-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathPolicy_Test_Gui.py
370 lines (327 loc) · 19.8 KB
/
Policy_Test_Gui.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
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Sun Aug 28 14:21:38 2022
@author: Nigel Swenson
"""
import PySimpleGUI as sg
import os
import pickle as pkl
import numpy as np
import json
from PIL import ImageGrab
# from itertools import islice
import threading
from mojograsp.simcore.run_from_file import run_pybullet
import pathlib
'''
Data Plotter
This is based on the Demo_PNG_Viewer by PySimpleGUI
'''
def save_element_as_file(element, filename):
"""
Saves any element as an image file. Element needs to have an underlyiong Widget available (almost if not all of them do)
:param element: The element to save
:param filename: The filename to save to. The extension of the filename determines the format (jpg, png, gif, ?)
"""
widget = element.Widget
box = (widget.winfo_rootx(), widget.winfo_rooty(), widget.winfo_rootx() + widget.winfo_width(), widget.winfo_rooty() + widget.winfo_height())
grab = ImageGrab.grab(bbox=box)
grab.save(filename)
class RNNGui():
def __init__(self):
self.toggle_btn_off = b'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'
self.toggle_btn_on = b'iVBORw0KGgoAAAANSUhEUgAAACgAAAAoCAYAAACM/rhtAAAABmJLR0QA/wD/AP+gvaeTAAAD+UlEQVRYCe1XzW8bVRCffbvrtbP+2NhOD7GzLm1VoZaPhvwDnKBUKlVyqAQ3/gAkDlWgPeVQEUCtEOIP4AaHSI0CqBWCQyXOdQuRaEFOk3g3IMWO46+tvZ+PeZs6apq4ipON1MNafrvreTPzfvub92bGAOEnZCBkIGQgZOClZoDrh25y5pdjruleEiX+A+rCaQo05bpuvJ/+IHJCSJtwpAHA/e269g8W5RbuzF6o7OVjF8D3Pr4tSSkyjcqfptPDMDKSleW4DKIggIAD5Yf+Oo4DNg6jbUBlvWLUNutAwZu1GnDjzrcXzGcX2AHw/emFUV6Sfk0pqcKpEydkKSo9q3tkz91uF5aWlo1Gs/mYc+i7tz4//19vsW2AU9O381TiioVCQcnlRsWeQhD3bJyH1/MiFLICyBHiuzQsD1arDvypW7DR9nzZmq47q2W95prm+I9fXfqXCX2AF2d+GhI98Y8xVX0lnxvl2UQQg0csb78ag3NjEeD8lXZ7pRTgftmCu4864OGzrq+5ZU0rCa3m+NzXlzvoAoB3+M+SyWQuaHBTEzKMq/3BMbgM+FuFCDBd9kK5XI5PJBKqLSev+POTV29lKB8rT0yMD0WjUSYLZLxzNgZvIHODOHuATP72Vwc6nQ4Uiw8MUeBU4nHS5HA6TYMEl02wPRcZBJuv+ya+UCZOIBaLwfCwQi1Mc4QXhA+PjWRkXyOgC1uIhW5Qd8yG2TK7kSweLcRGKKVnMNExWWBDTQsH9qVmtmzjiThQDs4Qz/OUSGTwcLwIQTLW58i+yOjpXDLqn1tgmDzXzRCk9eDenjo9yhvBmlizrB3V5dDrNTuY0A7opdndStqmaQLPC1WCGfShYRgHdLe32UrV3ntiH9LliuNrsToNlD4kruN8v75eafnSgC6Luo2+B3fGKskilj5muV6pNhk2Qqg5v7lZ51nBZhNBjGrbxfI1+La5t2JCzfD8RF1HTBGJXyDzs1MblONulEqPDVYXgwDIfNx91IUVbAbY837GMur+/k/XZ75UWmJ77ou5mfM1/0x7vP1ls9XQdF2z9uNsPzosXPNFA5m0/EX72TBSiqsWzN8z/GZB08pWq9VeEZ+0bjKb7RTD2i1P4u6r+bwypo5tZUumEcDAmuC3W8ezIqSGfE6g/sTd1W5p5bKjaWubrmWd29Fu9TD0GlYlmTx+8tTJoZeqYe2BZC1/JEU+wQR5TVEUPptJy3Fs+Vkzgf8lemqHumP1AnYoMZSwsVEz6o26i/G9Lgitb+ZmLu/YZtshfn5FZDPBCcJFQRQ+8ih9DctOFvdLIKHH6uUQnq9yhFu0bec7znZ+xpAGmuqef5/wd8hAyEDIQMjAETHwP7nQl2WnYk4yAAAAAElFTkSuQmCC'
self.data_dict = {'train': {'state':[],'label':[]}, 'validation':{'state':[],'label':[]}, 'test': {'state':[],'label':[]}}
# define menu layout
self.menu = [['File', ['Open Folder', 'Exit']], ['Help', ['About', ]]]
self.args = {}
self.shuffle_type = 'Episode'
self.save_path = '/'
self.expert_path = '/'
self.load_path = '/'
self.built = False
self.train_dataset, self.validation_dataset = None, None
# define layout, show and read the window
data_layout = [ [sg.Text('Model Type'), sg.OptionMenu(values=('DDPG', 'DDPGFD','DDPG+HER', 'DDPGFD+HER'), k='-model', default_value='DDPG')],
[sg.Text('Path to Expert Data if using FD')],
[sg.Button("Browse",key='-browse-expert',button_color='DarkBlue'),sg.Text("/", key='-expert-path')],
[sg.Text('Path to Save Data')],
[sg.Button("Browse",key='-browse-save',button_color='DarkBlue'),sg.Text("/", key='-save-path')],
[sg.Text('Path to Previous Policy if Transferring')],
