空间站广场

论文

Notebooks

比赛

课程

Apps

我的主页

我的Notebooks

我的论文库

我的足迹

我的工作空间

任务

节点

文件

数据集

镜像

项目

数据库

公开

使用MPerformer进行分子感知/键极预测

中文

化学信息学

AI4S

中文化学信息学AI4S

发布于 2023-09-19

推荐镜像 :mperformer:notebook

推荐机型 :c12_m92_1 * NVIDIA V100

数据集

XYZ(v1)

©️ Copyright 2023 @ Authors
作者： 王钒锰 📨
日期：2023-09-19
共享协议：本作品采用知识共享署名-非商业性使用-相同方式共享 4.0 国际许可协议进行许可。
快速开始：点击上方的 开始连接 按钮，选择 mperformer:notebook镜像 和GPU机型即可开始。

代码

文本

分子感知 (Molecular Perception)

代码

文本

分子感知的目的是基于仅知的部分分子信息(原子的3D坐标和原子类型)来预测出分子的其余结构特征(e.g., bond orders and formal charges)，从而重建出完整的分子结构

换句话说，分子感知的目的就是要实现从分子xyz文件到sdf文件的转换

为此，我们设计开发了一个基于SE(3)-Transformer的通用分子感知框架：MPerformer, 可以基于原子的3D坐标和原子类型准确高效地预测出bond orders、formal charges等重要特征，从而重构出完整的分子结构。

开源代码：https://github.com/FanmengWang/MPerformer
快速体验：https://labs.dp.tech/projects/mperforemr/

代码

文本

准备数据

代码

文本

[1]

# 输入的xyz file的文件夹路径

inputs_path = '/bohr/0919-ttdm/v1/xyz_fold/'

# 处理后得到的sdf file的文件夹路径

outputs_path = '/outputs'

代码

文本

基本设置

代码

文本

[2]

import os

import time

import shutil

t1 = time.time()

data_path = '/MPerformer'

cache_path = '/MPerformer/cache'

weight_path = '/MPerformer/weight/checkpoint.pt'

task_name = '/MPerformer/cache'

result = '/MPerformer/cache/weight_test_cpu.out.pkl'

seed = 0

task_num = 3

conf_size = 1

only_polar = 0

batch_size = 32

dict_name = 'dict.txt'

loss_func = 'finetune_focal_loss_all_constraint_infer'

noH = True

add_noise = False

noise_weight = 0

noise_valid = 0

if os.path.exists(cache_path):

shutil.rmtree(cache_path)

os.mkdir(cache_path)

if os.path.exists(outputs_path):

shutil.rmtree(outputs_path)

os.mkdir(outputs_path)

代码

文本

处理输入的xyz文件，提取原子3D坐标和原子类型信息，并将其存入到lmdb中

代码

文本

[3]

import re

import lmdb

import numpy as np

import pickle

def numpy_seed(seed, *addl_seeds):

"""

Context manager which seeds the NumPy PRNG with the specified seed and

restores the state afterward

"""

if seed is None:

yield

return

if len(addl_seeds) > 0:

seed = int(hash((seed, *addl_seeds)) % 1e6)

state = np.random.get_state()

np.random.seed(seed)

try:

yield

finally:

np.random.set_state(state)

def read_xyz(filename_path, skipH=True):

"""

Read xyz data

:param filename: filename of .xyz file

:param skipH: Do not read H atoms

:return: atomtypes, coordinates and title section

"""

with open(filename_path,'r') as fin:

natoms = int(fin.readline())

title = fin.readline()[:-1]

q=0

qin = re.search("(?:CHARGE|CHG)=([-+]?\d*\.\d+|\d+|[-+]?\d)",title,re.IGNORECASE)

if qin:

q = float(qin.group(1))

coords = []

atomtypes = []

for x in range(natoms):

line = fin.readline().split()

if (line[0].lower()=='h') and skipH: continue

atomtypes.append(line[0])

coords.append([float(line[1]),float(line[2]),float(line[3])])

return(atomtypes, coords, q, title)

outputfilename = os.path.join(cache_path, 'test.lmdb')

env = lmdb.open(

outputfilename,

subdir=False,

readonly=False,

lock=False,

readahead=False,

meminit=False,

max_readers=1,

map_size=int(1000e9),

)

txn_writer = env.begin(write=True)

numpy_seed(seed)

