In [2]:
from __future__ import print_function
from __future__ import division
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
%matplotlib inline
sns.set_context(rc={'figure.figsize': (14, 7) } )
figzize_me = figsize =(14, 7)
# import warnings; warnings.simplefilter('ignore')
import pandas as pd
pd.options.display.max_columns = 12
import os
import sys
# 使用insert 0即只使用github，避免交叉使用了pip安装的abupy，导致的版本不一致问题
sys.path.insert(0, os.path.abspath('../'))
import abupy
from abupy import xrange, range
# 打开测试数据环境，与书中的例子数据一致，使用RomDataBu下的df_kl.h5数据
abupy.env.enable_example_env_ipython()
In [2]:
print(sys.version)
In [3]:
import pandas as pd
In [4]:
stock_day_change = np.load('../gen/stock_day_change.npy')
stock_day_change.shape
Out[4]:
In [5]:
# 下面三种写法输出完全相同，输出如表4-1所示
pd.DataFrame(stock_day_change).head()
pd.DataFrame(stock_day_change).head(5)
pd.DataFrame(stock_day_change)[:5]
Out[5]:
In [6]:
# 股票0 -> 股票stock_day_change.shape[0]
stock_symbols = ['股票 ' + str(x) for x in
xrange(stock_day_change.shape[0])]
# 通过构造直接设置index参数，head(2)就显示两行，表4-2所示
pd.DataFrame(stock_day_change, index=stock_symbols).head(2)
Out[6]:
In [7]:
# 从2017-1-1向上时间递进，单位freq='1d'即1天
days = pd.date_range('2017-1-1',
periods=stock_day_change.shape[1], freq='1d')
# 股票0 -> 股票stock_day_change.shape[0]
stock_symbols = ['股票 ' + str(x) for x in
xrange(stock_day_change.shape[0])]
# 分别设置index和columns
df = pd.DataFrame(stock_day_change, index=stock_symbols, columns=days)
# 表4-3所示
df.head(2)
Out[7]:
In [8]:
# df做个转置
df = df.T
# 表4-4所示
df.head()
Out[8]:
In [9]:
# df_20 = df.resample('21D', how='mean')
from abupy import pd_resample
df_20 = pd_resample(df, '21D', how='mean')
# 表4-5所示
df_20.head()
Out[9]:
In [10]:
df_stock0 = df['股票 0']
# 打印df_stock0类型
print(type(df_stock0))
# 打印出Series的前5行数据, 与DataFrame一致
df_stock0.head()
Out[10]:
In [11]:
df_stock0.cumsum().plot()
Out[11]:
In [12]:
from abupy import pd_resample
# 以5天为周期重采样（周k）
# df_stock0_5 = df_stock0.cumsum().resample('5D', how='ohlc')
df_stock0_5 = pd_resample(df_stock0.cumsum(), '5D', how='ohlc')
# 以21天为周期重采样（月k），
# df_stock0_20 = df_stock0.cumsum().resample('21D', how='ohlc')
df_stock0_20 = pd_resample(df_stock0.cumsum(), '21D', how='ohlc')
# 打印5天重采样，如下输出2017-01-01, 2017-01-06, 2017-01-11, 表4-6所示
df_stock0_5.head()
Out[12]:
In [13]:
from abupy import ABuMarketDrawing
# 图4-2所示
ABuMarketDrawing.plot_candle_stick(df_stock0_5.index,
df_stock0_5['open'].values,
df_stock0_5['high'].values,
df_stock0_5['low'].values,
df_stock0_5['close'].values,
np.random.random(len(df_stock0_5)),
None, 'stock', day_sum=False,
html_bk=False, save=False)
In [14]:
print(type(df_stock0_5['open'].values))
print(df_stock0_5['open'].index)
print(df_stock0_5.columns)
In [15]:
from abupy import ABuSymbolPd
In [16]:
# n_folds=2两年
tsla_df = ABuSymbolPd.make_kl_df('usTSLA', n_folds=2)
# 表4-7所示
tsla_df.tail()
Out[16]:
In [17]:
