def index(_code=None):
if _code is None or len(_code) < 8:
return "error"
"""获取某个指数的详情"""
_market = 1 if _code[:2] == "SH" else 0
_idx_code = _code[2:]
stock_model = StockModel()
_filter = {"code": _idx_code, "market": _market, "date": {"$gte": date.datetime_to_str(date.years_ago(10))}}
_idx_list = stock_model.stock_list(_filter=_filter, _sort=[("date", pymongo.ASCENDING)],
_fields={"date": 1, "name": 1, "pe_ttm": 1, "pb": 1, "dr": 1})
item_df = pd.DataFrame(list(_idx_list), columns=["date", "name", "pe_ttm", "pb", "dr"])
item_df = item_df.dropna(0)
_pe_series = 1 / item_df["pe_ttm"]
pe_ranks = _pe_series.rank(pct=True)
_pb_series = 1 / item_df["pb"]
pb_ranks = _pb_series.rank(pct=True)
return render_template('indexes/detail.mako', ctx={
"index_name": item_df["name"].values[0],
"index_code": _code,
"start_date": date.datetime_to_str(date.years_ago(years=3)),
"pe_max": "%.2f" % item_df["pe_ttm"].max(),
"pe_min": "%.2f" % item_df["pe_ttm"].min(),
"pe_mean": "%.2f" % item_df["pe_ttm"].mean(),
"pe_high": "%.2f" % item_df["pe_ttm"].quantile(0.75),
"pe_middle": "%.2f" % item_df["pe_ttm"].quantile(0.5),
"pe_low": "%.2f" % item_df["pe_ttm"].quantile(0.2),
"pe_value": "%.2f" % item_df["pe_ttm"].values[-1],
"pe_rate": "%.2f%%" % ((1 - pe_ranks.values[-1]) * 100),
"pb_max": "%.2f" % item_df["pb"].max(),
"pb_min": "%.2f" % item_df["pb"].min(),
"pb_mean": "%.2f" % item_df["pb"].mean(),
"pb_high": "%.2f" % item_df["pb"].quantile(0.75),
"pb_middle": "%.2f" % item_df["pb"].quantile(0.5),
"pb_low": "%.2f" % item_df["pb"].quantile(0.2),
"pb_value": "%.2f" % item_df["pb"].values[-1],
"pb_rate": "%.2f%%" % ((1 - pb_ranks.values[-1]) * 100),
"dr_max": "%.2f%%" % (item_df["dr"].max() * 100),
"dr_min": "%.2f%%" % (item_df["dr"].min() * 100),
"dr_mean": "%.2f%%" % (item_df["dr"].mean() * 100),
"dr_value": "%.2f%%" % (item_df["dr"].values[-1] * 100)
})
def index_dr_json(_code=None):
if _code is None or len(_code) < 8:
return "error"
_code = _code.upper()
is_all = request.args.get('all')
data = []
_market = 1 if _code[:2] == "SH" else 0
_idx_code = _code[2:]
stock_model = StockModel()
_filter = {"code": _idx_code, "market": _market}
if is_all is None or len(is_all) <= 0:
from_date = date.datetime_to_str(date.years_ago(10))
_filter["date"] = {"$gte": from_date}
_idx_list = stock_model.stock_list(_filter=_filter, _sort=[("date", pymongo.ASCENDING)],
_fields={"date": 1, "dr": 1, "close": 1})
item_df = pd.DataFrame(list(_idx_list), columns=["date", "dr", "close"])
item_df = item_df.dropna(0)
for idx, row in item_df.iterrows():
data.append([row["date"], float("%.2f" % (row["dr"] * 100)), float("%.2f" % row["close"])])
return json.dumps(data)
def plot1(years, money):
tax = 0.00025 # 买卖按照万分之2.5计算
for item in config.get("indexes"):
_market = item[0]
_idx_code = item[1]
_money_per = money
click.echo("开始计算 %d - %s 指数的回测策略" % (_market, _idx_code))
_start_date = date.datetime_to_str(date.years_ago(years))
last_day_list = date.last_days(int(_start_date), 1,
True) # 每月最后一个交易日集合
data = []
db_client = MongoDBClient(
config.get("db").get("mongodb"),
config.get("db").get("database"))
_idx_list = db_client.find_stock_list(_filter={
"code": _idx_code,
"market": _market
},
_sort=[("date",
pymongo.ASCENDING)],
_fields={
"date": 1,
"close": 1
})
item_df = pd.DataFrame(list(_idx_list), columns=["date", "close"])
item_df = item_df.dropna(0)
total_count = int(0) # 累计数量
for idx, item_date in enumerate(last_day_list):
last_df = item_df[item_df["date"] == str(item_date)]