xyz_num = 0

for filename in os.listdir(inputs_path):

if not filename.endswith('.xyz'):

continue

filename_path = os.path.join(inputs_path, filename)

atomtypes, coords, q, title = read_xyz(filename_path, skipH=noH)

coords = np.array(coords, dtype=np.float32)

inner_output = {}

inner_output['data'] = filename.split('.')[0]

inner_output['atoms'] = atomtypes

inner_output['coordinates'] = [coords]

inner_output['target'] = np.array([2] * len(atomtypes), dtype=np.float32)

inner_output['atom_H_num'] = np.array([2] * len(atomtypes), dtype=np.float32)

bond = np.array(list(np.zeros((len(atomtypes),len(atomtypes)), dtype=int)))

for i in range(len(atomtypes)):

bond[i][i] = -1

inner_output['bond'] = bond

txn_writer.put(f"{inner_output['data']}".encode("ascii"), pickle.dumps(inner_output, protocol=-1))

xyz_num = xyz_num + 1

txn_writer.commit()

env.close()

print(f'The number of xyz file is {xyz_num}')

The number of xyz file is 800

代码

文本

调用MPerformer预测bond orders、formal charges等重要特征

代码

文本

[4]

cmd = "python /MPerformer/MPerformer/infer.py --user-dir {} {} --task-name {} --valid-subset test --results-path {} --noise-valid {} --num-workers 1 --ddp-backend=c10d --batch-size {} --task MPerformer --loss {} --arch MPerformer_base --classification-head-name {} --num-classes {} --dict-name {} --conf-size {} --only-polar {} --path {} --fp16 --fp16-init-scale 4 --fp16-scale-window 256 --log-interval 50 --log-format simple".format(os.path.join(data_path, 'MPerformer'), data_path, task_name, cache_path, noise_valid, batch_size, loss_func, task_name, task_num, dict_name, conf_size, only_polar, weight_path)

os.system(cmd)

代码

文本

基于预测结果重构出完整的分子结构

代码

文本

[5]

999

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

›

⌄

import pandas as pd

from tqdm import tqdm, trange

from multiprocessing import Pool

import dpdata

def create_sdfile(name, atomtypes, coords, row1, row2, bond, charge):

"""

Creates string with SD info

:param name: molecule name

:param atomtypes: atomic types

:param coords: coordinates

:return: molblock

"""

ins = name + "\n"

# comment block

ins += "MPerformer generated sdf\n"

ins += "\n"

ins += "%3d%3d 0 0 0 0 0 0 0 0 1 V2000\n" % (len(atomtypes), len(row1))

# atomb block

for at, xyz in zip(atomtypes, coords):

ins += "%10.4f%10.4f%10.4f %-2s 0 0 0 0 0\n" % (xyz[0], xyz[1], xyz[2], at)

# ins += "%2s %12.4f %12.4f %12.4f \n" % ( at,xyz[0], xyz[1], xyz[2])

# bond block

for index in range(len(row1)):

ins += "%3d%3d%3d 0 0 0 0\n" % (row1[index], row2[index], bond[index])

add_charge_num = 0

charge_ins = ""

for idx, charge_num in enumerate(charge):

if charge_num != 0:

charge_ins += f" {idx+1} {charge_num}"

add_charge_num = add_charge_num + 1

if add_charge_num != 0:

charge_ins = f"M CHG {add_charge_num}" + charge_ins

ins += charge_ins

ins += "\n"

print(charge_ins)

ins += "M END"

return(ins)

def check_consist_valid(mol_file):

mol_file= mol_file

pred_atom_charge = mol_file["pred_atom_charge"].view(-1)

target = mol_file["target"]

masked_tokens0 = target.ne(padding_idx)

sample_size0 = masked_tokens0.long().sum()

pred_atom_charge0 = pred_atom_charge[masked_tokens0]

charge_list = []

for i in range(pred_atom_charge0.shape[0]):

if pred_atom_charge0[i] < -0.5:

charge_list.append(-1)

continue

if pred_atom_charge0[i] > 0.5:

charge_list.append(1)

continue

charge_list.append(0)

pred_atom_H = mol_file["pred_atom_H"].view(-1)

atom_H_target = mol_file["atom_H_target"]

masked_atom_H_tokens = atom_H_target.ne(atom_H_pad_idx)

pred_atom_H0 = pred_atom_H[masked_atom_H_tokens]