[sg.Button("Browse",key='-browse-load',button_color='DarkBlue'),sg.Text("/", key='-load-path')],
[sg.Text('Object'), sg.OptionMenu(values=('Cube', 'Cylinder'), k='-object', default_value='Cube')],
[sg.Text('Hand'), sg.OptionMenu(values=('2v2', '2v2-B'), k='-hand', default_value='2v2')],
[sg.Text("Task"), sg.OptionMenu(values=('asterisk','random','full_random'), k='-task', default_value='full_random')],
[sg.Text('Replay Buffer Sampling'), sg.OptionMenu(values=('priority','random','random+expert'), k='-sampling', default_value='priority')]]
model_layout = [ [sg.Text('Num Epochs'), sg.Input(10000, key='-epochs'), sg.Text('Batch Size'), sg.Input(100, key='-batch-size')],
[sg.Text('Learning Rate'), sg.Input(0.0001,key='-learning'), sg.Text('Discount Factor'), sg.Input(0.995, key='-df')],
[sg.Text('Starting Epsilon'), sg.Input(0.7,key='-epsilon'), sg.Text('Epsilon Decay Rate'), sg.Input(0.998, key='-edecay')],
[sg.Text('Rollout Size'), sg.Input(5,key='-rollout_size'), sg.Text('Rollout Weight'), sg.Input(0.5, key='-rollout_weight')],
[sg.Text('Evaluation Period'), sg.Input(3,key='-eval'), sg.Text('Tau'), sg.Input(0.0005, key='-tau')],
[sg.Text('Timesteps per Episode'), sg.Input(150,key='-tsteps'), sg.Text('Timesteps in Evaluation'), sg.Input(150,key='-eval-tsteps')],
[sg.Text('State Training Noise'), sg.Input(0.05, key='-snoise'),sg.Text('Start Pos Range (mm)'), sg.Input(0, key='-start-noise')]]
plotting_layout = [[sg.Text('Model Title')],
[sg.Input('test1',key='-title')],
[sg.Text("State")],
[sg.Checkbox('Finger Tip Position', default=True, k='-ftp')],
[sg.Checkbox('Finger Base Position', default=False, k='-fbp')],
[sg.Checkbox('Finger Contact Position', default=False, k='-fcp')],
[sg.Checkbox('Joint Angle', default=False, k='-ja')],
[sg.Checkbox('Object Position', default=True, k='-op')],
[sg.Checkbox('Finger Object Distance', default=False, k='-fod')],
[sg.Checkbox('Finger Tip Angle',default=True, k='-fta')],
[sg.Checkbox('Goal Position',default=True, k='-gp')],
[sg.Text('Num Previous States'),sg.Input('0', k='-pv')],
[sg.Text("Reward"), sg.OptionMenu(values=('Sparse','Distance','Distance + Finger', 'Hinge Distance + Finger', 'Slope', 'Slope + Finger'), k='-reward',default_value='Distance + Finger'), sg.Text('Success Radius (mm)'), sg.Input(2, key='-sr'),],
[sg.Text("Distance Scale"), sg.Input(1,key='-distance_scale'), sg.Text('Contact Scale'), sg.Input(0.2,key='-contact_scale')],
[sg.Text("Action"), sg.OptionMenu(values=('Joint Velocity','Finger Tip Position'), k='-action',default_value='Finger Tip Position')],
[sg.Checkbox('Vizualize Simulation',default=False, k='-viz'), sg.Checkbox('Real World?',default=False, k='-rw')],
[sg.Button('Begin Training', key='-train', bind_return_key=True)],
[sg.Button('Build Config File WITHOUT Training', key='-build')],
[sg.Text('Work progress'), sg.ProgressBar(100, size=(20, 20), orientation='h', key='-PROG-')]]
layout = [[sg.Menu(menu)], [sg.(col)]]
self.data_type = None
self.window = sg.Window('RNN Gui', layout, return_keyboard_events=True, use_default_focus=False, finalize=True)
def build_args(self, values):
RW = bool(values['-rw'])
self.built = False
print('building RL arglist, real world setting: ',RW)
self.args = {'epochs': int(values['-epochs']),
'batch_size': int(values['-batch-size']),
'model': values['-model'],
'learning_rate': float(values['-learning']),
'discount': float(values['-df']),
'epsilon': float(values['-epsilon']),
'edecay': float(values['-edecay']),
'object': values['-object'],
'hand': values['-hand'],
'task': values['-task'],
'evaluate': int(values['-eval']),
'sampling': values['-sampling'],
