Python金融科技:cufflinks繪製金融圖表

前言

最近發現一個功能強大的繪圖工具庫cufflinks，其最吸引我的地方是內置了量化金融繪圖模塊，可以很方便地繪製K線和技術指標圖表。但遺憾的是，在網絡上並沒有找到cufflinks的參考手冊。雖然網絡上有一些介紹cufflinks的博客文章，但都沒有詳細介紹量化繪圖模塊的使用方法。

因此本蒟蒻參照cufflinks的github源碼，對cufflinks量化金融繪圖模塊的使用方法做出簡要的介紹，希望對您有些許幫助。

cufflinks介紹

cufflinks是對Python繪圖庫plotly的進一步封裝，是基於jupyter notebook的交互式繪圖工具，可以很方便地繪製各種交互式圖表，佈局優美，效果炫酷，操作方便，可以和pandas的數據類型無縫銜接。

cufflinks的量化金融繪圖模塊接收的金融交易數據爲，dataframe格式，index爲datetime類型，各列爲open、high、low、close、volume，如下：

使用方法

注意：目前cufflinks只能在jupyter notebook中使用，在其他的IDE中使用可能會無法 編譯執行。

因爲本蒟蒻水平有限，如有錯誤，歡迎批評指正。

筆者的環境：cufflinks 0.17.3、Python3.7 、Jupyter Notebook

更多信息可以參考cufflinks的源碼

安裝cufflinks：

pip install cufflinks

1.繪製基礎K線（蠟燭）圖

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
df=cf.datagen.ohlcv()#cufflinks提供的數據，也可以更改爲自定義數據
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.iplot()

效果圖:

2.添加趨勢線

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_trendline('2015-01-01','2015-04-10',on='close',text='趨勢線',textangle=0)#畫趨勢線，textangle設置文字角度
qf.iplot(up_color='red',down_color='green')

效果圖：

參數介紹（摘自源碼註釋）：

"""
		Parameters:
			date0 : string
				Trendline starting date
			date1 :  string
				Trendline end date
			on : string
				Indicate the data series in which the 
				trendline should be based.
					'close'
					'high'
					'low'
					'open'
			text : string
				If passed, then an annotation will be added 
				to the trendline (at mid point)
		kwargs:
			from_strfmt : string
				Defines the date formating in which 
				date0 and date1 are stated. 
				default: '%d%b%y'					
			to_strfmt : string
				Defines the date formatting 
				to which it should be converted. 
				This should match the same format as the timeseries index. 
				default : '%Y-%m-%d'		
"""

3.添加支撐線

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_support('2015-01-12',on='close',mode='toend',text='這是支持線')
qf.iplot()

參數介紹：

"""
		Parameters:
			date0 : string
				The support line will be drawn at the 'y' level 
				value that corresponds to this date. 
			on : string
				Indicate the data series in which the 
				support line should be based.
					'close'
					'high'
					'low'
					'open'
			mode : string
				Defines how the support/resistance will 
				be drawn
					'starttoened' : (x0,x1)
					'fromstart' : (x0,date)
					'toend' : (date,x1)
			text : string
				If passed, then an annotation will be added 
				to the support line (at mid point)
		
		kwargs:	
			from_strfmt : string
				Defines the date formating in which 
				date0 and date1 are stated. 
				default: '%d%b%y'					
			to_strfmt : string
				Defines the date formatting 
				to which it should be converted. 
				This should match the same format as the timeseries index. 
				default : '%Y-%m-%d'		
		"""

4.添加壓力線

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_resistance('2015-01-12',on='close',mode='fromstart',text='這是壓力線',textangle=30)
qf.iplot()

效果圖：
參數介紹：
與支撐線相似

5.添加註解

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_annotations({'2015-02-01':'我是註解'},fontcolor='red',fontsize=18,textangle=0)
qf.iplot()

效果圖：

"""
		Parameters:
			annotations : dict or list(dict,)
				Annotations can be on the form form of 
					{'date' : 'text'}
					and the text will automatically be placed at the 
					right level on the chart 
				or
					A Plotly fully defined annotation
		kwargs : 
			fontcolor : str
				Text color for annotations
			fontsize : int
				Text size for annotations
			textangle : int
				Textt angle 
			See https://plot.ly/python/reference/#layout-annotations 
			for a complete list of valid parameters.
		"""

6.添加交易量

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_volume()
qf.iplot(up_color='red',down_color='green')#可以設置主題顏色

