forked from beef-broccoli/deebo
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathchem_simulate.py
460 lines (395 loc) · 20.8 KB
/
chem_simulate.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
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
from chem_arms import simulate_propose_and_update, simulate_propose_and_update_interpolation
import algos_regret
import utils
import pathlib
import pandas as pd
import numpy as np
"""Functions to run simulations on different datasets"""
def deoxyf():
# fetch ground truth data
ground_truth = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/deoxyf.csv')
ground_truth['yield'] = ground_truth['yield'].apply(utils.scaler)
ground_truth = ground_truth[['base_name',
'fluoride_name',
'substrate_name',
'yield']]
bases = ground_truth['base_name'].unique()
fluorides = ground_truth['fluoride_name'].unique()
substrates = ground_truth['substrate_name'].unique()
#######################################################################################################################
# build dictionary for acquisition
scope_dict = {'base_name': bases,
'fluoride_name': fluorides,
'substrate_name': substrates}
arms_dict = {'base_name': bases,
'fluoride_name': fluorides}
#algo = algos_regret.ThompsonSamplingGaussianFixedVar(len(bases)*len(fluorides), assumed_sd=0.25)
algo = algos_regret.UCB1Tuned(len(bases)*len(fluorides))
wkdir = './dataset_logs/deoxyf/combo/'
num_sims = 400
num_round = 200
num_exp = 1
propose_mode = 'random'
#######################################################################################################################
dir_name = f'{wkdir}{algo.__str__()}-{num_sims}s-{num_round}r-{num_exp}e/'
p = pathlib.Path(dir_name)
p.mkdir(parents=True)
simulate_propose_and_update(scope_dict, arms_dict, ground_truth, algo,
dir=dir_name, num_sims=num_sims,
num_round=num_round, num_exp=num_exp, propose_mode=propose_mode)
def deoxyf_adversarial(): #TODO
# fetch ground truth data
ground_truth = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/deoxyf.csv')
seg1 = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/deoxyf-seg1.csv')
seg2 = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/deoxyf-seg2.csv')
seg3 = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/deoxyf-seg3.csv')
ground_truth['yield'] = ground_truth['yield'].apply(utils.scaler)
ground_truth = ground_truth[['base_name',
'fluoride_name',
'substrate_name',
'yield']]
bases = ground_truth['base_name'].unique()
fluorides = ground_truth['fluoride_name'].unique()
substrates_all = ground_truth['substrate_name'].unique()
substrates_1 = seg1['substrate_name'].unique()
substrates_2 = seg2['substrate_name'].unique()
substrates_3 = seg3['substrate_name'].unique()
#######################################################################################################################
# build dictionary for acquisition
scope_dict = {'base_name': bases,
'fluoride_name': fluorides,
'substrate_name': substrates_1} # substrate_1 to start
expansion_dict = {50: {'substrate_name': substrates_2},
100: {'substrate_name': substrates_3}}
arms_dict = {'base_name': bases,
'fluoride_name': fluorides,}
algo = algos_regret.BayesUCBGaussian(n_arms=20)
wkdir = './dataset_logs/deoxyf/adversarial/combo/'
num_sims = 400
num_round = 150
num_exp = 1
propose_mode = 'random'
#######################################################################################################################
dir_name = f'{wkdir}{algo.__str__()}-{num_sims}s-{num_round}r-{num_exp}e/'
p = pathlib.Path(dir_name)
p.mkdir(parents=True)
simulate_propose_and_update(scope_dict, arms_dict, ground_truth, algo,
dir=dir_name, num_sims=num_sims,
num_round=num_round, num_exp=num_exp, propose_mode=propose_mode,
expansion_dict=expansion_dict
)
def deoxyf_interpolation():
# fetch ground truth data
ground_truth = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/deoxyf.csv')
substrate_smiles_to_id = dict(zip(ground_truth['substrate_smiles'].unique(),
ground_truth['substrate_name'].unique()))
ground_truth['yield'] = ground_truth['yield'].apply(utils.scaler)
ground_truth = ground_truth[['base_name',
'fluoride_name',
'substrate_name',
'yield']]
bases = ground_truth['base_name'].unique()
fluorides = ground_truth['fluoride_name'].unique()
substrates = ground_truth['substrate_name'].unique()
