-
Notifications
You must be signed in to change notification settings - Fork 15
/
dc.py
323 lines (244 loc) · 11.9 KB
/
dc.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
import matplotlib.animation as animation
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from pykalman import KalmanFilter
class dc_calculator():
def __init__(self):
self.prices = None
self.time = None
self.TMV_list = []
self.T_list = []
self.colors = []
self.events = []
def compute_dc_variables(self, threshold: float = 0.0001):
"""
Method to compute all relevant DC parameters.
"""
if self.prices is None:
print('Please load the time series data first before proceeding with the DC parameters computation')
else:
self.TMV_list = []
self.T_list = []
self.colors = []
self.events = []
ext_point_n = self.prices[0]
curr_event_max = self.prices[0]
curr_event_min = self.prices[0]
time_point_max = 0
time_point_min = 0
trend_status = 'up'
T = 0
for i in range(len(self.prices)):
TMV = (self.prices[i] - ext_point_n) / (ext_point_n * threshold)
self.TMV_list.append(TMV)
self.T_list.append(T)
T += 1
if trend_status == 'up':
self.colors.append('lime')
self.events.append('Upward Overshoot')
if self.prices[i] < ((1 - threshold) * curr_event_max):
trend_status = 'down'
curr_event_min = self.prices[i]
ext_point_n = curr_event_max
T = i - time_point_max
num_points_change = i - time_point_max
for j in range(1, num_points_change + 1):
self.colors[-j] = 'red'
self.events[-j] = 'Downward DCC'
else:
if self.prices[i] > curr_event_max:
curr_event_max = self.prices[i]
time_point_max = i
else:
self.colors.append('lightcoral')
self.events.append('Downward Overshoot')
if self.prices[i] > ((1 + threshold) * curr_event_min):
trend_status = 'up'
curr_event_max = self.prices[i]
ext_point_n = curr_event_min
T = i - time_point_min
num_points_change = i - time_point_min
for j in range(1, num_points_change + 1):
self.colors[-j] = 'green'
self.events[-j] = 'Upward DCC'
else:
if self.prices[i] < curr_event_min:
curr_event_min = self.prices[i]
time_point_min = i
self.colors = np.array(self.colors)
print('DC variables computation has finished.')
def generate_event_data(self, output_csv_name: str):
"""
Method to write the detected events for each
point in the time series data to a .csv file
for external analysis. Again, feel free to
modify this method according to your needs.
"""
if isinstance(self.colors, list):
print('Please load the time series data and compute DC variables first before attempting to generate the event data.')
else:
df = pd.DataFrame({'Time': self.time, 'Rate': self.prices, 'Event': np.array(self.events)})
df.to_csv(output_csv_name + '.csv', index = False)
print("The event data file '" + output_csv_name + ".csv' has been generated.")
def generate_indicator_space_plot(self, title: str, output_plot_name = None):
"""
Method to generate the normalized TMV against normalized T indicator-
space plot after the DC parameters have been computed. This method
currently uses the detected event class to colour the points, and it
should be modified accordingly if one wishes to colour the points
according to the detected regime as described in the book.
"""
if len(self.TMV_list) == 0:
print('Please load the time series data and compute the DC variables before attempting to plot the indicator feature space.')
