5)
vol_change_rate_5_list = cal_vol_change_rate_n(vol_series, 5)
vol_relative_ratio_0_list = cal_vol_relative_ratio(
vol_series, 5, 5, 0)
vol_relative_ratio_1_list = cal_vol_relative_ratio(
vol_series, 5, 5, 1)
logvol.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
logvol_list))
vol_change_rate_5.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
vol_change_rate_5_list))
vol_relative_ratio_0.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
vol_relative_ratio_0_list))
vol_relative_ratio_1.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
vol_relative_ratio_1_list))
mfi_list = cal_mfi(high_series, low_series, close_series,
vol_series, 5)
mfi.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
mfi_list))
if len(vol_series) >= 10:
logvol_10_list = cal_logvol_n(vol_series, 10)
vol_change_rate_10_list = cal_vol_change_rate_n(vol_series, 10)
logvol_10.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
logvol_10_list))
vol_change_rate_10.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
vol_change_rate_10_list))
if len(vol_series) >= 14:
ar_14_list = cal_ar_n(open_series, high_series, low_series, 14)
br_14_list = cal_br_n(close_series, high_series, low_series, 14)
ar_14.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
ar_14_list))
br_14.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
br_14_list))
if len(vol_series) >= 15:
logvol_15_list = cal_logvol_n(vol_series, 15)
logvol_15.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
logvol_15_list))
if len(vol_series) >= 20:
logvol_20_list = cal_logvol_n(vol_series, 20)
vol_change_rate_20_list = cal_vol_change_rate_n(vol_series, 20)
logvol_20.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
logvol_20_list))
vol_change_rate_20.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
vol_change_rate_20_list))
if len(vol_series) >= 26:
ar_list = cal_ar_n(open_series, high_series, low_series, 26)
br_list = cal_br_n(close_series, high_series, low_series, 26)
ar.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
ar_list))
br.append_value(
si.get_column_tuple_list(code, code_date_serise[code],
br_list))
# 更新数据库中因子
si.update_factor_column(con, table_name, logvol.get_name(),
logvol.get_value())
si.update_factor_column(con, table_name, logvol_10.get_name(),
logvol_10.get_value())
si.update_factor_column(con, table_name, logvol_15.get_name(),
logvol_15.get_value())
si.update_factor_column(con, table_name, logvol_20.get_name(),
logvol_20.get_value())
si.update_factor_column(con, table_name, vol_change_rate.get_name(),
vol_change_rate.get_value())
si.update_factor_column(con, table_name, vol_change_rate_5.get_name(),
vol_change_rate_5.get_value())
si.update_factor_column(con, table_name, vol_change_rate_10.get_name(),
vol_change_rate_10.get_value())
si.update_factor_column(con, table_name, vol_change_rate_20.get_name(),
vol_change_rate_20.get_value())
si.update_factor_column(con, table_name, ar.get_name(), ar.get_value())
si.update_factor_column(con, table_name, br.get_name(), br.get_value())
si.update_factor_column(con, table_name, ar_14.get_name(),
ar_14.get_value())
si.update_factor_column(con, table_name, br_14.get_name(),
br_14.get_value())
si.update_factor_column(con, table_name, vol_relative_ratio_0.get_name(),
vol_relative_ratio_0.get_value())
si.update_factor_column(con, table_name, vol_relative_ratio_1.get_name(),
vol_relative_ratio_1.get_value())
si.update_factor_column(con, table_name, obv_0.get_name(),
obv_0.get_value())
si.update_factor_column(con, table_name, obv_1.get_name(),
obv_1.get_value())
si.update_factor_column(con, table_name, vma.get_name(), vma.get_value())
si.update_factor_column(con, table_name, pvt_0.get_name(),
pvt_0.get_value())
si.update_factor_column(con, table_name, pvt_1.get_name(),
pvt_1.get_value())
si.update_factor_column(con, table_name, mfi.get_name(), mfi.get_value())
def update_motive_factors(code_date_series):
con = pymysql.connect(host="localhost", user="******", passwd="mysql", db="stockvision")
table_name = "motive_factor"
pool = si.get_all_codes(con)
macd = Factor("macd", [])
cci = Factor("cci", [])
roc = Factor("roc", [])
trix = Factor("trix", [])
rsi = Factor("rsi", [])
dpo = Factor("dpo", [])
ma_close_dis = Factor("ma_close_dis", [])
ma_ma_dis = Factor("ma_ma_dis", [])
bias = Factor("bias", [])
for code in pool:
close_series = si.get_data_series(con, "daily", "close", code, code_date_series[code])
high_series = si.get_data_series(con, "daily", "high", code, code_date_series[code])
low_series = si.get_data_series(con, "daily", "low", code, code_date_series[code])
# 保证输入不为空
if len(close_series) > 0:
macd_list = cal_macd(close_series)
print(macd_list, end=',')
print(code + " " + macd.get_name())
macd.append_value(si.get_column_tuple_list(code, code_date_series[code], macd_list))
if len(close_series) >= 5:
ma_close_dis_list = cal_ma_close_dis(close_series, 5)
print(macd_list, end=',')
print(code + " " + ma_close_dis.get_name())
ma_close_dis.append_value(si.get_column_tuple_list(code, code_date_series[code], ma_close_dis_list))
if len(close_series) >= 6:
bias_list = cal_bias_n(close_series, 6)
bias.append_value(si.get_column_tuple_list(code, code_date_series[code], bias_list))