# make encodings for substrates
encodings = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deoxyF/mols/ECFP/substrate.csv'
)
encodings.set_index('substrate_SMILES', inplace=True)
ordered_ids = [substrate_smiles_to_id[s] for s in encodings.index] # smiles to id, maintain order
encodings = dict(zip(ordered_ids, encodings.to_numpy().tolist())) # {id: [ECFP1, ECFP2, ...]}
encodings = {'substrate_name': encodings} # one more level to specify it's substrates
#######################################################################################################################
# build dictionary for acquisition
scope_dict = {'base_name': bases,
'fluoride_name': fluorides,
'substrate_name': substrates}
arms_dict = {'base_name': bases,
'fluoride_name': fluorides}
# algo = algos_regret.ThompsonSamplingGaussianFixedVar(len(bases)*len(fluorides), assumed_sd=0.25)
algo = algos_regret.UCB1Tuned(len(bases) * len(fluorides))
wkdir = './dataset_logs/deoxyf/combo/interpolation/'
num_sims = 200
num_round = 200
num_exp = 1
propose_mode = 'random'
batch_size = 50
#######################################################################################################################
dir_name = f'{wkdir}{algo.__str__()}-{num_sims}s-{num_round}r-{batch_size}b/'
p = pathlib.Path(dir_name)
p.mkdir(parents=True)
simulate_propose_and_update_interpolation(scope_dict, arms_dict, encodings, ground_truth, algo,
dir=dir_name, n_sims=num_sims, n_horizon=num_round,
batch_size=batch_size)
def nickel_borylation():
# first version:
# 60% cutoff for yield, EtOH, use the top ligands identified by z score in paper
# top 6 ligand is identical to the ones in the paper
# second version:
# 50% cutoff for yield, EtOH, use the top ligands identified by z score in paper
# top 3 and top 8 ligands are identical to the ones in the paper
top_three = ['Cy-JohnPhos', 'P(p-Anis)3', 'PPh2Cy']
top_eight = ['PPh2Cy', 'CX-PCy', 'PPh3', 'P(p-F-Ph)3', 'P(p-Anis)3', 'Cy-JohnPhos', 'A-paPhos', 'Cy-PhenCar-Phos']
# fetch ground truth data
ground_truth = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/nib-etoh.csv', index_col=0)
ground_truth['yield'] = ground_truth['yield'].apply(utils.cutoffer, args=(50,))
ligands = ground_truth['ligand_name'].unique()
electrophiles = ground_truth['electrophile_id'].unique()
#######################################################################################################################
# build dictionary for acquisition
scope_dict = {'electrophile_id': electrophiles,
'ligand_name': ligands,}
arms_dict = {'ligand_name': ligands}
# bayes ucb beta keeps picking the same arm with n_exp=100, but looks to be optimal
algos = [algos_regret.ThompsonSamplingBeta(len(ligands)),
algos_regret.ThompsonSamplingGaussianFixedVarSquared(len(ligands)),
algos_regret.UCB1Tuned(len(ligands)),
algos_regret.UCB1(len(ligands)),
algos_regret.BayesUCBGaussianSquared(len(ligands)),
algos_regret.BayesUCBBeta(len(ligands)),
algos_regret.BayesUCBBetaPPF(len(ligands)),
algos_regret.AnnealingEpsilonGreedy(len(ligands)),
algos_regret.Random(len(ligands))
]
algos = [
algos_regret.Random(len(ligands))
]
wkdir = './dataset_logs/nib/etoh-50cutoff/'
num_sims = 500
num_round = 100
num_exp = 1
propose_mode = 'random'
#######################################################################################################################
for algo in algos:
dir_name = f'{wkdir}{algo.__str__()}-{num_sims}s-{num_round}r-{num_exp}e/'
p = pathlib.Path(dir_name)
p.mkdir(parents=True)
simulate_propose_and_update(scope_dict, arms_dict, ground_truth, algo,
dir=dir_name, num_sims=num_sims,
num_round=num_round, num_exp=num_exp, propose_mode=propose_mode,
)
return
def cn():
# fetch ground truth data
ground_truth = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/cn-processed.csv')
ground_truth['yield'] = ground_truth['yield'].apply(utils.scaler)
ground_truth = ground_truth[['base_name',
'ligand_name',
'substrate_id',
'additive_id',
'yield']]
bases = ground_truth['base_name'].unique()
ligands = ground_truth['ligand_name'].unique()
additives = ground_truth['additive_id'].unique()
substrates = ground_truth['substrate_id'].unique()
#######################################################################################################################
# build dictionary for acquisition
scope_dict = {'base_name': bases,
'ligand_name': ligands,