效果圖：

參數介紹：

"""
		Parameters:
			colorchange : bool
				If True then each volume bar will have a fill color 
				depending on if 'base' had a positive or negative
				change compared to the previous value
				If False then each volume bar will have a fill color 
				depending on if the volume data itself had a positive or negative
				change compared to the previous value
			column :string
				Defines the data column name that contains the volume data. 
				Default: 'volume'
			name : string
				Name given to the study
			str : string
				Label factory for studies
				The following wildcards can be used:
					{name} : Name of the column
					{study} : Name of the study
					{period} : Period used
				Examples:
					'study: {study} - period: {period}'
			
		kwargs : 
			base : string
				Defines the column which will define the
				positive/negative changes (if colorchange=True).
				Default = 'close'
			up_color : string
				Color for positive bars
			down_color : string
				Color for negative bars
		"""

7.添加異同移動平均線MACD

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_macd()
qf.iplot()

效果圖：

參數介紹：

"""
		Parameters:
			fast_period : int
				MACD Fast Period
			slow_period : int
				MACD Slow Period
			signal_period : int
				MACD Signal Period
			column :string
				Defines the data column name that contains the 
				data over which the study will be applied. 
				Default: 'close'
			name : string
				Name given to the study
			str : string
				Label factory for studies
				The following wildcards can be used:
					{name} : Name of the column
					{study} : Name of the study
					{period} : Period used
				Examples:
					'study: {study} - period: {period}'
		kwargs: 
			legendgroup : bool
				If true, all legend items are grouped into a 
				single one
			All formatting values available on iplot()
		"""

8.添加簡單移動均線SMA

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_sma(periods=[10,30],color=['red','blue'])
qf.iplot()

參數介紹：

"""
		Parameters:
			periods : int or list(int)
				Number of periods
			column :string
				Defines the data column name that contains the 
				data over which the study will be applied. 
				Default: 'close'
			name : string
				Name given to the study
			str : string
				Label factory for studies
				The following wildcards can be used:
					{name} : Name of the column
					{study} : Name of the study
					{period} : Period used
				Examples:
					'study: {study} - period: {period}'
		kwargs: 
			legendgroup : bool
				If true, all legend items are grouped into a 
				single one
			All formatting values available on iplot()
		"""

9.添加EMA

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_ema(periods=[10,20])
qf.iplot()

效果圖：

參數介紹：

"""
		Parameters:
			periods : int or list(int)
				Number of periods
			column :string
				Defines the data column name that contains the 
				data over which the study will be applied. 
				Default: 'close'
			name : string
				Name given to the study
			str : string
				Label factory for studies
				The following wildcards can be used:
					{name} : Name of the column
					{study} : Name of the study
					{period} : Period used
				Examples:
					'study: {study} - period: {period}'
		kwargs: 
			legendgroup : bool
				If true, all legend items are grouped into a 
				single one
			All formatting values available on iplot()
		"""

10.添加RSI

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_rsi(periods=10)
qf.iplot()

效果圖：

參數介紹：

"""
		Parameters:
			periods : int or list(int)
				Number of periods
			rsi_upper : int 
				bounds [0,100]
				Upper (overbought) level
			rsi_lower : int
				bounds [0,100]
				Lower (oversold) level
			showbands : boolean
				If True, then the rsi_upper and
				rsi_lower levels are displayed
			column :string
				Defines the data column name that contains the 
				data over which the study will be applied. 
				Default: 'close'
			name : string
				Name given to the study
			str : string
				Label factory for studies
				The following wildcards can be used:
					{name} : Name of the column
					{study} : Name of the study
					{period} : Period used
				Examples:
					'study: {study} - period: {period}'
		kwargs: 
			legendgroup : bool
				If true, all legend items are grouped into a 
				single one
			All formatting values available on iplot()
		"""

11.添加布林線

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_bollinger_bands(periods=10)
qf.iplot()

效果圖：

參數介紹：

"""
		Parameters:
			periods : int or list(int)
				Number of periods
			boll_std : int
				Number of standard deviations for
				the bollinger upper and lower bands
			fill : boolean
				If True, then the innner area of the 
				bands will filled
			column :string
				Defines the data column name that contains the 
				data over which the study will be applied. 
				Default: 'close'
			name : string
				Name given to the study
			str : string
				Label factory for studies
				The following wildcards can be used:
					{name} : Name of the column
					{study} : Name of the study
					{period} : Period used
				Examples:
					'study: {study} - period: {period}'
		kwargs: 
			legendgroup : bool
				If true, all legend items are grouped into a 
				single one
			fillcolor : string
				Color to be used for the fill color.
				Example:
					'rgba(62, 111, 176, .4)'
			All formatting values available on iplot()
		"""

12.添加順勢指標/商品通道指標CCI

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_cci(periods=[10,20])
qf.iplot()

效果圖：

參數介紹：

"""
		Parameters:
			periods : int or list(int)
				Number of periods
			cci_upper : int
				Upper bands level
				default : 100
			cci_lower : int
				Lower band level
				default : -100
			showbands : boolean
				If True, then the cci_upper and
				cci_lower levels are displayed
			name : string
				Name given to the study
			str : string
				Label factory for studies
				The following wildcards can be used:
					{name} : Name of the column
					{study} : Name of the study
					{period} : Period used
				Examples:
					'study: {study} - period: {period}'
		kwargs: 
			legendgroup : bool
				If true, all legend items are grouped into a 
				single one
			All formatting values available on iplot()
		"""

13.添加平均趨向指數ADX

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_adx(periods=5)
qf.iplot()

效果圖：

參數介紹：

"""
		Parameters:
			periods : int or list(int)
				Number of periods
			name : string
				Name given to the study
			str : string
				Label factory for studies
				The following wildcards can be used:
					{name} : Name of the column
					{study} : Name of the study
					{period} : Period used
				Examples:
					'study: {study} - period: {period}'
		kwargs: 
			legendgroup : bool
				If true, all legend items are grouped into a 
				single one
			All formatting values available on iplot()
		"""

14.添加均幅指標ATR

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_atr(periods=5)
qf.iplot()

效果圖：

參數介紹：

"""
		Parameters:
			periods : int or list(int)
				Number of periods
			name : string
				Name given to the study
			str : string
				Label factory for studies
				The following wildcards can be used:
					{name} : Name of the column
					{study} : Name of the study
					{period} : Period used
				Examples:
					'study: {study} - period: {period}'
		kwargs: 
			legendgroup : bool
				If true, all legend items are grouped into a 
				single one
			All formatting values available on iplot()
		"""

15.添加趨向指標DMI

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_dmi(periods=5)
qf.iplot()

效果圖：

參數介紹：

"""
		Parameters:
			periods : int or list(int)
				Number of periods
			name : string
				Name given to the study
			str : string
				Label factory for studies
				The following wildcards can be used:
					{name} : Name of the column
					{study} : Name of the study
					{period} : Period used
				Examples:
					'study: {study} - period: {period}'
		kwargs: 
			legendgroup : bool
				If true, all legend items are grouped into a 
				single one
			All formatting values available on iplot()
		"""

拓展內容

1.剔除非交易日期

當我們使用自己的數據繪圖時，其默認會把 非交易日也包含在圖中。例如下面從tushare中獲取數據：

import cufflinks as cf
import tushare as ts
cf.set_config_file(offline=True, world_readable=True)
df=ts.get_hist_data('000001',start='2020-04-15',end='2020-05-13')
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.iplot()

效果圖：

可以看到，繪製的圖像包含了非交易日。如果我們想要剔除這些非交易日，需要把代碼修改爲：

import cufflinks as cf
import tushare as ts
cf.set_config_file(offline=True, world_readable=True)
df=ts.get_hist_data('000001',start='2020-04-15',end='2020-05-13')
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
layout = dict(
    xaxis=dict(
        categoryorder="category ascending",#篩除非交易日
        type='category')
)
qf.iplot(layout=layout)

效果圖：

可以看到，在添加了layout後，圖像底部多出了一個選擇橫座標範圍的小條。同時在layout中將xaxis的type設置爲‘category’便可以剔除非交易日了。

2.拋物轉向指標add_ptps報錯

除了上面介紹的方法外，cufflinks還有添加拋物轉向指標add_ptps(),但我測試的時候，該方法一直報錯。

import cufflinks as cf
cf.set_config_file(offline=True, world_readable=True)
df=cf.datagen.ohlcv()
qf=cf.QuantFig(df,title='cufflinks金融繪圖樣例',legend='top',name='QF')
qf.add_ptps()
qf.iplot()

錯誤最終指向：

提示說是計算過程中出現了None。
初步考察，應該是cufflinks的源碼的問題(如果不是還請大佬指出）
打開cufflinks中的ta.py文件

可以看到，在286行，其將某一單元的初值設置爲None，而該單元恰好是在報錯的位置出現的變量。
解決方法：註釋掉286行。