else:
TMV_array = np.array(self.TMV_list)
T_array = np.array(self.T_list)
norm_TMV = (TMV_array - np.min(TMV_array)) / (np.max(TMV_array) - np.min(TMV_array))
norm_T = (T_array - np.min(T_array)) / (np.max(T_array) - np.min(T_array))
fig, ax = plt.subplots()
ax.scatter(norm_T, norm_TMV, c = self.colors, edgecolors = 'k')
legend_elements = [plt.Line2D([0], [0], marker = 'o', color = 'w', markerfacecolor = 'green', markersize = 5, label = 'Upward DCC Event'),
plt.Line2D([0], [0], marker = 'o', color = 'w', markerfacecolor = 'lime', markersize = 5, label = 'Upward Overshoot Event'),
plt.Line2D([0], [0], marker = 'o', color = 'w', markerfacecolor = 'red', markersize = 5, label = 'Downward DCC Event'),
plt.Line2D([0], [0], marker = 'o', color = 'w', markerfacecolor = 'lightcoral', markersize = 5, label = 'Downward Overshoot Event')]
ax.legend(handles = legend_elements, loc = 'upper right', fontsize = 'small')
ax.set_xlim(np.min(norm_T) - 0.1, np.max(norm_T) + 0.1)
ax.set_ylim(np.min(norm_TMV) - 0.1, np.max(norm_TMV) + 0.1)
ax.set_title(title)
ax.set_xlabel('Normalized T')
ax.set_ylabel('Normalized TMV')
if output_plot_name is not None and isinstance(output_plot_name, str):
plt.savefig(output_plot_name + '.jpg')
else:
plt.show()
print("The indicator feature space plot '" + title + "' has been generated.")
def generate_original_time_series_plot(self, title: str, x_axis_label: str, y_axis_label: str, output_plot_name = None):
"""
Method to plot out the original time series data with no event
annotation for visualization purpose. If output_plot_name is
supplied, then it is assumed that the plot should be written to
file, and the corresponding plot .jpg file will be created.
"""
if self.prices is None:
print('Please load the time series data first before plotting the original time series data.')
else:
fig2, ax2 = plt.subplots()
ax2.ticklabel_format(style = 'plain', axis = 'y', useOffset = False)
for i in range(len(self.prices)):
ax2.plot(self.time[i : i + 2], self.prices[i : i + 2], color = 'black')
ax2.set_xlim(0, len(self.prices) - 1)
ax2.set_ylim(self.prices.min() * 0.9999, self.prices.max() * 1.0001)
ax2.set_title(title)
ax2.set_xlabel(x_axis_label)
ax2.set_ylabel(y_axis_label)
if output_plot_name is not None and isinstance(output_plot_name, str):
plt.savefig(output_plot_name + '.jpg')
else:
plt.show()
print("The original time series plot '" + title + "' has been generated.")
def generate_time_series_animation(self, title: str, x_axis_label: str, y_axis_label: str, fps: int = 60, output_gif_name = None):
"""
Method to generate the event-annotated time series animation GIF after the DC parameters have been computed.
"""
if isinstance(self.colors, list):
print('Please load the time series data and compute DC variables first before attempting to generate the time series animation.')
else:
green_patch = plt.Line2D([0], [0], color = 'green', label = 'Upward DCC Event')
lime_patch = plt.Line2D([0], [0], color = 'lime', label = 'Upward Overshoot Event')
red_patch = plt.Line2D([0], [0], color = 'red', label = 'Downward DCC Event')
lightcoral_patch = plt.Line2D([0], [0], color = 'lightcoral', label = 'Downward Overshoot Event')
fig, ax = plt.subplots()
ax.ticklabel_format(style = 'plain', axis = 'y', useOffset = False)
lines = [ax.plot([], [], color = color)[0] for color in self.colors]
ax.set_xlim(0, len(self.prices) - 1)
ax.set_ylim(self.prices.min() * 0.9999, self.prices.max() * 1.0001)
ax.set_title(title)
ax.set_xlabel(x_axis_label)
ax.set_ylabel(y_axis_label)
ax.legend(handles = [green_patch, lime_patch, red_patch, lightcoral_patch], loc = 'upper right', fontsize = 'small')
def init():
for line in lines:
line.set_data([], [])
return lines
def update(i):
if i == 0:
return lines
else:
lines[i - 1].set_data(self.time[i - 1 : i + 1], self.prices[i - 1 : i + 1])
return lines
ani = animation.FuncAnimation(fig, update, frames = len(self.prices), init_func = init, blit = True)
if output_gif_name is not None and isinstance(output_gif_name, str):
ani.save(output_gif_name + '.gif', writer = 'pillow', fps = fps)
else:
plt.show()
print("The animation '" + title + "' has been generated.")
def generate_time_series_plot(self, title: str, x_axis_label: str, y_axis_label: str, output_plot_name = None):
"""
Method to generate the event-annotated time series plot after the DC parameters have been computed.
"""
if isinstance(self.colors, list):
print('Please load the time series data and compute DC variables first before attempting to generate the time series plot.')
else:
green_patch = plt.Line2D([0], [0], color = 'green', label = 'Upward DCC Event')
lime_patch = plt.Line2D([0], [0], color = 'lime', label = 'Upward Overshoot Event')
red_patch = plt.Line2D([0], [0], color = 'red', label = 'Downward DCC Event')
lightcoral_patch = plt.Line2D([0], [0], color = 'lightcoral', label = 'Downward Overshoot Event')
fig1, ax1 = plt.subplots()
ax1.ticklabel_format(style = 'plain', axis = 'y', useOffset = False)
for i, color in enumerate(self.colors):
ax1.plot(self.time[i : i + 2], self.prices[i : i + 2], color = color)
ax1.set_xlim(0, len(self.prices) - 1)
ax1.set_ylim(self.prices.min() * 0.9999, self.prices.max() * 1.0001)
ax1.set_title(title)
ax1.set_xlabel(x_axis_label)
ax1.set_ylabel(y_axis_label)
ax1.legend(handles = [green_patch, lime_patch, red_patch, lightcoral_patch], loc = 'upper right', fontsize = 'small')
if output_plot_name is not None and isinstance(output_plot_name, str):
plt.savefig(output_plot_name + '.jpg')
else:
plt.show()
print("The plot '" + title + "' has been generated.")
def load_time_series_data_from_file(self, file: str, data_point_limit = None, kalman_filter: bool = False):
"""
Method to load the time series data from a file. Please modify this method to suit your input data format.
The default method assumes that the data is stored in the format of a .csv file of which the columns
are similar to that of the historical data provided by TrueFX (visit
https://www.truefx.com/truefx-historical-downloads/ for more details). In addition, this method also
provides an option to use the kalman filter on the time series data to smoothen it first before proceeding
with the DC parameters computation
"""
try:
df = pd.read_csv(file, header = None, names = ['Currency', 'Date', 'Short', 'Long'])
self.prices = df['Long'].to_numpy()
if data_point_limit is not None and data_point_limit < len(self.prices):
self.prices = self.prices[:data_point_limit]
if kalman_filter:
kf = KalmanFilter(initial_state_mean = 0, n_dim_obs = 1)
kf = kf.em(self.prices, n_iter = 10)
(self.prices, _) = kf.smooth(self.prices)
self.time = np.arange(len(self.prices))
print("The time series data file '" + file + "' is loaded and ready for processing.")
except:
print('Please input a valid file name, and ensure that the file contains the data in an appropriate format.')
if __name__ == '__main__':
calc = dc_calculator()
# Example use of the methods of the dc_calculator class object
calc.load_time_series_data_from_file('Truncated EURUSD-2023-09.csv', data_point_limit = 1000)
calc.generate_original_time_series_plot('EUR-USD September 2023 Time Series',
'Time', 'EUR-USD',
output_plot_name = 'EURUSD-2023-09 Time Series Plot'
)
calc.compute_dc_variables()
calc.generate_indicator_space_plot('EUR-USD September 2023 Indicator Feature Space',
output_plot_name = 'EURUSD-2023-09 Indicator Feature Space Plot'
)
calc.generate_event_data('EURUSD-2023-09 Event Data')
calc.generate_time_series_plot('Annotated EUR-USD September 2023 Time Series',
'Time', 'EUR-USD',
output_plot_name = 'Annotated EURUSD-2023-09 Plot'
)
calc.generate_time_series_animation('Annotated EUR-USD September 2023 Time Series Animation',
'Time', 'EUR-USD',
output_gif_name = 'Annotated EURUSD-2023-09 Animation'
)