joinquant code snippet to visualize on pe,pb history rank for several index

index_pepb.py

import pandas as pd
import datetime as dt
import os
import numpy as np
import jqdata
import bisect
import matplotlib.pyplot as plt
from pyecharts import Page,online,Line,Grid
import warnings
warnings.filterwarnings('ignore')

online()

origin = '2005-01-01'
today = dt.datetime.strftime(pd.datetime.today()-dt.timedelta(1),'%Y-%m-%d')

def convert_code(code):
    if code.endswith('XSHG'):
        return 'sh' + code[0:6]
    elif code.endswith('XSHE'):
        return 'sz' + code[0:6] 

def get_ff(index_code,date=today):
    stock_list = get_index_stocks(index_code, date)
    if len(stock_list)>0:
        q = query(valuation).filter(valuation.code.in_(list(stock_list)))
        df = get_fundamentals(q, date)
    else:
        df = pd.DataFrame()
    return df

def get_ff_w(index_code,date=today,weight='C'):
    result = get_ff(index_code,date)
    if result.empty:
        return result
    result=result.loc[:,['code','pe_ratio','pb_ratio','market_cap']]
    num = len(result.iloc[:,0])
    if weight == 'E': #equal weight
        return result.assign(weight=pd.Series([1/float(num) for _ in range(num)]))
    elif weight == 'C': #based on capital
        total = sum([result.iloc[i].market_cap for i in range(num)])
        wei = [result.iloc[i].market_cap for i in range(num)]/total
        return result.assign(weight=pd.Series(wei))
    else:
        wei = weight(result)
        return result.assign(weight=pd.Series(wei))
    

def get_pepb(index_code,date=today,weight='C'):
    result = get_ff_w(index_code,date,weight)
    if result.empty:
        return [0,0]
    num = len(result.iloc[:,0])
    if num>0:
        pede = np.array([result.iloc[i,4]/result.iloc[i,1] for i in range(num)])
        pbde = np.array([result.iloc[i,4]/result.iloc[i,2] for i in range(num)])
        pede =pede[pede<100]
        pbde =pbde[pbde<100]
        return [1/sum(pede),1/sum(pbde)]
    else:
        return [0,0]
    
    
def get_pepb_range(index_code,start=origin,end=today,weight='C'):
    pebs = pd.DataFrame()
    for date in pd.DatetimeIndex(jqdata.get_trade_days(start_date=start,end_date=end)):
        s = get_pepb(index_code, dt.datetime.strftime(date,'%Y-%m-%d'), weight)
        if s != [0,0]:
            pebs[dt.datetime.strftime(date,'%Y-%m-%d')]=pd.Series(s,index=['pe','pb'])
    return pebs.dropna()




class index:
    
    def __init__(self,code,start=origin,end=today,weight='C'):
        self.code, self.start, self.end, self.weight = code, dt.datetime.strptime(start,'%Y-%m-%d'), dt.datetime.strptime(end,'%Y-%m-%d'), weight
        self.datapath = '%s%spepb.csv'%(convert_code(self.code),self.weight)
        if (self.start-dt.datetime.strptime(origin,'%Y-%m-%d')).total_seconds()<0 or (self.end-self.start).total_seconds()<0 or (self.end-pd.datetime.today()+dt.timedelta(1)).total_seconds()>0:
            raise Exception('date range error')
    
    def __repr__(self):
        return self.code
    __str__ = __repr__

    def update_pepbs(self):
        if os.path.exists(self.datapath):
            cps = pd.DataFrame.from_csv(self.datapath)
            start_date = dt.datetime.strptime(cps.iloc[:,-1].name,'%Y-%m-%d') + dt.timedelta(1)
            origin_date = dt.datetime.strptime(cps.iloc[:,0].name,'%Y-%m-%d')
            omit = (self.end-start_date).days
            omitb = (origin_date-self.start).days
            if omit>=-1:
                self.pepbs = pd.concat([cps, get_pepb_range(self.code,start_date,self.end,self.weight)],axis=1) 
                self.pepbs.to_csv(self.datapath)
            else:
                self.pepbs = cps.loc[:,:dt.datetime.strftime(self.end,'%Y-%m-%d')]
            if omitb>=0:
                self.start = origin_date
            else:
                self.pepbs = self.pepbs.loc[:,dt.datetime.strftime(self.start,'%Y-%m-%d'):]
        else:
            self.pepbs = get_pepb_range(self.code,self.start,self.end,self.weight)
            if self.start-dt.datetime.strptime(origin,'%Y-%m-%d') == dt.timedelta(0):
                self.pepbs.to_csv(self.datapath)
            self.start = self.pepbs.iloc[:,0].name
    
    def img_line(self):
        all_index = get_all_securities(['index'])
        name = all_index.ix[self.code].display_name
        line = Line(name,"PB chart")
        ref=[self.pepbs.iloc[1].quantile(i/10.0) for i in range(11)]
        for i in range(11):
            line.add('ref',list(self.pepbs.iloc[0].index),[ref[i] for _ in self.pepbs.iloc[1].index],is_symbol_show=False,line_opacity=0,mark_line=['average'],mark_line_symbolsize=0)
        line.add('pb',list(self.pepbs.iloc[0].index),list(self.pepbs.iloc[1].values),is_symbol_show=True,is_smooth=True,is_more_utils=True)
#         page.add(line)
        line2 = Line(name,"PE chart")
        ref=[self.pepbs.iloc[0].quantile(i/10.0) for i in range(11)]
        for i in range(11):
            line2.add('ref',list(self.pepbs.iloc[0].index),[ref[i] for _ in self.pepbs.iloc[0].index],is_symbol_show=False,line_opacity=0,mark_line=['average'],mark_line_symbolsize=0)
        line2.add('pe',list(self.pepbs.iloc[0].index),list(self.pepbs.iloc[0].values),is_symbol_show=True,is_smooth=True,is_more_utils=True)
        line2
#         page.add(line2)
        return [line,line2]
    def img_page(self):
        page=Page()
        line = self.img_line()
        page.add(line[0])
        page.add(line[1])
        return page

def pos(pepbs,pe,pb):
    idx = bisect.bisect(sort(list(pepbs.loc['pe'])),pe)
    idy = bisect.bisect(sort(list(pepbs.loc['pb'])),pb)
    return [idx/float(len(pepbs.iloc[0])),idy/float(len(pepbs.iloc[0]))]
def pos_hist(pepbs,date):
    pepbs=pepbs.loc[:,:date]
    pe=pepbs.loc['pe',date]
    pb=pepbs.loc['pb',date]
    return pos(pepbs,pe,pb)
def hist_line(pepbs,start=origin):
    h = pd.DataFrame()
    for date in pd.DatetimeIndex(jqdata.get_trade_days(start_date=start)):
        h[date]=pd.Series(pos_hist(pepbs,dt.datetime.strftime(date,'%Y-%m-%d')),index=['pe','pb'])
    return h.transpose()
   
   
'''
page=Page()
for code in ['000016.XSHG','000300.XSHG','000905.XSHG','399006.XSHE','399971.XSHE','399975.XSHE','000932.XSHG','000925.XSHG']:
    demo=index(code)
    demo.update_pepbs()
    a=demo.img_line()
    page.add(a[0])
    page.add(a[1])
for code in ['000922.XSHG','000831.XSHG','000827.XSHG','000978.XSHG','399812.XSHE']:
    demo=index(code,weight='E')
    demo.update_pepbs()
    a=demo.img_line()
    page.add(a[0])
    page.add(a[1])
page.render('graph.html')
# page
'''