Bitmex price and volume chart using UDF datasource, with Python and Bokeh for chart

bitmex-plotinit-plotly.py

#!/usr/bin/python3
# coding=utf-8
#
# Dev Log
# bitmex-plotinit-plotly.py
# =======
# 2020-07-19 - setup script to pull data from Bitmex UDF and create candles and highlights
#            - intent is for a base script I can re-run to highlight specific things
# 2020-08-08 - switching to a different chart library
#
##########################

import os
import sys
import time
import datetime as dt
import json
import requests
import pandas as df
import numpy as np
import matplotlib.pyplot as plt
from plotly.subplots import make_subplots
import plotly.graph_objects as go

# debug to true or false
debug = "true"
showchart = "true"

# we need today to determine yesterday because we need to get OHLC until yesterday
today = dt.date.today()
yesterday = today - dt.timedelta(days=1)

# need start and end date in YYYY-MM-DD format
myStartDate = dt.datetime(2020, 1, 1, 0, 0).strftime('%Y-%m-%d')
#myEndDate = dt.datetime(2020, 8, 8, 0, 0).strftime('%Y-%m-%d')
myEndDate = today.strftime('%Y-%m-%d')

# price range, set to relevant figures for range being displayed
priceLow = 2000
priceHigh = 14000

# the Bitmex UDF response with OHLC is limited to 10080 replies and won't accept resolution of 1d, only 60,
# we'll create epochs and call in the data for each year, appending to the data frame each time
# earliest record for an hourly candle appears to be epoch datetime 1443182400
# this is GMT: Friday, September 25, 2015 12:00:00 PM
# period2015 will return ~30k rows from 25 Sept to 31 December 2015

# 2015
period2015Start = dt.datetime(2015, 9, 25, 0, 0).strftime('%s')
period2015End = dt.datetime(2015, 12, 31, 11, 59).strftime('%s')
period2015Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2015Start+"&to="+period2015End).json()
period2015Array = df.DataFrame(period2015Request)

# 2016
period2016Start = dt.datetime(2016, 1, 1, 0, 0).strftime('%s')
period2016End = dt.datetime(2016, 12, 31, 11, 59).strftime('%s')
period2016Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2016Start+"&to="+period2016End).json()
period2016Array = df.DataFrame(period2016Request)

# 2017
period2017Start = dt.datetime(2017, 1, 1, 0, 0).strftime('%s')
period2017End = dt.datetime(2017, 12, 31, 11, 59).strftime('%s')
period2017Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2017Start+"&to="+period2017End).json()
period2017Array = df.DataFrame(period2017Request)

# 2018
period2018Start = dt.datetime(2018, 1, 1, 0, 0).strftime('%s')
period2018End = dt.datetime(2018, 12, 31, 11, 59).strftime('%s')
period2018Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2018Start+"&to="+period2018End).json()
period2018Array = df.DataFrame(period2018Request)

# 2019 
period2019Start = dt.datetime(2019, 1, 1, 0, 0).strftime('%s')
period2019End = dt.datetime(2019, 12, 31, 11, 59).strftime('%s')
period2019Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2019Start+"&to="+period2019End).json()
period2019Array = df.DataFrame(period2019Request)

# 2020
period2020Start = dt.datetime(2020, 1, 1, 0, 0).strftime('%s')
# set to yesterday date, keeping manual for history purposes
#period2020End = dt.datetime(2020, 7, 17, 11, 59).strftime('%s')
#period2020End = yesterday.strftime('%s')
# 2020-08-08: setting to today
period2020End = today.strftime('%s')
period2020Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2020Start+"&to="+period2020End).json()
period2020Array = df.DataFrame(period2020Request)

# 2021 prep
#period2021Start = dt.datetime(2021, 1, 1, 0, 0).strftime('%s')
# set to yesterday date, keeping manual for history purposes
#period2021End = yesterday.strftime('%s')
#period2021Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2021Start+"&to="+period2021End).json()
#period2021Array = df.DataFrame(period2021Request)

# then we combined all the dataframes into an array of annual dataframes
##2021 prep, add period2021Array
myBigArray = [period2015Array, period2016Array, period2017Array, period2018Array, period2019Array, period2020Array]

# concat the dataframes, with ignore_index=True, so we get all rows
myCombinedResult = df.concat(myBigArray, axis=0, join='outer', ignore_index=True, keys=None, levels=None, names=None, verify_integrity=False, copy=True)

# remove the http ok 200 status column
del myCombinedResult['s']

# set the column headings
myCombinedResult.columns = ["datetime", "close", "open", "high", "low", "volume"]
myCombinedResult = myCombinedResult[["datetime", "open", "high", "low", "close", "volume"]]