'additive_id': additives,
'substrate_id': substrates}
arms_dict = {'base_name': bases,
'ligand_name': ligands}
n_arms = len(bases)*len(ligands)
algos = [algos_regret.UCB1Tuned(n_arms),
algos_regret.UCB1(n_arms),
algos_regret.AnnealingEpsilonGreedy(n_arms),
algos_regret.ThompsonSamplingGaussianFixedVar(n_arms, assumed_sd=0.25),
algos_regret.ThompsonSamplingGaussianFixedVarSquared(n_arms),
algos_regret.BayesUCBGaussian(n_arms, assumed_sd=0.25, c=2),
algos_regret.BayesUCBGaussianSquared(n_arms, c=2),
algos_regret.Random(n_arms)]
# algo = algos_regret.ThompsonSamplingGaussianFixedVar(len(bases)*len(fluorides), assumed_sd=0.25)
wkdir = './dataset_logs/cn/'
num_sims = 500
num_round = 100
num_exp = 1
propose_mode = 'random'
#######################################################################################################################
for algo in algos:
dir_name = f'{wkdir}{algo.__str__()}-{num_sims}s-{num_round}r-{num_exp}e/'
p = pathlib.Path(dir_name)
p.mkdir(parents=True)
simulate_propose_and_update(scope_dict, arms_dict, ground_truth, algo,
dir=dir_name, num_sims=num_sims,
num_round=num_round, num_exp=num_exp, propose_mode=propose_mode)
def arylation():
# fetch ground truth data
ground_truth = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/aryl-scope-ligand.csv')
ground_truth['yield'] = ground_truth['yield'].apply(utils.scaler)
ground_truth = ground_truth[['ligand_name',
'electrophile_id',
'nucleophile_id',
'yield']]
ligands = ground_truth['ligand_name'].unique()
elecs = ground_truth['electrophile_id'].unique()
nucs = ground_truth['nucleophile_id'].unique()
#######################################################################################################################
# build dictionary for acquisition
scope_dict = {'ligand_name': ligands,
'electrophile_id': elecs,
'nucleophile_id': nucs}
arms_dict = {'ligand_name': ligands}
n_arms = len(ligands)
# algos = [algos_regret.UCB1Tuned(n_arms),
# algos_regret.UCB1(n_arms),
# algos_regret.AnnealingEpsilonGreedy(n_arms),
# algos_regret.ThompsonSamplingGaussianFixedVar(n_arms, assumed_sd=0.25),
# algos_regret.ThompsonSamplingGaussianFixedVarSquared(n_arms),
# algos_regret.BayesUCBGaussian(n_arms, assumed_sd=0.25, c=2),
# algos_regret.BayesUCBGaussianSquared(n_arms, c=2),
# algos_regret.Random(n_arms)]
algos = [algos_regret.BayesUCBGaussianSquared(n_arms, c=2),
algos_regret.Random(n_arms)]
# algo = algos_regret.ThompsonSamplingGaussianFixedVar(len(bases)*len(fluorides), assumed_sd=0.25)
wkdir = './dataset_logs/aryl-scope-ligand/'
num_sims = 500
num_round = 200
num_exp = 1
propose_mode = 'random'
#######################################################################################################################
for algo in algos:
dir_name = f'{wkdir}{algo.__str__()}-{num_sims}s-{num_round}r-{num_exp}e/'
p = pathlib.Path(dir_name)
p.mkdir(parents=True)
simulate_propose_and_update(scope_dict, arms_dict, ground_truth, algo,
dir=dir_name, num_sims=num_sims,
num_round=num_round, num_exp=num_exp, propose_mode=propose_mode,
predict=False)
def arylation_expansion():
# fetch ground truth data
ground_truth = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/aryl-scope-ligand.csv')
ground_truth['yield'] = ground_truth['yield'].apply(utils.scaler)
ground_truth = ground_truth[['ligand_name',
'electrophile_id',
'nucleophile_id',
'yield']]
ligands = ground_truth['ligand_name'].unique()
elecs = ground_truth['electrophile_id'].unique()
nucs = ground_truth['nucleophile_id'].unique()
#######################################################################################################################
scope_dict = {'nucleophile_id': ['nA', 'nB', 'nC', 'nD'],
'electrophile_id': ['e1', 'e2', 'e3', 'e4'],
'ligand_name': ligands}
expansion_dict = {50: {'nucleophile_id': ['nE', 'nF', 'nG', 'nI']},
100: {'electrophile_id': ['e5', 'e7', 'e9', 'e10']}}
arms_dict = {'ligand_name': ligands}
n_arms = len(ligands)
algos = [algos_regret.UCB1Tuned(n_arms),
algos_regret.UCB1(n_arms)]
wkdir = './dataset_logs/aryl-scope-ligand/expansion/scenario1/'
num_sims = 500
num_round = 300
num_exp = 1
propose_mode = 'random'
#######################################################################################################################
for algo in algos:
dir_name = f'{wkdir}{algo.__str__()}-{num_sims}s-{num_round}r-{num_exp}e/'