'reward': values['-reward'],
'action': values['-action'],
'rollout_size': int(values['-rollout_size']),
'rollout_weight': float(values['-rollout_weight']),
'tau': float(values['-tau']),
'pv': int(values['-pv']),
'viz': int(values['-viz']),
'sr': int(values['-sr']),
'state_noise': float(values['-snoise']),
'start_noise': float(values['-start-noise']),
'tsteps': int(values['-tsteps']),
'eval-tsteps':int(values['-eval-tsteps']),
'distance_scaling': float(values['-distance_scale']),
'contact_scaling': float(values['-contact_scale'])}
state_len = 0
state_mins = []
state_maxes = []
state_list = []
if values['-ftp']:
if not RW:
state_mins.extend([-0.072, 0.018, -0.072, 0.018])
state_maxes.extend([0.072, 0.172, 0.072, 0.172])
elif RW:
state_mins.extend([-0.108, 0.132, -0.108, 0.132])
state_maxes.extend([0.108, 0.348, 0.108, 0.348])
state_len += 4
state_list.append('ftp')
if values['-fbp']:
if not RW:
state_mins.extend([-0.072, 0.018, -0.072, 0.018])
state_maxes.extend([0.072, 0.172, 0.072, 0.172])
elif RW:
state_mins.extend([-0.108, 0.132, -0.108, 0.132])
state_maxes.extend([0.108, 0.348, 0.108, 0.348])
state_len += 4
state_list.append('fbp')
if values['-fcp']:
if not RW:
state_mins.extend([-0.072, 0.018, -0.072, 0.018])
state_maxes.extend([0.072, 0.172, 0.072, 0.172])
elif RW:
state_mins.extend([-0.108, 0.132, -0.108, 0.132])
state_maxes.extend([0.108, 0.348, 0.108, 0.348])
state_len += 4
state_list.append('fcp')
if values['-op']:
if not RW:
state_mins.extend([-0.072, 0.018])
state_maxes.extend([0.072, 0.172])
elif RW:
state_mins.extend([-0.108, 0.132])
state_maxes.extend([0.108, 0.348])
state_len += 2
state_list.append('op')
if values['-ja']:
state_mins.extend([-np.pi/2, -np.pi, -np.pi/2, 0])
state_maxes.extend([np.pi/2, 0, np.pi/2, np.pi])
state_len += 4
state_list.append('ja')
if values['-fod']:
if not RW:
state_mins.extend([-0.001, -0.001])
state_maxes.extend([0.072, 0.072])
elif RW:
state_mins.extend([-0.001, -0.001])
state_maxes.extend([0.108, 0.108])
state_len += 2
state_list.append('fod')
if values['-fta']:
state_mins.extend([-np.pi/2, -np.pi/2])
state_maxes.extend([np.pi/2, np.pi/2])
state_len += 2
state_list.append('fta')
if values['-gp']:
if not RW:
state_mins.extend([-0.07, -0.07])
state_maxes.extend([0.07, 0.07])
elif RW:
state_mins.extend([-0.105, -0.105])
state_maxes.extend([0.105, 0.105])
state_len += 2
state_list.append('gp')
if self.args['pv'] > 0:
state_len += state_len * self.args['pv']
temp_mins = state_mins.copy()
temp_maxes = state_maxes.copy()
for i in range(self.args['pv']):
state_mins.extend(temp_mins)
state_maxes.extend(temp_maxes)
if state_len == 0:
print('No selected state space')
return False
if (self.args['task'] == 'asterisk') or (self.args['task'] == 'random'):
if not values['-gp']:
print('Goal position needed for multigoal tasks')
return False
self.args['state_dim'] = state_len
self.args['state_mins'] = state_mins
self.args['state_maxes'] = state_maxes
self.args['state_list'] = state_list
if self.args['action'] =='Joint Velocity' or self.args['action'] =='Finger Tip Position':
self.args['action_dim'] = 4
if 'FD' in self.args['model']:
exists = os.path.isfile(self.expert_path + 'episode_all.pkl')
if not exists:
print('Selected FD model but no expert data loaded')
return False
else:
self.args['edata'] = values['-browse-expert'] + 'episode_all.pkl'
if os.path.isdir(self.save_path) and self.save_path != '/':
self.args['save_path'] = self.save_path + '/'
self.args['load_path'] = self.load_path + '/'
else:
print('save path is not a valid directory')
return False
overall_path = pathlib.Path(__file__).parent.resolve()
