h&d.py

import pandas as pd
import numpy as np
import math
import matplotlib.dates as mpl_dates
import matplotlib.pyplot as plt
from test import get_data
from scipy.stats import linregress
import plotly.graph_objects as go
from plotly.subplots import make_subplots
from datetime import datetime

pairs=['BTC-USDT','ETH-USDT','ADA-USDT','AXS-USDT','SOL-USDT','IOST-USDT','FLOW-USDT','IOTX-USDT','XMR-USDT','ETC-USDT','XTZ-USDT'
        ,'EGLD-USDT','SAND-USDT','BNB-USDT','ADA-USDT','XRP-USDT','CAKE-USDT','DOGE-USDT','DOT-USDT','AVAX-USDT','MATIC-USDT'
        ,'ALGO-USDT','ICP-USDT','VET-USDT','AAVE-USDT','AXS-USDT','UNI-USDT','FIL-USDT','SHIB-USDT','EOS-USDT','KCS-USDT'
        ,'NEAR-USDT','LTC-USDT','ATOM-USDT','LINK-USDT','BCH-USDT','TRX-USDT','XLM-USDT','MANA-USDT','HBAR-USDT','APE-USDT','FTM-USDT','GRT-USDT'
        ,'THETA-USDT','MKR-USDT']
for pair in pairs:
    df = get_data(pair)
    #Check if NA values are in data
    df=df[df['volume']!=0]
    df.reset_index(drop=True, inplace=True)
    df['date'] = [datetime.fromtimestamp(float(time)/1000) for time in df['date']]
    df.rename(columns = {'date' : 'time'}, inplace = True)
    
    df.isna().sum()


    
        
    def pivotid(df1, l, n1, n2): #n1 n2 before and after candle l
        if l-n1 < 0 or l+n2 >= len(df1):
            return 0
        pividlow=1
        pividhigh=1
        for i in range(l-n1, l+n2+1):
            if(df1.low[l]>df1.low[i]):
                pividlow=0
            if(df1.high[l]<df1.high[i]):
                pividhigh=0
        if pividlow and pividhigh:
            return 3
        elif pividlow:
            return 1
        elif pividhigh:
            return 2
        else:
            return 0
        
    df['pivot'] = df.apply(lambda x: pivotid(df, x.name,15,15), axis=1)
    df['shortpivot'] = df.apply(lambda x: pivotid(df, x.name,5,5), axis=1)
    
    
    
    def pointpos(x):
        if x['pivot']==1:
            return x['low']-1e-3
        elif x['pivot']==2:
            return x['high']+1e-3
        else:
            return np.nan

    def shortpointpos(x):
        if x['shortpivot']==1:
            return x['low']-2e-3
        elif x['shortpivot']==2:
            return x['high']+2e-3
        else:
            return np.nan
        
    df['pointpos'] = df.apply(lambda row: pointpos(row), axis=1)
    df['shortpointpos'] = df.apply(lambda row: shortpointpos(row), axis=1)
    dfpl = df
    fig = go.Figure(data=[go.Candlestick(x=dfpl.index,
                    open=dfpl['open'],
                    high=dfpl['high'],
                    low=dfpl['low'],
                    close=dfpl['close'])])

    fig.add_scatter(x=dfpl.index, y=dfpl['pointpos'], mode="markers",
                    marker=dict(size=5, color="MediumPurple"),
                    name="pivot")
    fig.add_scatter(x=dfpl.index, y=dfpl['shortpointpos'], mode="markers",
                    marker=dict(size=5, color="red"),
                    name="shortpivot")
    fig.update_layout(xaxis_rangeslider_visible=False)
    fig.show()
    backcandles= 28 #!!!should be less than pivot candles

    for candleid in range(99, len(df)-backcandles):
        if df.iloc[candleid].pivot != 2 or df.iloc[candleid].shortpivot != 2:
            continue
        
        maxim = np.array([])
        minim = np.array([])
        xxmin = np.array([])
        xxmax = np.array([])
        minbcount=0 #minimas before head
        maxbcount=0 #maximas before head
        minacount=0 #minimas after head
        maxacount=0 #maximas after head
        
        for i in range(candleid-backcandles, candleid+backcandles):
            if df.iloc[i].shortpivot == 1:
                minim = np.append(minim, df.iloc[i].low)
                xxmin = np.append(xxmin, i) #could be i instead df.iloc[i].name             
                if i < candleid:
                    minbcount=+1
                elif i>candleid:
                    minacount+=1
            if df.iloc[i].shortpivot == 2:
                maxim = np.append(maxim, df.iloc[i].high)
                xxmax = np.append(xxmax, i) # df.iloc[i].name
                if i < candleid:
                    maxbcount+=1
                elif i>candleid:
                    maxacount+=1
        
        if minbcount<1 or minacount<1 or maxbcount<1 or maxacount<1:
            continue

        slmin, intercmin, rmin, pmin, semin = linregress(xxmin, minim)
        headidx = np.argmax(maxim, axis=0)
        if maxim[headidx]-maxim[headidx-1]>1.5e-3 and maxim[headidx]-maxim[headidx+1]>1.5e-3 and abs(slmin)<=1e-4 and xxmin[0]>xxmax[headidx-1] and xxmin[1]<xxmax[headidx+1]:# and (maxim[headidx]-maxim[headidx+1])>(maxim[headidx+1]-minim[headidx+1]) and (maxim[headidx]-maxim[headidx-1])>(maxim[headidx-1]-minim[headidx-1]) :
            print(minbcount,minacount,maxbcount,maxacount, slmin, candleid)
            #print(maxim)
            #print(xxmax)
            #print(minim)
            #print(xxmin)
            break
            
        if candleid % 1000 == 0:
            print(candleid)