Как можно строить свечные графики в питоне.

dataFile = 'YNDX_190114_190115.csv'
timeZoneDiffSecs = 3 * 3600
pt = 0.01

plt.errorbar( xs[ blackBars ], totalCentres[ blackBars ], yerr = totalSpans[ blackBars ], ecolor = 'k', elinewidth = 0.5, capsize = 0, ls = 'none' )
plt.errorbar( xs[ blackBars ], bodyCentres[ blackBars ], yerr = np.abs( bodySpans[ blackBars ] ), ecolor = 'k', elinewidth = 0.5, capsize = 2, ls = 'none' )

bodyCentres = 0.5 * ( bars[ :, 1 ] + bars[ :, 4 ] )
bodySpans = 0.5 * ( bars[ :, 4 ] — bars[ :, 1 ] )
totalCentres = 0.5 * ( bars[ :, 2 ] + bars[ :, 3 ] )
totalSpans = 0.5 * ( bars[ :, 2 ] — bars[ :, 3 ] )

blackBars = np.abs( bodySpans ) < 0.25 * pt
greenBars = np.logical_and( np.logical_not( blackBars ),
                            bodySpans >= 0.25 * pt )
redBars = np.logical_not( np.logical_or( blackBars, greenBars ) )

import os.path, time
import numpy as np
import matplotlib.style
import matplotlib as mpl
mpl.style.use('classic')
mpl.rcParams['axes.formatter.useoffset'] = False
import matplotlib.pyplot as plt

dataFile = 'YNDX_190114_190115.csv'
#dataFile = 'YNDX_190114_190115-1.csv'
#dataFile = 'YNDX_190114_190115-2.csv'
#dataFile = 'YNDX_190114_190115-3.csv'
#dataFile = 'YNDX_181201_190115-4.csv'

def my_split( s, seps ): # this function splits line to parts separated with given separators
    res = [s]
    for sep in seps:
        s, res = res, []
        for seq in s:
            res += seq.split( sep )
    i = 0
    while i < len( res ):
        if res[i] == '':
            res.pop(i)
            continue
        i += 1
    return res

def loadFinamCsv( fname ):
    if not os.path.isfile( fname ):
        raise ValueError( 'wrong file name: %s' % fname )  

    counter = 0

    fi = open( fname, 'r' )

    tickerNameIsFound = False

    for line in fi: # this loop counts number of bars and reads ticker name from the first bar
        firstSymbol = line[ :1 ]
        if firstSymbol == '' or firstSymbol == '<':
            continue
        if not tickerNameIsFound:
            parsed = my_split( line, ',\n' )
            ticker = parsed[0]
            period = parsed[1]
            tickerNameIsFound = True
        counter += 1

    bars = np.zeros( ( counter, 6 ), dtype = np.float64 ) # create matrix for reading the whole file

    print( counter )

    fi.seek( 0, 0 ) # move file pointer to the beginning 

    counter = 0
                 
    for line in fi:
        firstSymbol = line[ :1 ]
        if firstSymbol == '' or firstSymbol == '<':
            continue
        
        parsed = my_split( line, ';,\n' )     

        timeStamp = parsed[2] + parsed[3]
        dtime = time.strptime( timeStamp + '+0300', '%Y%m%d%H%M%S%z' )
        timeEpoch = time.mktime( dtime )

        bars[ counter, : ] = np.array( ( timeEpoch, np.float64(parsed[4]), np.float64(parsed[5]),
                                         np.float64(parsed[6]), np.float64(parsed[7]), np.float64(parsed[8]) ) )

        counter += 1
        if counter % 1000 == 0:
            print( int( counter / 1000 ), end = ' ' )
        
    fi.close()
    print( '\n' )

    return { 'ticker': ticker, 'period': period, 'bars': bars }

def convertPeriodString( periodRaw ):
    try:
        numMins = int( periodRaw )
        if numMins % 60 != 0:
            return 'M%d' % numMins
        else:
            return 'H%d' % ( numMins // 60 )
    except ValueError:
        return periodRaw

timeZoneDiffSecs = 3 * 3600 # we need to know in advance the time zone difference between UTC
pt = 0.01 # we need to know in advance the min price step (point)

data = loadFinamCsv( dataFile )
ticker = data[ 'ticker' ]
period = data[ 'period' ]
bars = data[ 'bars' ]

periodFine = convertPeriodString( period )


xs = np.array( range( bars.shape[0] ) )

bodyCentres = 0.5 * ( bars[ :, 1 ] + bars[ :, 4 ] )
bodySpans = 0.5 * ( bars[ :, 4 ] - bars[ :, 1 ] )
totalCentres = 0.5 * ( bars[ :, 2 ] + bars[ :, 3 ] )
totalSpans = 0.5 * ( bars[ :, 2 ] - bars[ :, 3 ] )

blackBars = np.abs( bodySpans ) < 0.25 * pt
greenBars = np.logical_and( np.logical_not( blackBars ),
                            bodySpans >= 0.25 * pt )
redBars = np.logical_not( np.logical_or( blackBars, greenBars ) )

plt.clf()

plt.errorbar( xs[ blackBars ], totalCentres[ blackBars ], yerr = totalSpans[ blackBars ], ecolor = 'k', elinewidth = 0.5, capsize = 0, ls = 'none' )
plt.errorbar( xs[ blackBars ], bodyCentres[ blackBars ], yerr = np.abs( bodySpans[ blackBars ] ), ecolor = 'k', elinewidth = 0.5, capsize = 2, ls = 'none' )

plt.errorbar( xs[ greenBars ], totalCentres[ greenBars ], yerr = totalSpans[ greenBars ], ecolor = 'g', elinewidth = 0.5, capsize = 0, ls = 'none' )
plt.errorbar( xs[ greenBars ], bodyCentres[ greenBars ], yerr = np.abs( bodySpans[ greenBars ] ), ecolor = 'g', elinewidth = 0.5, capsize = 2, ls = 'none' )

plt.errorbar( xs[ redBars ], totalCentres[ redBars ], yerr = totalSpans[ redBars ], ecolor = 'r', elinewidth = 0.5, capsize = 0, ls = 'none' )
plt.errorbar( xs[ redBars ], bodyCentres[ redBars ], yerr = np.abs( bodySpans[ redBars ] ), ecolor = 'r', elinewidth = 0.5, capsize = 2, ls = 'none' )

plt.xlabel( 'Bar No., %s' % periodFine )
plt.ylabel( ticker ) # before was: fname[ : fname.find( ' ' ) ]

plt.xlim( xs[0] - 0.5, xs[-1] + 0.5 )

plt.annotate( 'start: %s' % ( time.strftime( '%Y-%m-%d %H:%M:%S', time.gmtime( bars[ 0, 0 ] + timeZoneDiffSecs ) ) ),
            xy=(0.1,0.95), xycoords='axes fraction',
            fontsize=11, horizontalalignment='left', verticalalignment='top' )

plt.annotate( 'end: %s' % ( time.strftime( '%Y-%m-%d %H:%M:%S', time.gmtime( bars[ -1, 0 ] + timeZoneDiffSecs ) ) ),
            xy=(0.1,0.90), xycoords='axes fraction',
            fontsize=11, horizontalalignment='left', verticalalignment='top' )

plt.savefig( dataFile[ : dataFile.rfind( '.' ) ] + '.png')

plt.show()