tsla_df[['close', 'volume']].plot(subplots=True, style=['r', 'g'], grid=True);
In [18]:
tsla_df.info()
In [19]:
tsla_df.describe()
Out[19]:
In [20]:
# 2014-07-23至2014-07-31 开盘价格序列
tsla_df.loc['2014-07-23':'2014-07-31', 'open']
Out[20]:
In [21]:
# 2014-07-23至2014-07-31 所有序列，表4-9所示
tsla_df.loc['2014-07-23':'2014-07-31']
Out[21]:
In [22]:
# [1:5]：(1，2，3，4)，[2:6]: (2, 3, 4, 5)
# 表4-10所示
tsla_df.iloc[1:5, 2:6]
Out[22]:
In [23]:
# 切取所有行[2:6]: (2, 3, 4, 5)列
tsla_df.iloc[:, 2:6]
# 选取所有的列[35:37]:(35, 36)行，表4-11所示
tsla_df.iloc[35:37]
Out[23]:
In [24]:
# 指定一个列
print(tsla_df.close[0:3])
# 通过组成一个列表选择多个列，表4-12所示
tsla_df[['close', 'high', 'low']][0:3]
Out[24]:
In [25]:
# abs为取绝对值的意思，不是防抱死，表4-13所示
tsla_df[np.abs(tsla_df.p_change) > 8]
Out[25]:
In [26]:
tsla_df[(np.abs(tsla_df.p_change) > 8) & (tsla_df.volume > 2.5 * tsla_df.volume.mean())]
Out[26]:
In [27]:
# deprecated sort_index use sort_values
# tsla_df.sort_index(by='p_change')[:5]
tsla_df.sort_values(by='p_change')[:5]
Out[27]:
In [28]:
# deprecated sort_index use sort_values
# tsla_df.sort_index(by='p_change', ascending=False)[:5]
tsla_df.sort_values(by='p_change', ascending=False)[:5]
Out[28]:
In [29]:
# 如果一行的数据中存在na就删除这行
tsla_df.dropna()
# 通过how控制 如果一行的数据中全部都是na就删除这行
tsla_df.dropna(how='all')
# 使用指定值填充na， inplace代表就地操作，即不返回新的序列在原始序列上修改
tsla_df.fillna(tsla_df.mean(), inplace=True).head()
Out[29]:
In [30]:
tsla_df.close[:3]
Out[30]:
In [31]:
tsla_df.close.pct_change()[:3]
Out[31]:
In [32]:
(223.54 - 222.49) / 222.49, (223.57 - 223.54) / 223.54
Out[32]:
In [33]:
# pct_change对序列从第二项开始向前做减法在除以前一项，这样的针对close做pct_change后的结果就是涨跌幅
change_ratio = tsla_df.close.pct_change()
change_ratio.tail()
Out[33]:
In [34]:
# 将change_ratio转变成与tsla_df.p_change字段一样的百分百，同样保留两位小数
np.round(change_ratio[-5:] * 100, 2)
Out[34]:
In [36]:
format = lambda x: '%.2f' % x
tsla_df.atr21.map(format).tail()
Out[36]:
In [35]:
tsla_df.to_csv('../gen/tsla_df.csv', columns=tsla_df.columns, index=True)
In [36]:
tsla_df_load = pd.read_csv('../gen/tsla_df.csv', parse_dates=True, index_col=0)
tsla_df_load.head()
Out[36]:
In [37]:
tsla_df.p_change.hist(bins=80);
In [38]:
cats = pd.qcut(np.abs(tsla_df.p_change), 10)
cats.value_counts()
Out[38]:
In [39]:
# 将涨跌幅数据手工分类，从负无穷到－7，－5，－3，0， 3， 5， 7，正无穷
bins = [-np.inf, -7.0, -5, -3, 0, 3, 5, 7, np.inf]
cats = pd.cut(tsla_df.p_change, bins)
cats.value_counts()
Out[39]:
In [40]:
# cr_dummies为列名称前缀
change_ration_dummies = pd.get_dummies(cats, prefix='cr_dummies')
change_ration_dummies.head()
Out[40]:
In [41]:
pd.concat([tsla_df, change_ration_dummies], axis=1).tail()
Out[41]:
In [42]:
# pd.concat的连接axis＝0：纵向连接atr>14的df和p_change > 10的df
pd.concat([tsla_df[tsla_df.p_change > 10],
tsla_df[tsla_df.atr14 > 16]], axis=0)
# 直接使用DataFrame对象append，结果与上面pd.concat的结果一致, 表4-20所示
tsla_df[tsla_df.p_change > 10].append(
tsla_df[tsla_df.atr14 > 16])
Out[42]:
In [43]:
stock_a = pd.DataFrame({'stock_a': ['a', 'b', 'c', 'd', 'a'],
'data': list(range(5))})