p = pathlib.Path(dir_name)
p.mkdir(parents=True)
simulate_propose_and_update(scope_dict, arms_dict, ground_truth, algo,
dir=dir_name, num_sims=num_sims,
num_round=num_round, num_exp=num_exp, propose_mode=propose_mode,
expansion_dict=expansion_dict)
def amidation():
# fetch ground truth data
ground_truth = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/amidation.csv')
ground_truth['yield'] = ground_truth['yield'].apply(utils.scaler)
short_name_dict = {
'1-Methylimidazole': 'MeIm',
'2,6-Lutidine': 'lutidine',
'N-methylmorpholine': 'MeMorph',
'Diisopropylethylamine': 'DIPEA'
}
nuc_id_dict = dict(zip(ground_truth['nucleophile_name'].unique(),
[f'n{num}' for num in np.arange(len(ground_truth['nucleophile_name'].unique())) + 1]))
ground_truth['nucleophile_id'] = ground_truth['nucleophile_name'].apply(lambda x: nuc_id_dict[x])
ground_truth = ground_truth[['nucleophile_id', 'base_name', 'activator_name', 'solvent_name','yield']]
bases = ground_truth['base_name'].unique()
activators = ground_truth['activator_name'].unique()
solvents = ground_truth['solvent_name'].unique()
nucs = ground_truth['nucleophile_id'].unique()
#######################################################################################################################
# build dictionary for acquisition
scope_dict = {'nucleophile_id': nucs,
'base_name': bases,
'activator_name': activators,
'solvent_name': solvents}
arms_dict = {'activator_name': activators,
'base_name': bases,}
n_arms = len(activators)*len(bases)
algos = [algos_regret.UCB1Tuned(n_arms),
algos_regret.UCB1(n_arms),
algos_regret.AnnealingEpsilonGreedy(n_arms),
algos_regret.ThompsonSamplingGaussianFixedVar(n_arms, assumed_sd=0.25),
algos_regret.ThompsonSamplingGaussianFixedVarSquared(n_arms),
algos_regret.BayesUCBGaussian(n_arms, assumed_sd=0.25, c=2),
algos_regret.BayesUCBGaussianSquared(n_arms, c=2),
algos_regret.Random(n_arms)]
wkdir = './dataset_logs/amidation/combo/'
num_sims = 500
num_round = 96
num_exp = 1
propose_mode = 'random'
#######################################################################################################################
for algo in algos:
dir_name = f'{wkdir}{algo.__str__()}-{num_sims}s-{num_round}r-{num_exp}e/'
p = pathlib.Path(dir_name)
p.mkdir(parents=True)
simulate_propose_and_update(scope_dict, arms_dict, ground_truth, algo,
dir=dir_name, num_sims=num_sims,
num_round=num_round, num_exp=num_exp, propose_mode=propose_mode,
predict=False)
def cn_maldi():
# fetch ground truth data
ground_truth = pd.read_csv(
'https://raw.githubusercontent.com/beef-broccoli/ochem-data/main/deebo/maldi-amine.csv')
eic_max = ground_truth['EIC(+)[M+H] Product Area'].max()
ground_truth['yield'] = ground_truth['EIC(+)[M+H] Product Area'] / eic_max # being naughty here, "yield"
ground_truth = ground_truth[['substrate_id',
'condition',
'yield',]]
conditions = ground_truth['condition'].unique()
substrates = ground_truth['substrate_id'].unique()
#######################################################################################################################
# build dictionary for acquisition
scope_dict = {'condition': conditions,
'substrate_id': substrates}
arms_dict = {'condition': conditions}
n_arms = len(conditions)
# algos = [algos_regret.UCB1Tuned(n_arms),
# algos_regret.UCB1(n_arms),
# algos_regret.AnnealingEpsilonGreedy(n_arms),
# algos_regret.ThompsonSamplingGaussianFixedVar(n_arms, assumed_sd=0.25),
# algos_regret.ThompsonSamplingGaussianFixedVarSquared(n_arms),
# algos_regret.BayesUCBGaussian(n_arms, assumed_sd=0.25, c=2),
# algos_regret.BayesUCBGaussianSquared(n_arms, c=2),
# algos_regret.Random(n_arms)]
algos = [algos_regret.Random(n_arms)]
# algo = algos_regret.ThompsonSamplingGaussianFixedVar(len(bases)*len(fluorides), assumed_sd=0.25)
wkdir = './test/merck-maldi/amine/'
num_sims = 1000
num_round = 200
num_exp = 1
propose_mode = 'random'
#######################################################################################################################
for algo in algos:
dir_name = f'{wkdir}{algo.__str__()}-{num_sims}s-{num_round}r-{num_exp}e/'
p = pathlib.Path(dir_name)
p.mkdir(parents=True)
simulate_propose_and_update(scope_dict, arms_dict, ground_truth, algo,
dir=dir_name, num_sims=num_sims,
num_round=num_round, num_exp=num_exp, propose_mode=propose_mode)
if __name__ == '__main__':
cn_maldi()