H_list = []

for i in range(pred_atom_H0.shape[0]):

if pred_atom_H0[i] > 2.5:

H_list.append(3)

continue

if pred_atom_H0[i] < 0.5:

H_list.append(0)

continue

if pred_atom_H0[i] < 1.5:

H_list.append(1)

continue

H_list.append(2)

pred_atom_bond = mol_file["pred_atom_bond"][:, :, 0]

bond_masked_tokens = masked_tokens0

masked_pred_atom_bond = pred_atom_bond[bond_masked_tokens]

masked_bond_target = mol_file['bond_target'][bond_masked_tokens]

non_pad_pos = (masked_bond_target >= 0) & (masked_bond_target != bond_pad_idx)

masked_bond_target = masked_bond_target[non_pad_pos]

masked_pred_atom_bond = masked_pred_atom_bond[non_pad_pos]

bond_list = []

for i in range(masked_pred_atom_bond.shape[0]):

if masked_pred_atom_bond[i] < 0.5:

bond_list.append(0)

continue

if masked_pred_atom_bond[i] > 2.5:

bond_list.append(3)

continue

if masked_pred_atom_bond[i] < 1.25:

bond_list.append(1)

continue

if masked_pred_atom_bond[i] > 1.75:

bond_list.append(2)

continue

bond_list.append(4)

row1 = []

row2 = []

bond_type = []

index = 0

for i in range(sample_size0):

for j in range(sample_size0):

if j == i:

continue

if bond_list[index] != 0 and i < j:

row1.append(i + 1)

row2.append(j + 1)

bond_type.append(bond_list[index])

index = index + 1

ins = create_sdfile(mol_file['id_name'], mol_file['atoms'], mol_file['coords'], row1, row2, bond_type, charge_list)

filename = os.path.join(outputs_path, mol_file['id_name'] + '.sdf')

xyz_filename = os.path.join(inputs_path, mol_file['id_name'] + '.xyz')

with open(filename, 'w') as f:

f.write(ins)

try:

suppl = Chem.SDMolSupplier(filename)

mol = [mol for mol in suppl if mol][0]

except:

try:

suppl = Chem.SDMolSupplier(filename, sanitize=False)

mol = [mol for mol in suppl if mol][0]

system = dpdata.BondOrderSystem(rdkit_mol=mol)

Chem.Kekulize(mol)

except:

return None

with Chem.SDWriter(filename) as w:

w.write(mol)

padding_idx = 0

bond_pad_idx = 6

atom_H_pad_idx = 6

env = lmdb.open(

outputfilename,

subdir=False,

readonly=True,

lock=False,

readahead=True,

meminit=False,

max_readers=8,

map_size=int(1000e9),

)

txn = env.begin()

predicts = pd.read_pickle(result)

mol_file_list = []

for epoch in range(len(predicts)):

predict = predicts[epoch]

bsz = predicts[epoch]['bsz']

for index in range(bsz):

mol_file = {}

mol_file['id_name'] = predict['id_name'][index]

value = pickle.loads(txn.get(mol_file['id_name'].encode()))

mol_file['atoms'] = list(np.array(value['atoms']))

mol_file['coords'] = list(np.array(value['coordinates'][0]))

mol_file['logit_output'] = predict['logit_output'][index].float()

mol_file['target'] = predict['target'][index]

mol_file['logit_atom_H_output'] = predict['logit_atom_H_output'][index].float()

mol_file['atom_H_target'] = predict['atom_H_target'][index]

mol_file['logit_bond_output'] = predict['logit_bond_output'][index].float()

mol_file['bond_target'] = predict['bond_target'][index]

mol_file['pred_atom_charge'] = predict['pred_atom_charge'][index].float()

mol_file['pred_atom_H'] = predict['pred_atom_H'][index].float()

mol_file['pred_atom_bond'] = predict['pred_atom_bond'][index].float()

mol_file_list.append(mol_file)

with Pool() as pool:

for inner_output in tqdm(pool.imap(check_consist_valid, mol_file_list), total=len(mol_file_list)):

pass

/opt/conda/lib/python3.8/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html
  from .autonotebook import tqdm as notebook_tqdm
  0%|          | 0/800 [00:00<?, ?it/s]




  0%|          | 1/800 [00:00<01:37,  8.18it/s]



