if last_df is None or len(last_df) <= 0:
continue
item_point = last_df.ix[last_df.index.values[0]]["close"]
item_count = _money_per * (1 - tax) / item_point
total_count += item_count
data.append([item_date, "买入", item_point, item_count, total_count])
item = data.pop(-1)
item_count = total_count * item[2] * (1 - tax) / item[2]
data.append([item[0], "卖出", item[2], item_count, 0])
result_df = pd.DataFrame(data,
columns=["日期", "操作", "单价", "数量", "累计数量"])
result_df.to_csv(config.get("files").get("plots") %
(PROJECT_ROOT, "plot1", _market, _idx_code),
index=False,
mode="w",
header=True,
encoding='gb18030',
float_format="%.3f")
def dividend_rate_with(_date: int, _code: str, _price: float):
"""
求某只股票在指定时间点的三年平均股息率
:param _date:
:param _code:
:param _price:
:return:
"""
end = date.str_to_datetime(str(_date))
start = date.years_ago(years=3, from_date=end)
bonus_df = bonus_with(_code=_code).sort_values(['date'], ascending=True)
filter_df = bonus_df[(bonus_df["type"] == 1) & (
bonus_df["date"] >= int(date.datetime_to_str(start))) &
(bonus_df["date"] <= int(date.datetime_to_str(end)))]
total_money = 0.0
for idx, item in filter_df.iterrows():
# 每股分红的三年之和
total_money += item["money"] / 10
return total_money / (3 * _price)
def stock_pe_json(_code: str = None):
if _code is None or len(_code) < 8:
return "error"
_code = _code.upper()
is_all = request.args.get('all')
data = []
_market = 1 if _code[:2] == "SH" else 0
_idx_code = _code[2:]
stockModel = StockModel()
_filter = {"code": _idx_code, "market": _market}
if is_all is None or len(is_all) <= 0:
from_date = date.datetime_to_str(date.years_ago(10))
_filter["date"] = {"$gte": from_date}
_idx_list = stockModel.stock_list(_filter=_filter,
_sort=[("date", pymongo.ASCENDING)],
_fields={
"date": 1,
"pe_ttm": 1,
"close": 1
})
item_df = pd.DataFrame(list(_idx_list),
columns=["date", "pe_ttm", "close"])
item_df = item_df.dropna(0)
item_df = item_df.reset_index()
_pe_series = 1 / item_df["pe_ttm"]
pe_ranks = _pe_series.rank(pct=True)
_pe_list = []
for idx, row in item_df.iterrows():
_pe_list.append(row["pe_ttm"])
_pe_high = np.percentile(_pe_list, 75)
_pe_middle = np.percentile(_pe_list, 50)
_pe_low = np.percentile(_pe_list, 20)
_item_rank = float("%.2f" % ((1 - pe_ranks.values[idx]) * 100))
data.append([
row["date"],
float("%.2f" % row["pe_ttm"]),
float("%.2f" % _pe_high),
float("%.2f" % _pe_middle),
float("%.2f" % _pe_low), _item_rank,
float("%.2f" % row["close"])
])
return json.dumps(data)
def indexes():
"""显示指数估值信息"""
_lp_file = None
try:
_lp_file = open("%s/last_update" % PROJECT_ROOT, 'r')
_content = _lp_file.read()
last_date = _content.strip()
finally:
if _lp_file:
_lp_file.close()
_tmp_date = date.datetime_to_str(date.years_ago(years=3))
_tmp_date = _tmp_date if date.is_work_day(int(_tmp_date)) else str(date.next_day(int(_tmp_date)))
last_date = last_date if date.is_work_day(int(last_date)) else str(date.prev_day(int(last_date)))
return render_template('indexes/list.mako', ctx={
"last_date": last_date,
"start_date": _tmp_date
})
def home():
"""首页 """
return render_template('home.mako', ctx={
"title": "宽基PB走势图",
"start_date": date.datetime_to_str(date.years_ago(years=3))
})
def indexes_json():
"""显示指数估值信息"""
# '上证50', '沪深300', '中证500', '中证1000', '上证红利', '中证红利'，'中小板指', '创业扳指'
# '食品饮料'
idx_list = config.get("indexes")
stock_model = StockModel()
result = []
for idx, item in enumerate(idx_list):
index_obj = {"idx": idx + 1}
_market = 'SZ' if item[0] == 0 else 'SH'