# convert the datetime column into a datetime dtype
myCombinedResult["datetime"] = df.to_datetime(myCombinedResult["datetime"], errors='raise', dayfirst=False, yearfirst=False, utc=None, format=None, exact=True, unit='s', infer_datetime_format=False, origin='unix', cache=True)

# set the datetime to an index
myCombinedResult.set_index(["datetime"], drop=True, append=False, inplace=True, verify_integrity=False)

# resample hourly to daily
# myCombinedResult = myCombinedResult.resample('1440Min', axis=0).bfill()

# resample to 4 hours
myFourHourResult = myCombinedResult.resample('240Min', axis=0).bfill()

# Calculate Bollinger Bands
# https://school.stockcharts.com/doku.php?id=technical_indicators:bollinger_bands
# Middle Band = 20-day simple moving average (SMA)
# Upper Band = 20-day SMA + (20-day standard deviation of price x 2) 
# Lower Band = 20-day SMA - (20-day standard deviation of price x 2)
myFourHourResult["BBSMA"] = myFourHourResult["close"].rolling(window=20).mean()
myFourHourResult["BBSTD"] = myFourHourResult["close"].rolling(window=20).std()
myFourHourResult["BBUPPER"] = myFourHourResult["BBSMA"] + (myFourHourResult["BBSTD"] * 2)
myFourHourResult["BBLOWER"] = myFourHourResult["BBSMA"] - (myFourHourResult["BBSTD"] * 2)

# if the debug variable is true, print out some debug info
if debug == "true":
    print(myCombinedResult)
    print(myFourHourResult)
    print(myStartDate)
    print(myEndDate)

# if the showchart variable is true, plot a chart and open in browser
if showchart == "true":
    fig = go.Figure()
    fig.add_trace(go.Candlestick(x=myCombinedResult.index,
        open=myCombinedResult["open"],
        high=myCombinedResult["high"],
        low=myCombinedResult["low"],
        close=myCombinedResult["close"]))
    fig.add_trace(go.Scatter(
        x=myFourHourResult.index,
        y=myFourHourResult["BBUPPER"],
        name="upper bband",
        line_color='red',
        opacity=0.5))
    fig.update_layout(xaxis_rangeslider_visible=False)
    fig.update_layout(xaxis_range=[myStartDate,myEndDate],title_text="Bitmex price versus 4h upper bband with date range limited to selected time period")
    fig.update_layout(yaxis_range=[priceLow,priceHigh])
    fig.show()

bitmex-python-bokeh-chart.py

#!/usr/bin/python3
# coding=utf-8
#
# Dev Log
# plotinit.py
# =======
# 2020-07-19 - setup script to pull data from Bitmex UDF and create candles and highlights
#            - intent is for a base script I can re-run to highlight specific things
#
##########################

import os
import sys
import time
import datetime as dt
import json
import requests
import pandas as df
from bokeh.layouts import column
from bokeh.plotting import figure, show, output_file

# debug to true or false
debug = "true"
showchart = "true"

# we need today to determine yesterday because we need to get OHLC until yesterday
today = dt.date.today()
yesterday = today - dt.timedelta(days=1)

# the Bitmex UDF response with OHLC is limited to 10080 replies and won't accept resolution of 1d, only 60,
# we'll create epochs and call in the data for each year, appending to the data frame each time
# earliest record for an hourly candle appears to be epoch datetime 1443182400
# this is GMT: Friday, September 25, 2015 12:00:00 PM
# period2015 will return ~30k rows from 25 Sept to 31 December 2015

# 2015
period2015Start = dt.datetime(2015, 9, 25, 0, 0).strftime('%s')
period2015End = dt.datetime(2015, 12, 31, 11, 59).strftime('%s')
period2015Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2015Start+"&to="+period2015End).json()
period2015Array = df.DataFrame(period2015Request)

# 2016
period2016Start = dt.datetime(2016, 1, 1, 0, 0).strftime('%s')
period2016End = dt.datetime(2016, 12, 31, 11, 59).strftime('%s')
period2016Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2016Start+"&to="+period2016End).json()
period2016Array = df.DataFrame(period2016Request)

# 2017
period2017Start = dt.datetime(2017, 1, 1, 0, 0).strftime('%s')
period2017End = dt.datetime(2017, 12, 31, 11, 59).strftime('%s')
period2017Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2017Start+"&to="+period2017End).json()
period2017Array = df.DataFrame(period2017Request)

