def generate_optimized_molecules(self,
scoring_function,
number_molecules,
starting_population=None):
print("generating %s samples..." % number_molecules)
all_smiles = set()
width = 24
max_depth = 100
c = 5
num_simulations = 10000
self.best_smiles = None
self.best_score = -1.0
def eval_function(text):
generated = ''.join(text)
try:
decoded = DeepSMILESLanguageModelUtils.decode(generated,
start='<s>',
end='</s>')
smiles = DeepSMILESLanguageModelUtils.sanitize(decoded)
except Exception:
return -1.0
if smiles in all_smiles:
score = -1.0
else:
score = scoring_function.score(smiles)
all_smiles.add(smiles)
if self.best_score < score:
self.best_score = score
self.best_smiles = smiles
return score
mcts = LanguageModelMCTSWithPUCTTerminating(self.lm,
width,
max_depth,
eval_function,
cpuct=c,
terminating_symbol='</s>')
mcts.search(["<s>"], num_simulations)
return [self.best_smiles]
def generate(self, number_samples):
print("generating %s samples..." % number_samples)
all_smiles = set()
samples = []
width = 24
max_depth = 100
c = 5
def eval_function(text):
generated = ''.join(text)
try:
decoded = DeepSMILESLanguageModelUtils.decode(generated,
start='<s>',
end='</s>')
smiles = DeepSMILESLanguageModelUtils.sanitize(decoded)
except Exception:
samples.append("invalid")
return -1.0
samples.append(smiles)
if smiles in all_smiles:
score = -1.0
else:
score = 1.0
all_smiles.add(smiles)
return score
mcts = LanguageModelMCTSWithPUCTTerminating(self.lm,
width,
max_depth,
eval_function,
cpuct=c,
terminating_symbol='</s>')
mcts.search(["<s>"], number_samples)
return samples
num_valid += 1
if smiles in all_smiles:
score = -1.0
else:
score = scorer.score(smiles)
all_smiles[smiles] = (score, generated)
logger.debug("%s, %s" % (smiles, str(score)))
log_best(i, all_smiles, num_valid, logger)
return score
mcts = LanguageModelMCTSWithPUCTTerminating(lm,
width,
max_depth,
eval_function,
cpuct=c,
terminating_symbol='</s>')
state = start_state
logger.info("beginning search...")
start = time.time()
mcts.search(state, num_simulations)
end = time.time()
logger.info("--done--")
logger.info("num valid: %d" % num_valid)
best = mcts.get_best_sequence()
generated_text = ''.join(best[0])
logger.info("best generated text: %s" % generated_text)
# in practice, the log probs are rarely less than -45; so the min tot_score can be: -45 + (sigma*-1.0)
rescale_min = -45 - sigma
if tot_score < rescale_min:
logger.info("WARNING: total score lower than %s" % rescale_min)
# because probabilities are in the range [0,1], the max log prob is log(1) i.e. 0
# so the max tot_score can be: 0 + sigma*1.0
rescale_max = sigma
# scaling x into [a,b]: (b-a)*((x - min(x))/(max(x) - min(x))+a
ret_score = (1 - (-1)) * ((tot_score - rescale_min) /
(rescale_max - rescale_min)) + (-1)
return ret_score
mcts = LanguageModelMCTSWithPUCTTerminating(prior,
width,
max_depth,
eval_function,
cpuct=c,
terminating_symbol='</s>')
state = start_state
logger.info("beginning search...")
start = time.time()
mcts.search(state, num_simulations)
t.cancel()
end = time.time()
logger.info("--done--")
logger.info("num valid: %d" % num_valid)
logger.info("num unique: %s" % len(all_smiles))