item_detail = stock_model.stock_item(_filter={"code": item[1], "market": item[0]},
_sort=[("date", pymongo.DESCENDING)])
index_obj["name"] = "<a href='/index/%s%s' title='%s的估值信息'> %s </a>" % \
(_market, item_detail["code"], item_detail["name"], item_detail["name"])
index_obj["interest"] = "%.2f%%" % (1 / item_detail["pe_ttm"] * 100)
index_obj["dr"] = "%.2f%%" % (item_detail["dr"] * 100)
index_obj["roe"] = "%.2f%%" % (item_detail["roe"] * 100)
index_obj["PE"] = "%.2f" % item_detail["pe_ttm"]
index_obj["PB"] = "%.2f" % item_detail["pb"]
_filter = {"code": item[1], "market": item[0], "date": {"$gte": date.datetime_to_str(date.years_ago(10))}}
item_list = stock_model.stock_list(_filter=_filter, _sort=[("date", pymongo.ASCENDING)],
_fields={"pe_ttm": 1, "pb": 1})
item_df = pd.DataFrame(list(item_list), columns=["pe_ttm", "pb"])
item_df = item_df.dropna(0)
_pe_series = 1 / item_df["pe_ttm"]
pe_ranks = _pe_series.rank(pct=True)
_pb_series = 1 / item_df["pb"]
pb_ranks = _pb_series.rank(pct=True)
index_obj["PE_RATE"] = "%.2f%%" % ((1 - pe_ranks.values[-1]) * 100)
index_obj["PB_RATE"] = "%.2f%%" % ((1 - pb_ranks.values[-1]) * 100)
# 本周以来的涨跌幅
now = datetime.today()
_tmp_date = (now - timedelta(days=now.weekday())).strftime('%Y%m%d')
item_last_week = stock_model.stock_item(_filter={"code": item[1], "market": item[0],
"date": _tmp_date if date.is_work_day(int(_tmp_date))
else str(date.next_day(int(_tmp_date)))})
if item_last_week is None:
item_last_week = item_detail
# 本月以来
_tmp_date = datetime(now.year, now.month, 1).strftime('%Y%m%d')
item_last_month = stock_model.stock_item(_filter={"code": item[1], "market": item[0],
"date": _tmp_date if date.is_work_day(int(_tmp_date))
else str(date.next_day(int(_tmp_date)))})
if item_last_month is None:
item_last_month = item_detail
# 本季度以来
month = (now.month - 1) - (now.month - 1) % 3 + 1
_tmp_date = datetime(now.year, month, 1).strftime('%Y%m%d')
item_last_quarter = stock_model.stock_item(_filter={"code": item[1], "market": item[0],
"date": _tmp_date if date.is_work_day(int(_tmp_date))
else str(date.next_day(int(_tmp_date)))})
if item_last_quarter is None:
item_last_quarter = item_detail
# 本年度以来
_tmp_date = datetime(now.year, 1, 1).strftime('%Y%m%d')
item_last_year = stock_model.stock_item(_filter={"code": item[1], "market": item[0],
"date": _tmp_date if date.is_work_day(int(int(_tmp_date)))
else str(date.next_day(int(_tmp_date)))})
if item_last_year is None:
item_last_year = item_detail
index_obj["last_week"] = "%.2f%%" % (
(item_detail["close"] - item_last_week["close"]) / item_last_week["close"] * 100)
index_obj["last_month"] = "%.2f%%" % (
(item_detail["close"] - item_last_month["close"]) / item_last_month["close"] * 100)
index_obj["last_quarter"] = "%.2f%%" % (
(item_detail["close"] - item_last_quarter["close"]) / item_last_quarter["close"] * 100)
index_obj["last_year"] = "%.2f%%" % (
(item_detail["close"] - item_last_year["close"]) / item_last_year["close"] * 100)
result.append(index_obj)
# 追加A股市场的估值信息
stock_a_df = pd.read_csv(config.get("files").get("stock_a"), header=0, encoding="utf8")
stock_a_df = stock_a_df.dropna(0)
if len(stock_a_df) > 0:
item_last = stock_a_df.loc[stock_a_df.index[-1]]
pe_array = 1 / stock_a_df["pe"]
pb_array = 1 / stock_a_df["pb"]
pe_ranks = pe_array.rank(pct=True)
pb_ranks = pb_array.rank(pct=True)
stock_a_item = {"idx": len(idx_list) + 1, "name": "A股市场", "interest": "%.2f%%" % (1 / item_last["pe"] * 100),
"dr": "%.2f%%" % (item_last["dr"] * 100), "roe": "%.2f%%" % (item_last["roe"] * 100),
