hi

import asyncio
import csv
import json
import requests
import pandas as pd
import numpy as np
import os
import glob
import nest_asyncio
import gzip
from datetime import datetime

import matplotlib.pyplot as plt
import matplotlib.ticker as mtick
from matplotlib.ticker import FormatStrFormatter

from tardis_client import TardisClient, Channel
import ccxt

# BYBIT

todate = 1676694984#int(datetime.utcnow().timestamp())
oldtodate = todate-10000

bybit_oi_dfs = []

while True:
    print(todate, oldtodate)
    bybit_oi = requests.get("https://api2.bybit.com/v2/public/open-interest?symbol=GMTUSDT&period=5min&limit=200&to={todate}".format(todate=todate)).json()
    data = pd.DataFrame(bybit_oi["result"])
    print(data.shape)
    todate = int(data.timestamp.max())
    print("todate: ", todate, oldtodate)
    if todate > oldtodate:
        bybit_oi_dfs.append(data)
    else:
        print(todate, oldtodate)
        break
    oldtodate=todate
    todate += 200*5*60


todate = 1676694984#int(datetime.utcnow().timestamp())
oldtodate = todate-10000

bybit_price_dfs = []

while True:
    print(todate, oldtodate)
    bybit_price = requests.get("https://api2.bybit.com/v2/public/open-interest?symbol=GMTUSDT&period=5min&limit=200&to={todate}".format(todate=todate)).json()
    data = pd.DataFrame(bybit_price["result"])
    print(data.shape)
    todate = int(data.timestamp.max())
    print("todate: ", todate, oldtodate)
    if todate > oldtodate:
        bybit_price_dfs.append(data)
    else:
        print(todate, oldtodate)
        break
    oldtodate=todate
    todate += 200*5*60


boidf = pd.concat(bybit_oi_dfs)

boidf["date"] = pd.to_datetime(boidf.timestamp, unit='s')

exch = ccxt.bybit()

symbol = "STXUSDT"
timer = "1m"
beginning=1676694984000
# UTC now - 1 minute
now = int((datetime.utcnow() - datetime(1970,1,1)).total_seconds()-60)*1000

dfs = []
import time
while beginning < now:
    data = exch.fetchOHLCV(
        f"{symbol}", timer, since=int(beginning)
    )
    df = pd.DataFrame(
        data, columns=["time", "open", "high", "low", "close", "volume"]
    )
    print(df.shape)
    dfs.append(df)
    print('max = ', df.time.max())
    beginning = df.time.max()
#     time.sleep(5)

df = pd.concat(dfs, ignore_index=True)

df["time_utc"] = pd.to_datetime(df.time, unit="ms").astype(str)
df["symbol"] = symbol


df["logret"] = np.log(df.close/df.close.shift(1))
df["ret"] = df.close/df.close.shift(1)-1

df["date"] = pd.to_datetime(df.time, unit='ms')

# ax = plt.figure().gca()
# df.set_index("date")[["ret", "logret"]].cumsum().plot(grid=True, ax=ax)
# plt.title(symbol + " Returns")

# ax.yaxis.set_major_formatter(mtick.PercentFormatter(1.0))

df["wap"] = (df.close*df.volume).rolling(5).sum()

df["roll_volume"] = df.volume.rolling(5).sum()

df["px"] = df.wap/df.roll_volume

ff = pd.merge(df, boidf, on="date")

ff["oi_delta"] = ff.open_interest-ff.open_interest.shift(1)

pos = ff.loc[ff.oi_delta>0].set_index("date")
times = ["1h", "4h", "8h", "24h", "48h", "96h"]

for time in times:
    pos["oi_wap_{}".format(time)] = (pos.oi_delta*pos.px).rolling(time).sum()
    pos["oi_sum_{}".format(time)] = (pos.oi_delta).rolling(time).sum()
    pos["oi_px_{}".format(time)] = pos["oi_wap_{}".format(time)]/pos["oi_sum_{}".format(time)] 

plt.figure()
pos.px.plot(color='black', alpha=0.6, label="Current Price")


plt.grid(True)
for time in times:
    pos["oi_px_{}".format(time)].plot(label=time, alpha=0.5, linestyle='dashed')
plt.legend()
plt.grid(True)

plt.title("$STX Bybit Time Lagged OI weighted avg price \n (periods with open OI only)")