  3%|▎         | 25/800 [00:00<00:06, 123.82it/s]




























  6%|▌         | 49/800 [00:00<00:04, 155.92it/s]


















  9%|▉         | 73/800 [00:00<00:04, 167.40it/s]

























 12%|█▏        | 96/800 [00:00<00:03, 181.70it/s]























 15%|█▌        | 120/800 [00:00<00:03, 179.53it/s]
























 18%|█▊        | 148/800 [00:00<00:03, 204.38it/s]


























 21%|██▏       | 170/800 [00:00<00:03, 202.56it/s]





















 24%|██▍       | 194/800 [00:01<00:02, 209.31it/s]
























 27%|██▋       | 218/800 [00:01<00:02, 197.92it/s]

























 31%|███       | 247/800 [00:01<00:02, 221.47it/s]

















 34%|███▍      | 270/800 [00:01<00:02, 220.85it/s]































 37%|███▋      | 293/800 [00:01<00:02, 214.90it/s]

















 40%|███▉      | 318/800 [00:01<00:02, 222.29it/s]




























 43%|████▎     | 343/800 [00:01<00:02, 224.77it/s]


























 46%|████▋     | 370/800 [00:01<00:01, 218.08it/s]






















 49%|████▉     | 393/800 [00:01<00:01, 212.62it/s]





























 52%|█████▏    | 418/800 [00:02<00:01, 214.92it/s]





















 55%|█████▌    | 441/800 [00:02<00:01, 205.78it/s]




















 58%|█████▊    | 464/800 [00:02<00:01, 207.41it/s]





















 61%|██████    | 488/800 [00:02<00:01, 193.75it/s]




















 64%|██████▍   | 510/800 [00:02<00:01, 196.76it/s]























 67%|██████▋   | 534/800 [00:02<00:01, 205.26it/s]





















 70%|██████▉   | 557/800 [00:02<00:01, 200.57it/s]























 72%|███████▏  | 579/800 [00:02<00:01, 189.88it/s]


























 76%|███████▌  | 605/800 [00:03<00:00, 199.76it/s]
























 78%|███████▊  | 628/800 [00:03<00:00, 206.08it/s]






















 82%|████████▏ | 653/800 [00:03<00:00, 217.29it/s]













 84%|████████▍ | 675/800 [00:03<00:00, 214.67it/s]




























 87%|████████▋ | 697/800 [00:03<00:00, 207.36it/s]

















 90%|████████▉ | 718/800 [00:03<00:00, 197.86it/s]






















 92%|█████████▎| 740/800 [00:03<00:00, 196.28it/s]

























 95%|█████████▌| 761/800 [00:03<00:00, 189.41it/s]

































100%|██████████| 800/800 [00:03<00:00, 203.71it/s]

代码

文本

统计结果

代码

文本

[6]

t2 = time.time()

data_process_time = t2 - t1

print(os.listdir(outputs_path))

print(f'input_xyz_fold: {filename}')

print(f'output_sdf_fold: {outputs_path}')

print(f'used_time: {data_process_time / 60} min')