stock_b = pd.DataFrame({'stock_b': ['a', 'b', 'c'],
'data2': list(range(3))})
pd.merge(stock_a, stock_b, left_on='stock_a', right_on='stock_b')
Out[43]:
In [44]:
tsla_df['positive'] = np.where(tsla_df.p_change > 0, 1, 0)
tsla_df.tail()
Out[44]:
In [45]:
xt = pd.crosstab(tsla_df.date_week, tsla_df.positive)
xt
Out[45]:
In [46]:
xt_pct = xt.div(xt.sum(1).astype(float), axis=0)
xt_pct
Out[46]:
In [47]:
xt_pct.plot(
figsize=(8, 5),
kind='bar',
stacked=True,
title='date_week -> positive')
plt.xlabel('date_week')
plt.ylabel('positive')
Out[47]:
In [48]:
tsla_df.pivot_table(['positive'], index=['date_week'])
Out[48]:
In [49]:
tsla_df.groupby(['date_week', 'positive'])['positive'].count()
Out[49]:
In [50]:
jump_threshold = tsla_df.close.median() * 0.03
jump_threshold
Out[50]:
In [51]:
jump_pd = pd.DataFrame()
def judge_jump(p_today):
global jump_pd
if p_today.p_change > 0 and (p_today.low - p_today.pre_close) > jump_threshold:
"""
符合向上跳空
"""
# jump记录方向 1向上
p_today['jump'] = 1
# 向上跳能量＝（今天最低 － 昨收）／ 跳空阀值
p_today['jump_power'] = (p_today.low - p_today.pre_close) / jump_threshold
jump_pd = jump_pd.append(p_today)
elif p_today.p_change < 0 and (p_today.pre_close - p_today.high) > jump_threshold:
"""
符合向下跳空
"""
# jump记录方向 －1向下
p_today['jump'] = -1
# 向下跳能量＝（昨收 － 今天最高）／ 跳空阀值
p_today['jump_power'] = (p_today.pre_close - p_today.high) / jump_threshold
jump_pd = jump_pd.append(p_today)
for kl_index in np.arange(0, tsla_df.shape[0]):
# 通过ix一个一个拿
today = tsla_df.ix[kl_index]
judge_jump(today)
# filter按照顺序只显示这些列, 表4-26所示
jump_pd.filter(['jump', 'jump_power', 'close', 'date', 'p_change', 'pre_close'])
Out[51]:
In [52]:
jump_pd = pd.DataFrame()
# axis=1即行数据，tsla_df的每一条行数据即为每一个交易日数据
tsla_df.apply(judge_jump, axis=1)
jump_pd
Out[52]:
In [53]:
from abupy import ABuMarketDrawing
# view_indexs传入jump_pd.index，即在k图上使用圆来标示跳空点
ABuMarketDrawing.plot_candle_form_klpd(tsla_df, view_indexs=jump_pd.index)
In [3]:
# disable_example_env_ipython不再使用沙盒数据，因为沙盒里面没有相关tsla行业的数据啊
abupy.env.disable_example_env_ipython()
In [59]:
abupy.env.g_data_fetch_mode = abupy.env.EMarketDataFetchMode.E_DATA_FETCH_FORCE_NET
In [4]:
from abupy import ABuIndustries
r_symbol = 'usTSLA'
# 这里获取了和TSLA电动车处于同一行业的股票组成pandas三维面板Panel数据
p_date, _ = ABuIndustries.get_industries_panel_from_target(r_symbol, show=False)
In [5]:
type(p_date)
Out[5]:
In [6]:
p_date
Out[6]:
In [7]:
p_date['usTTM'].head()
Out[7]:
In [8]:
p_data_it = p_date.swapaxes('items', 'minor')
p_data_it
Out[8]:
In [9]:
p_data_it_close = p_data_it['close'].dropna(axis=0)
p_data_it_close.tail()
Out[9]:
In [10]:
from abupy import ABuScalerUtil
# ABuScalerUtil.scaler_std将所有close的切面数据做(group - group.mean()) / group.std()标示化，为了可视化在同一范围
p_data_it_close = ABuScalerUtil.scaler_std(p_data_it_close)
p_data_it_close.plot()
plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)
plt.ylabel('Price')
plt.xlabel('Time')
Out[10]:
In [ ]:
In [ ]: