binance_historical_klines_to_csv.py

import time
import dateparser
import pytz
import json
import csv
import datetime
from dateutil.rrule import rrule, MONTHLY
from binance.client import Client
import os
import sys

pathname = os.path.dirname(sys.argv[0])
full_path = '{}/'.format(str(os.path.abspath(pathname)))

def grab_currencies_list():
	base_currencies_list = list()
	currencies_list = list()
	info = Client(None,None)
	pair_query = info.get_all_tickers()
	list_preference = input('Which base currencies would you like to grab data for?(eg. BTC or ETH,BNB or BTC,ETH,BNB,USDT) ')
	split_preference = list_preference.split(',')
	for x in split_preference:
		base_currency = x.strip().upper()
		base_currencies_list.append(base_currency)
	for y in base_currencies_list:
		base = y
		for z in range(0,len(pair_query)):
			symbol = pair_query[z]['symbol']
			if symbol[-len(base):] == base:
				currencies_list.append(symbol)
	return currencies_list

def grab_date_interval():
	print ('What date range would you like to pull data from?\nIn MM/YYYY format,except you can enter now for end date to get most recent.')
	start_date_input = input('Start date: ')
	start_date_split = start_date_input.split('/')
	month = int(start_date_split[0])
	day = 1
	year = int(start_date_split[1])
	start_date = datetime.date(year,month,day)
	end_date_input = input('End date: ')
	if end_date_input.lower() in ['now']:
		end_date = datetime.datetime.now()
	else:
		end_date_split = end_date_input.split('/')
		month = int(end_date_split[0])
		year = int(end_date_split[1])
		if month in [1,3,5,7,8,10,12]:
			day = 31
		elif month in [4,6,9,11]:
			day = 30
		elif month in [2] and year not in [2020,2024,2028,2032,2036,2040]:
			day = 28
		else:
			day = 29
		end_date = datetime.date(year,month,day)
	return start_date,end_date

def grab_kline_interval():
	kline_interval = input('What Kline Interal would you prefer? Options: 1m,3m,5m,15m,30m,1h,2h,4h,6h,8h,12h ')
	if kline_interval in ['1m','3m','5m','15m','30m','1h','2h','4h','6h','8h','12h']:
		return kline_interval
	else:
		print ('{} is an invalid option, please try again'.format(str(kline_interval)))
		kline_interval_2 = input('What Kline Interal would you prefer? Options: 1m,3m,5m,15m,30m,1h,2h,4h,6h,8h,12h ')
		if kline_interval_2 in ['1m','3m','5m','15m','30m','1h','2h','4h','6h','8h','12h']:
			return kline_interval_2


def create_directories(pair_list):
	main_directory_finder = [x[0] for x in os.walk(full_path)]
	historical_price_data_directory = '{}historical_price_data'.format(str(full_path))
	if historical_price_data_directory not in main_directory_finder:
		os.makedirs(historical_price_data_directory)
	pair_directory_finder = [x[0] for x in os.walk('{}historical_price_data/'.format(str(full_path)))]
	for x in range(0,len(pair_list)):
		symbol = pair_list[x]
		if '{}/{}'.format(str(historical_price_data_directory),str(symbol)) not in pair_directory_finder:
			os.makedirs('{}historical_price_data/{}'.format(str(full_path),str(symbol)))

def date_to_milliseconds(date_str):
	epoch = datetime.datetime.utcfromtimestamp(0).replace(tzinfo=pytz.utc)
	d = dateparser.parse(date_str)
	if d.tzinfo is None or d.tzinfo.utcoffset(d) is None:
		d = d.replace(tzinfo=pytz.utc)
	return int((d - epoch).total_seconds() * 1000.0)


def interval_to_milliseconds(interval):
	ms = None
	seconds_per_unit = {'m': 60,'h': 60 * 60,'d': 24 * 60 * 60,'w': 7 * 24 * 60 * 60}
	unit = interval[-1]
	if unit in seconds_per_unit:
		try:
			ms = int(interval[:-1]) * seconds_per_unit[unit] * 1000
		except ValueError:
			pass
	return ms


def get_historical_klines(symbol, interval, start_str, end_str=None):
	client = Client(None, None)
	output_data = []
	limit = 500
	timeframe = interval_to_milliseconds(interval)
	start_ts = date_to_milliseconds(start_str)
	end_ts = None
	if end_str:
		end_ts = date_to_milliseconds(end_str)
	idx = 0
	symbol_existed = False
	while True:
		try:
			temp_data = client.get_klines(symbol=symbol,interval=interval,limit=limit,startTime=start_ts,endTime=end_ts)
			if not symbol_existed and len(temp_data):
				symbol_existed = True
			if symbol_existed:
				output_data += temp_data
				start_ts = temp_data[len(temp_data) - 1][0] + timeframe
			else:
				start_ts += timeframe
			idx += 1
		except IndexError:
			idx+=1
		if len(temp_data) < limit:
			break
		if idx % 3 == 0:
			time.sleep(1)
	return output_data

def grab_data(pair_list,dates,start_date,end_date,interval):
	for x in range(0,len(pair_list)):
		symbol = pair_list[x]
		for x in range(0,len(dates)):
			date_splitter = str(dates[x]).split('-')
			year = date_splitter[0]
			numerical_month = date_splitter[1]
			month_abbreviation_dict = {'01':'Jan','02':'Feb','03':'Mar','04':'Apr','05':'May','06':'Jun','07':'Jul','08':'Aug','09':'Sept','10':'Oct','11':'Nov','12':'Dec'}
			for num,abbr in month_abbreviation_dict.items():
				if numerical_month==num:
					calendar_month = abbr
			klines_date = '{}, {}'.format(str(calendar_month),str(year))
			csv_month = '{}-{}-'.format(str(year),str(numerical_month))
			if numerical_month in ['01','03','05','07','08','10','12']:
				day = range(1,32)
			elif numerical_month in ['04','06','09','11']:
				day = range(1,31)
			elif numerical_month in ['02'] and year not in ['2020','2024','2028','2032','2036','2040','2044']:
				day = range(1,29)
			else:
				day = range(1,30)
			for y in day:
				next_day = y+1
				start = '{} {}'.format(str(y),str(klines_date))
				if y == day[-1]:
					next_day = 1
					if int(numerical_month) in range(1,12):
						next_numerical_month = int(numerical_month)+1
					else:
						next_numerical_month = 1
						year = int(year)+1
					next_numerical_month_abbreviation_dict = {'1':'Jan','2':'Feb','3':'Mar','4':'Apr','5':'May','6':'Jun','7':'Jul','8':'Aug','9':'Sept','10':'Oct','11':'Nov','12':'Dec'}
					for key,val in next_numerical_month_abbreviation_dict.items():
						if str(next_numerical_month)==key:
							calendar_month = val
					klines_date = '{}, {}'.format(str(calendar_month),str(year))
				end = '{} {}'.format(str(next_day),str(klines_date))
				print ('symbol {} start {} end {}'.format(str(symbol),str(start),str(end)))
				klines = get_historical_klines(symbol, interval, start, end)
				if len(klines)>0:
					titles = ('Date','Open','High','Low','Close','Volume')
					if y in range(1,10):
						csv_day = '0{}'.format(str(y))
					else:
						csv_day = str(y)
					with open('{}historical_price_data/{}/{}{}_{}.csv'.format(str(full_path),str(symbol),str(csv_month),str(csv_day),str(interval)), 'a') as f:
							writer = csv.writer(f)
							writer.writerow(titles)
					history_list = list()
					for x in range(0,(len(klines)-1)):
						open_epoch_timestamp = float(klines[x][0])/1000
						open_time = datetime.datetime.utcfromtimestamp(open_epoch_timestamp).strftime('%Y-%m-%d %H:%M:%S.%f')
						open_price = klines[x][1]
						high = klines[x][2]
						low = klines[x][3]
						close_price = klines[x][4]
						volume = klines[x][5]
						close_epoch_timestamp = float(klines[x][6])/1000
						close_time = datetime.datetime.utcfromtimestamp(close_epoch_timestamp).strftime('%Y-%m-%d %H:%M:%S.%f')
						quote_asset_volume = klines[x][7]
						number_of_trades = klines[x][8]
						taker_buy_base_asset_volume = klines[x][9]
						taker_buy_quote_asset_volume = klines[x][10]
						fields = (open_time,open_price,high,low,close_price,volume)
						history_list.append(fields)
						with open('{}historical_price_data/{}/{}{}_{}.csv'.format(str(full_path),str(symbol),str(csv_month),str(csv_day),str(interval)), 'a') as f:
							writer = csv.writer(f)
							writer.writerow(fields)

def main():
	pair_list = grab_currencies_list()
	interval = grab_kline_interval()
	create_needed_directories = create_directories(pair_list)
	start_date,end_date = grab_date_interval()
	dates = [date for date in rrule(MONTHLY, dtstart=start_date, until=end_date)]
	grab_data(pair_list,dates,start_date,end_date,interval)

if __name__ == '__main__':
	main()