# 2018
period2018Start = dt.datetime(2018, 1, 1, 0, 0).strftime('%s')
period2018End = dt.datetime(2018, 12, 31, 11, 59).strftime('%s')
period2018Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2018Start+"&to="+period2018End).json()
period2018Array = df.DataFrame(period2018Request)

# 2019 
period2019Start = dt.datetime(2019, 1, 1, 0, 0).strftime('%s')
period2019End = dt.datetime(2019, 12, 31, 11, 59).strftime('%s')
period2019Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2019Start+"&to="+period2019End).json()
period2019Array = df.DataFrame(period2019Request)

# 2020
period2020Start = dt.datetime(2020, 1, 1, 0, 0).strftime('%s')
# set to yesterday date, keeping manual for history purposes
#period2020End = dt.datetime(2020, 7, 17, 11, 59).strftime('%s')
period2020End = yesterday.strftime('%s')
period2020Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2020Start+"&to="+period2020End).json()
period2020Array = df.DataFrame(period2020Request)

# 2021 prep
#period2021Start = dt.datetime(2021, 1, 1, 0, 0).strftime('%s')
# set to yesterday date, keeping manual for history purposes
#period2021End = yesterday.strftime('%s')
#period2021Request = requests.get("https://www.bitmex.com/api/udf/history?symbol=XBTUSD&resolution=60&from="+period2021Start+"&to="+period2021End).json()
#period2021Array = df.DataFrame(period2021Request)

# then we combined all the dataframes into an array of annual dataframes
##2021 prep, add period2021Array
myBigArray = [period2015Array, period2016Array, period2017Array, period2018Array, period2019Array, period2020Array]

# concat the dataframes, with ignore_index=True, so we get all rows
myCombinedResult = df.concat(myBigArray, axis=0, join='outer', ignore_index=True, keys=None, levels=None, names=None, verify_integrity=False, copy=True)

# remove the http ok 200 status column
del myCombinedResult['s']

# set the column headings
myCombinedResult.columns = ["datetime", "close", "open", "high", "low", "volume"]
myCombinedResult = myCombinedResult[["datetime", "open", "high", "low", "close", "volume"]]

# convert the datetime column into a datetime dtype
myCombinedResult["datetime"] = df.to_datetime(myCombinedResult["datetime"], errors='raise', dayfirst=False, yearfirst=False, utc=None, format=None, exact=True, unit='s', infer_datetime_format=False, origin='unix', cache=True)

# set the datetime to an index
myCombinedResult.set_index(["datetime"], drop=True, append=False, inplace=True, verify_integrity=False)

# resample hourly to daily
# myCombinedResult = myCombinedResult.resample('1440Min', axis=0).bfill()

# resample to 4 hours
# myCombinedResult = myCombinedResult.resample('240Min', axis=0).bfill()

# if the debug variable is true, print out some debug info
if debug == "true":
    print(myCombinedResult)

# is the shotchart variable is true, plot a chart and open in browser
if showchart == "true":
    # plot results
    # w = 1*60*60*1000 # 1h in ms
    w = 24*60*60*1000 # 1d in ms
    chartTools = "pan,wheel_zoom,box_zoom,reset,save"

    # price chart
    pricechart = figure(x_axis_type="datetime", tools=chartTools, plot_width=1400, title = "Bitmex XBTUSD - Title Goes Here", output_backend="webgl", sizing_mode="fixed")
    pricechart.grid.grid_line_alpha=0.3
    # add price line from close
    pricechart.line(x = 'datetime', y = 'close', source = myCombinedResult, color = 'black', line_width = 1)
    # configure visual properties on a plot's title attribute
    pricechart.title.align = "center"
    pricechart.title.text_color = "black"
    pricechart.title.text_font_size = "30px"

    # volume chart - use a bar chart to show volume
    volchart = figure(x_axis_type="datetime", tools=chartTools, plot_height=100, title = "Bitmex XBTUSD daily trade volumes", output_backend="webgl", sizing_mode="scale_width")
    volchart.vbar(x = 'datetime', top = 'volume', source = myCombinedResult, width = w, color = "pink")
    volchart.grid.grid_line_alpha=0.3
    volchart.title.align = "center"
    volchart.title.text_color = "black"
    volchart.title.text_font_size = "18px"

    # save to file
    output_file("report.html", title="Title Goes Here")

    # open a browser
    show(column(pricechart,volchart,sizing_mode="scale_height"))

newer file is Bitmex price chart using UDF datasource, with Python and Plotly for chart, with 4H bbands, but no volume charts yet.