['437.sdf', '176.sdf', '347.sdf', '433.sdf', '340.sdf', '662.sdf', '719.sdf', '462.sdf', '109.sdf', '84.sdf', '484.sdf', '745.sdf', '515.sdf', '156.sdf', '6.sdf', '578.sdf', '80.sdf', '246.sdf', '353.sdf', '291.sdf', '653.sdf', '2.sdf', '42.sdf', '317.sdf', '331.sdf', '632.sdf', '231.sdf', '266.sdf', '714.sdf', '428.sdf', '549.sdf', '607.sdf', '568.sdf', '740.sdf', '680.sdf', '418.sdf', '446.sdf', '4.sdf', '531.sdf', '48.sdf', '518.sdf', '32.sdf', '206.sdf', '706.sdf', '521.sdf', '792.sdf', '290.sdf', '670.sdf', '40.sdf', '16.sdf', '663.sdf', '486.sdf', '767.sdf', '699.sdf', '777.sdf', '287.sdf', '411.sdf', '289.sdf', '51.sdf', '119.sdf', '30.sdf', '161.sdf', '658.sdf', '665.sdf', '196.sdf', '19.sdf', '420.sdf', '467.sdf', '754.sdf', '208.sdf', '163.sdf', '483.sdf', '154.sdf', '496.sdf', '572.sdf', '625.sdf', '36.sdf', '678.sdf', '627.sdf', '123.sdf', '455.sdf', '323.sdf', '452.sdf', '554.sdf', '440.sdf', '648.sdf', '593.sdf', '24.sdf', '14.sdf', '383.sdf', '576.sdf', '292.sdf', '268.sdf', '530.sdf', '10.sdf', '0.sdf', '219.sdf', '454.sdf', '197.sdf', '730.sdf', '527.sdf', '365.sdf', '200.sdf', '120.sdf', '668.sdf', '68.sdf', '514.sdf', '186.sdf', '765.sdf', '776.sdf', '22.sdf', '729.sdf', '5.sdf', '443.sdf', '8.sdf', '377.sdf', '217.sdf', '368.sdf', '720.sdf', '757.sdf', '300.sdf', '256.sdf', '46.sdf', '50.sdf', '336.sdf', '180.sdf', '43.sdf', '305.sdf', '798.sdf', '536.sdf', '778.sdf', '613.sdf', '547.sdf', '87.sdf', '143.sdf', '38.sdf', '233.sdf', '468.sdf', '89.sdf', '711.sdf', '1.sdf', '480.sdf', '773.sdf', '477.sdf', '129.sdf', '451.sdf', '524.sdf', '286.sdf', '421.sdf', '402.sdf', '415.sdf', '229.sdf', '29.sdf', '770.sdf', '115.sdf', '78.sdf', '580.sdf', '332.sdf', '326.sdf', '60.sdf', '552.sdf', '671.sdf', '712.sdf', '182.sdf', '476.sdf', '696.sdf', '288.sdf', '753.sdf', '75.sdf', '155.sdf', '96.sdf', '360.sdf', '638.sdf', '114.sdf', '616.sdf', '350.sdf', '211.sdf', '529.sdf', '414.sdf', '623.sdf', '273.sdf', '498.sdf', '460.sdf', '439.sdf', '172.sdf', '278.sdf', '373.sdf', '131.sdf', '263.sdf', '406.sdf', '689.sdf', '94.sdf', '91.sdf', '214.sdf', '566.sdf', '502.sdf', '592.sdf', '634.sdf', '425.sdf', '339.sdf', '582.sdf', '795.sdf', '384.sdf', '142.sdf', '76.sdf', '100.sdf', '691.sdf', '657.sdf', '444.sdf', '367.sdf', '490.sdf', '732.sdf', '53.sdf', '784.sdf', '203.sdf', '610.sdf', '703.sdf', '378.sdf', '351.sdf', '645.sdf', '74.sdf', '688.sdf', '585.sdf', '750.sdf', '705.sdf', '262.sdf', '789.sdf', '205.sdf', '771.sdf', '167.sdf', '335.sdf', '125.sdf', '427.sdf', '346.sdf', '434.sdf', '73.sdf', '436.sdf', '160.sdf', '128.sdf', '615.sdf', '64.sdf', '708.sdf', '709.sdf', '701.sdf', '629.sdf', '302.sdf', '599.sdf', '617.sdf', '666.sdf', '191.sdf', '27.sdf', '330.sdf', '319.sdf', '293.sdf', '338.sdf', '86.sdf', '388.sdf', '59.sdf', '759.sdf', '643.sdf', '228.sdf', '322.sdf', '717.sdf', '517.sdf', '526.sdf', '394.sdf', '251.sdf', '153.sdf', '158.sdf', '258.sdf', '164.sdf', '679.sdf