if len(close_series) >= 10:
# dpo_list = cal_dpo(close_series, 10)
# dpo.append_value(si.get_column_tuple_list(code, code_date_series[code], dpo_list))
ma_ma_dis_list = cal_ma_ma_dis(close_series, 5, 10)
ma_ma_dis.append_value(si.get_column_tuple_list(code, code_date_series[code], ma_ma_dis_list))
if len(close_series) >= 12:
trix_list = cal_trix(close_series, 12)
trix.append_value(si.get_column_tuple_list(code, code_date_series[code], trix_list))
roc_list = cal_roc_n(close_series, 12)
roc.append_value(si.get_column_tuple_list(code, code_date_series[code], roc_list))
if len(close_series) >= 14:
rsi_list = cal_rsi(close_series, 14)
rsi.append_value(si.get_column_tuple_list(code, code_date_series[code], rsi_list))
cci_list = cal_cci_n(close_series, high_series, low_series, 14)
cci.append_value(si.get_column_tuple_list(code, code_date_series[code], cci_list))
si.update_factor_column(con, table_name, macd.get_name(), macd.get_value())
si.update_factor_column(con, table_name, cci.get_name(), cci.get_value())
si.update_factor_column(con, table_name, roc.get_name(), roc.get_value())
si.update_factor_column(con, table_name, trix.get_name(), trix.get_value())
si.update_factor_column(con, table_name, rsi.get_name(), rsi.get_value())
# si.update_factor_column(con, table_name, dpo.get_name(), dpo.get_value())
si.update_factor_column(con, table_name, ma_close_dis.get_name(), ma_close_dis.get_value())
si.update_factor_column(con, table_name, ma_ma_dis.get_name(), ma_ma_dis.get_value())
si.update_factor_column(con, table_name, bias.get_name(), bias.get_value())
def update_fluctuate_factor(code_date_series):
con = pymysql.connect(host="localhost", user="******", passwd="mysql", db="stockvision")
table_name = "fluctuate_factor"
pool = si.get_all_codes(con)
amplitude = Factor("amplitude", [])
amplitude_10 = Factor("amplitude_10", [])
price_efficiency = Factor("price_efficiency", [])
price_efficiency_10 = Factor("price_efficiency_10", [])
price_center = Factor("price_center", [])
avg_daily_amplitude = Factor("avg_daily_amplitude", [])
atr = Factor("atr", [])
boll_upperband = Factor("boll_upperband", [])
boll_middleband = Factor("boll_middleband", [])
boll_lowerband = Factor("boll_lowerband", [])
# 更新因子数据
for code in pool:
# 获取基础数据
close_series = si.get_data_series(con, "daily", "close", code, code_date_series[code])
high_series = si.get_data_series(con, "daily", "high", code, code_date_series[code])
low_series = si.get_data_series(con, "daily", "low", code, code_date_series[code])
# 避免时间序列不够长无法进行计算
if len(close_series) >= 5:
# 得到因子
amplitude_list = cal_amplitude_n(close_series, high_series, low_series, 5)
amplitude.append_value(si.get_column_tuple_list(code, code_date_series[code], amplitude_list))
price_efficiency_list = cal_price_efficiency_n(close_series, 5)
price_efficiency.append_value(si.get_column_tuple_list(code, code_date_series[code], price_efficiency_list))
price_center_list = cal_price_center_n_ma(close_series, 3, 5)
price_center.append_value(si.get_column_tuple_list(code, code_date_series[code], price_center_list))
avg_daily_amplitude_list = cal_avg_daily_amplitude_n(close_series, high_series, low_series, 5)
avg_daily_amplitude.append_value(
si.get_column_tuple_list(code, code_date_series[code], avg_daily_amplitude_list))
atr_list = cal_atr(high_series, low_series, close_series, 5)
atr.append_value(si.get_column_tuple_list(code, code_date_series[code], atr_list))
if len(close_series) >= 10:
amplitude_10_list = cal_amplitude_n(close_series, high_series, low_series, 10)
amplitude_10.append_value(si.get_column_tuple_list(code, code_date_series[code], amplitude_10_list))
price_efficiency_10_list = cal_price_efficiency_n(close_series, 10)
price_efficiency_10.append_value(
si.get_column_tuple_list(code, code_date_series[code], price_efficiency_10_list))
if len(close_series) >= 20:
boll_upperband_list, boll_middleband_list, boll_lowerband_list = cal_boll(close_series, 20)
boll_upperband.append_value(si.get_column_tuple_list(code, code_date_series[code], boll_upperband_list))
boll_middleband.append_value(si.get_column_tuple_list(code, code_date_series[code], boll_middleband_list))
boll_lowerband.append_value(si.get_column_tuple_list(code, code_date_series[code], boll_lowerband_list))
# 更新数据库中因子
si.update_factor_column(con, table_name, amplitude.get_name(), amplitude.get_value())
si.update_factor_column(con, table_name, amplitude_10.get_name(), amplitude_10.get_value())
si.update_factor_column(con, table_name, price_efficiency.get_name(), price_efficiency.get_value())
si.update_factor_column(con, table_name, price_efficiency_10.get_name(), price_efficiency_10.get_value())
si.update_factor_column(con, table_name, price_center.get_name(), price_center.get_value())
si.update_factor_column(con, table_name, avg_daily_amplitude.get_name(), avg_daily_amplitude.get_value())
si.update_factor_column(con, table_name, atr.get_name(), atr.get_value())
si.update_factor_column(con, table_name, boll_upperband.get_name(), boll_upperband.get_value())
si.update_factor_column(con, table_name, boll_middleband.get_name(), boll_middleband.get_value())
si.update_factor_column(con, table_name, boll_lowerband.get_name(), boll_lowerband.get_value())