"PE": "%.2f" % item_last["pe"], "PB": "%.2f" % item_last["pb"],
"PE_RATE": "%.2f%%" % ((1 - pe_ranks.values[-1]) * 100),
"PB_RATE": "%.2f%%" % ((1 - pb_ranks.values[-1]) * 100)}
result.append(stock_a_item)
# 追加国债收益供参考
bond_df = pd.read_csv(config.get("files").get("bond"), header=0, encoding="utf8")
if len(bond_df) > 0:
item_last = bond_df.loc[bond_df.index[-1]]
bond_item = {"idx": len(idx_list) + 2,
"name": "<a href='http://www.pbc.gov.cn/rmyh/108976/index.html#Container' "
"title='%s的历史利率' target='_blank'> 十年国债利率 </a>",
"interest": "%.2f%%" % item_last["10year"]}
result.append(bond_item)
return json.dumps(result)
def plot2(years, money):
"""
红利指数定投进阶方案
80% <= pe百分位: 全卖
60% <= pe百分位 < 80%: 定赎回
40% <= pe百分位 < 60%: 正常定投
20% <= pe百分位 < 40%: 1.5倍定投
pe百分位 < 20%: 2倍定投
:return:
"""
tax = 0.00025 # 买卖按照万分之2.5计算
for item in config.get("indexes"):
data = []
_market = item[0]
_idx_code = item[1]
_money_per = money
click.echo("开始计算 %d - %s 指数的回测策略" % (_market, _idx_code))
_start_date = date.datetime_to_str(date.years_ago(years))
last_day_list = date.last_days(int(_start_date), 1,
True) # 每月最后一个交易日集合
db_client = MongoDBClient(
config.get("db").get("mongodb"),
config.get("db").get("database"))
_idx_list = db_client.find_stock_list(_filter={
"code": _idx_code,
"market": _market
},
_sort=[("date",
pymongo.ASCENDING)],
_fields={
"date": 1,
"pe_ttm": 1,
"close": 1
})
item_df = pd.DataFrame(list(_idx_list),
columns=["date", "pe_ttm", "close"])
item_df = item_df.dropna(0)
_pe_series = 1 / item_df["pe_ttm"]
item_df.insert(0, 'rank', _pe_series.rank(pct=True))
item_df = item_df.reset_index()
total_count = 0 # 累计数量
for idx, item_date in enumerate(last_day_list):
last_df = item_df[item_df["date"] == str(item_date)]
if last_df is None or len(last_df) <= 0:
continue
_tmp_rank = last_df.ix[last_df.index.values[0]]["rank"]
item_point = last_df.ix[last_df.index.values[0]]["close"]
item_rank = float("%.2f" % ((1 - _tmp_rank) * 100))
if item_rank >= 80 and total_count > 0:
# 全部卖出, 卖出后要扣税费
item_count = total_count * item_point * (1 - tax) / item_point
data.append(
[item_date, item_rank, "卖出", -item_point, item_count, 0])
total_count = 0
elif 60 <= item_rank < 80:
# 卖出1份
_op = "卖出"
# 卖出的这一份应该包含税费，所以要多卖出一部分税费的数量
item_count = _money_per * (1 + tax) / item_point
if total_count >= item_count:
total_count -= item_count
elif total_count > 0:
item_count = total_count
total_count = 0
else:
_op = "/"
item_count = 0
total_count = 0
data.append([
item_date, item_rank, _op, -item_point, item_count,
total_count
])
elif 40 <= item_rank < 60:
# 买入1份
item_count = _money_per * (1 - tax) / item_point
total_count += item_count
data.append([
item_date, item_rank, "买入", item_point, item_count,
total_count
])
elif 20 <= item_rank < 40:
# 买入1.5份
item_count = (_money_per * 1.5) * (1 - tax) / item_point
total_count += item_count
data.append([
item_date, item_rank, "买入", item_point, item_count,
total_count
])
elif 0 <= item_rank < 20:
# 买入2份
item_count = (_money_per * 2) * (1 - tax) / item_point
total_count += item_count
data.append([
item_date, item_rank, "买入", item_point, item_count,
total_count
])
else:
data.append(
[item_date, item_rank, "/", item_point, 0, total_count])
result_df = pd.DataFrame(
data, columns=["日期", "百分位", "操作", "单价", "数量", "累计数量"])
result_df.to_csv(config.get("files").get("plots") %
(PROJECT_ROOT, "plot2", _market, _idx_code),
index=False,
mode="w",
header=True,
encoding='gb18030',
float_format="%.3f")