', '295.sdf', '207.sdf', '199.sdf', '240.sdf', '370.sdf', '503.sdf', '746.sdf', '525.sdf', '141.sdf', '344.sdf', '399.sdf', '280.sdf', '532.sdf', '69.sdf', '630.sdf', '220.sdf', '677.sdf', '587.sdf', '700.sdf', '500.sdf', '565.sdf', '724.sdf', '23.sdf', '21.sdf', '639.sdf', '499.sdf', '533.sdf', '445.sdf', '194.sdf', '401.sdf', '170.sdf', '575.sdf', '449.sdf', '550.sdf', '337.sdf', '598.sdf', '661.sdf', '95.sdf', '431.sdf', '413.sdf', '382.sdf', '97.sdf', '509.sdf', '138.sdf', '345.sdf', '124.sdf', '387.sdf', '13.sdf', '748.sdf', '304.sdf', '398.sdf', '118.sdf', '725.sdf', '758.sdf', '333.sdf', '430.sdf', '309.sdf', '586.sdf', '67.sdf', '31.sdf', '407.sdf', '122.sdf', '747.sdf', '495.sdf', '134.sdf', '794.sdf', '756.sdf', '324.sdf', '390.sdf', '216.sdf', '92.sdf', '553.sdf', '640.sdf', '559.sdf', '571.sdf', '126.sdf', '136.sdf', '684.sdf', '570.sdf', '621.sdf', '133.sdf', '202.sdf', '458.sdf', '151.sdf', '465.sdf', '769.sdf', '655.sdf', '650.sdf', '349.sdf', '716.sdf', '422.sdf', '651.sdf', '470.sdf', '741.sdf', '744.sdf', '132.sdf', '667.sdf', '277.sdf', '159.sdf', '237.sdf', '357.sdf', '395.sdf', '386.sdf', '321.sdf', '783.sdf', '355.sdf', '230.sdf', '225.sdf', '41.sdf', '672.sdf', '62.sdf', '110.sdf', '453.sdf', '310.sdf', '755.sdf', '261.sdf', '513.sdf', '797.sdf', '82.sdf', '70.sdf', '276.sdf', '713.sdf', '168.sdf', '799.sdf', '285.sdf', '56.sdf', '631.sdf', '376.sdf', '601.sdf', '224.sdf', '252.sdf', '727.sdf', '728.sdf', '618.sdf', '113.sdf', '739.sdf', '102.sdf', '190.sdf', '269.sdf', '173.sdf', '259.sdf', '54.sdf', '535.sdf', '560.sdf', '556.sdf', '461.sdf', '391.sdf', '588.sdf', '569.sdf', '417.sdf', '264.sdf', '764.sdf', '375.sdf', '247.sdf', '265.sdf', '218.sdf', '303.sdf', '624.sdf', '563.sdf', '537.sdf', '171.sdf', '209.sdf', '55.sdf', '239.sdf', '169.sdf', '779.sdf', '673.sdf', '489.sdf', '410.sdf', '329.sdf', '793.sdf', '204.sdf', '736.sdf', '248.sdf', '772.sdf', '683.sdf', '676.sdf', '242.sdf', '215.sdf', '419.sdf', '656.sdf', '358.sdf', '223.sdf', '762.sdf', '166.sdf', '675.sdf', '561.sdf', '7.sdf', '600.sdf', '511.sdf', '790.sdf', '174.sdf', '695.sdf', '307.sdf', '140.sdf', '210.sdf', '584.sdf', '494.sdf', '456.sdf', '35.sdf', '58.sdf', '316.sdf', '397.sdf', '145.sdf', '694.sdf', '116.sdf', '198.sdf', '396.sdf', '707.sdf', '579.sdf', '481.sdf', '282.sdf', '474.sdf', '702.sdf', '654.sdf', '697.sdf', '761.sdf', '52.sdf', '488.sdf', '633.sdf', '135.sdf', '408.sdf', '9.sdf', '296.sdf', '611.sdf', '478.sdf', '88.sdf', '594.sdf', '491.sdf', '426.sdf', '614.sdf', '179.sdf', '644.sdf', '226.sdf', '181.sdf', '775.sdf', '137.sdf', '544.sdf', '362.sdf', '112.sdf', '342.sdf', '320.sdf', '548.sdf', '244.sdf', '733.sdf', '379.sdf', '162.sdf', '558.sdf', '101.sdf', '107.sdf', '760.sdf', '270.sdf', '212.sdf', '464.sdf', '274.sdf', '34.sdf', '619.sdf', '341.sdf', '604.sdf', '583.sdf', '685.sdf', '298.sdf', '501.sdf', '641.sdf', '591.sdf', '33.sdf', '234.sdf', '581.sdf', '121.sdf', '652