resource_path = overall_path.joinpath('demos/rl_demo/resources')
run_path = overall_path.joinpath('demos/rl_demo/runs')
self.args['tname'] = str(run_path.joinpath(values['-title']))
if values['-hand'] == '2v2':
self.args['hand_path'] = str(resource_path.joinpath('2v2_Hand_A/hand/2v2_50.50_50.50_1.1_53.urdf'))
elif values['-hand'] == '2v2-B':
self.args['hand_path'] = str(resource_path.joinpath('2v2_Hand_B/hand/2v2_65.35_65.35_1.1_53.urdf'))
if values['-object'] == 'Cube':
self.args['object_path'] = str(resource_path.joinpath('object_models/2v2_mod/2v2_mod_cuboid_small.urdf'))
elif values['-object'] == 'Cylinder':
self.args['object_path'] = str(resource_path.joinpath('resources/object_models/2v2_mod/2v2_mod_cylinder_small_alt.urdf'))
if values['-action'] == 'Joint Velocity':
self.args['max_action'] = 1.57
elif values['-action'] == 'Finger Tip Position':
self.args['max_action'] = 0.01
if values['-task'] == 'full_random':
self.args['points_path'] = str(resource_path.joinpath('points.csv'))
else:
self.args['points_path'] = str(resource_path.joinpath('train_points.csv'))
self.built = True
return True
def train(self):
run_pybullet(self.args['save_path'] + 'experiment_config.json', self.window)
print('model finished, saving now')
def log_params(self):
if self.built:
print('saving configuration')
with open(self.args['save_path'] + '/experiment_config.json', 'w') as conf_file:
json.dump(self.args, conf_file, indent=4)
try:
os.mkdir(self.args['save_path'] + '/Train/')
except FileExistsError:
pass
try:
os.mkdir(self.args['save_path'] + '/Test/')
except FileExistsError:
pass
else:
print('config not built, parameters not saved')
def run_gui(self):
while True:
event, values = self.window.read()
# print(values.keys())
# --------------------- Button & Keyboard ---------------------
if event == sg.WIN_CLOSED:
break
elif event == 'shuffle-type':
self.shuffle_type = values['shuffle-type']
elif event == '-load-model':
newfolder = sg.popup_get_file('Select Model File', no_window=True)
if newfolder is None:
continue
if newfolder.lower().endswith('.pt'):
self.model_path = newfolder
elif event == 'Exit':
break
# ----------------- Menu choices -----------------
if event == '-browse-expert':
newfolder = sg.popup_get_folder('Select Folder Containing Expert Data', no_window=True)
if newfolder is None:
continue
folder = newfolder
print(type(folder))
self.expert_path = folder
self.window.refresh()
elif event == '-browse-save':
newfolder = sg.popup_get_folder('Select Folder To Save Data In', no_window=True)
if newfolder is None:
continue
folder = newfolder
print(type(folder))
self.save_path = folder
self.window.refresh()
elif event == '-browse-load':
newfolder = sg.popup_get_folder('Select Folder To Save Data In', no_window=True)
if newfolder is None:
continue
folder = newfolder
print(type(folder))
self.load_path = folder
self.window.refresh()
elif event == '-train':
ready = self.build_args(values)
if ready:
self.log_params()
thread = threading.Thread(target=self.train, daemon=True)
thread.start()
else:
print('Parameters incorrect, cant start training')
elif event == '-build':
ready = self.build_args(values)
if ready:
self.log_params()
print('Build Successful')
else:
print('Build Not Successful')
# elif event == '-update':
# print(values['-update'])
elif event == 'About':
sg.popup('Why you click me?',
'Go harrass Nigel with questions. swensoni@oregonstate.edu')
self.window['-save-path'].update(self.save_path)
self.window['-expert-path'].update(self.expert_path)
self.window['-load-path'].update(self.load_path)
self.window.close()
def main():
backend = RNNGui()
backend.run_gui()
if __name__ == '__main__':
main()