.sdf', '348.sdf', '787.sdf', '567.sdf', '249.sdf', '65.sdf', '272.sdf', '577.sdf', '519.sdf', '366.sdf', '766.sdf', '18.sdf', '482.sdf', '442.sdf', '99.sdf', '400.sdf', '106.sdf', '504.sdf', '236.sdf', '361.sdf', '435.sdf', '416.sdf', '385.sdf', '432.sdf', '403.sdf', '734.sdf', '622.sdf', '165.sdf', '45.sdf', '57.sdf', '77.sdf', '742.sdf', '81.sdf', '479.sdf', '774.sdf', '520.sdf', '506.sdf', '473.sdf', '674.sdf', '192.sdf', '751.sdf', '235.sdf', '325.sdf', '609.sdf', '698.sdf', '538.sdf', '573.sdf', '780.sdf', '603.sdf', '510.sdf', '681.sdf', '508.sdf', '315.sdf', '83.sdf', '15.sdf', '497.sdf', '93.sdf', '313.sdf', '649.sdf', '37.sdf', '380.sdf', '686.sdf', '551.sdf', '294.sdf', '731.sdf', '354.sdf', '44.sdf', '308.sdf', '441.sdf', '597.sdf', '522.sdf', '117.sdf', '245.sdf', '605.sdf', '146.sdf', '493.sdf', '105.sdf', '636.sdf', '193.sdf', '404.sdf', '715.sdf', '343.sdf', '786.sdf', '540.sdf', '647.sdf', '184.sdf', '104.sdf', '311.sdf', '412.sdf', '507.sdf', '738.sdf', '271.sdf', '188.sdf', '381.sdf', '327.sdf', '749.sdf', '472.sdf', '108.sdf', '424.sdf', '328.sdf', '405.sdf', '111.sdf', '457.sdf', '602.sdf', '608.sdf', '241.sdf', '438.sdf', '423.sdf', '626.sdf', '356.sdf', '352.sdf', '178.sdf', '555.sdf', '546.sdf', '130.sdf', '692.sdf', '152.sdf', '542.sdf', '718.sdf', '781.sdf', '541.sdf', '150.sdf', '318.sdf', '63.sdf', '743.sdf', '312.sdf', '185.sdf', '752.sdf', '791.sdf', '183.sdf', '606.sdf', '393.sdf', '195.sdf', '250.sdf', '528.sdf', '177.sdf', '299.sdf', '253.sdf', '175.sdf', '788.sdf', '213.sdf', '763.sdf', '283.sdf', '710.sdf', '127.sdf', '469.sdf', '721.sdf', '471.sdf', '28.sdf', '660.sdf', '723.sdf', '459.sdf', '664.sdf', '284.sdf', '475.sdf', '26.sdf', '371.sdf', '737.sdf', '334.sdf', '275.sdf', '642.sdf', '61.sdf', '221.sdf', '628.sdf', '735.sdf', '369.sdf', '690.sdf', '254.sdf', '782.sdf', '72.sdf', '25.sdf', '47.sdf', '590.sdf', '545.sdf', '90.sdf', '485.sdf', '392.sdf', '543.sdf', '448.sdf', '574.sdf', '187.sdf', '492.sdf', '450.sdf', '363.sdf', '785.sdf', '157.sdf', '612.sdf', '279.sdf', '523.sdf', '534.sdf', '71.sdf', '463.sdf', '596.sdf', '139.sdf', '516.sdf', '297.sdf', '11.sdf', '557.sdf', '359.sdf', '227.sdf', '409.sdf', '620.sdf', '374.sdf', '314.sdf', '260.sdf', '85.sdf', '389.sdf', '103.sdf', '267.sdf', '147.sdf', '539.sdf', '3.sdf', '659.sdf', '637.sdf', '98.sdf', '687.sdf', '39.sdf', '20.sdf', '768.sdf', '281.sdf', '693.sdf', '589.sdf', '306.sdf', '222.sdf', '144.sdf', '487.sdf', '796.sdf', '243.sdf', '429.sdf', '704.sdf', '189.sdf', '466.sdf', '682.sdf', '149.sdf', '201.sdf', '238.sdf', '49.sdf', '564.sdf', '726.sdf', '635.sdf', '372.sdf', '595.sdf', '722.sdf', '512.sdf', '257.sdf', '148.sdf', '669.sdf', '447.sdf', '646.sdf', '255.sdf', '232.sdf', '301.sdf', '17.sdf', '66.sdf', '12.sdf', '364.sdf', '505.sdf', '562.sdf', '79.sdf']
input_xyz_fold: 99.xyz
output_sdf_fold: /outputs
used_time: 0.708737599849701 min

代码

文本

中文

化学信息学

AI4S

中文化学信息学AI4S

点个赞吧