def process(fname):
results = []
label = int(os.path.basename(fname).replace('.json', ''))
with open(fname, 'r') as f:
data = json.load(f)
ok = []
for d in data:
smi = d['smiles']
if smi is None: continue
# Validate SMILES
errs = molvs.validate_smiles(smi)
if errs:
# print('Validation error(s):', errs)
continue
# Standardize SMILES
smi = molvs.standardize_smiles(smi)
# Check if exists already
if smi in pubchem:
# print('Exists in PubChem')
continue
ok.append(smi)
#print('Kept:', len(ok))
atc_codes = [atc_lookup[i] for i in atc_model.predict(ok)]
for smi, atc_code in zip(ok, atc_codes):
mol = Chem.MolFromSmiles(smi)
formula = CalcMolFormula(mol)
h = md5(smi.encode('utf8')).hexdigest()
im = Draw.MolToImage(mol)
im_path = os.path.join(images_dir, '{}.png'.format(h))
im.save(im_path)
results.append({
'label': label,
'smiles': smi,
'formula': formula,
'image': im_path,
'atc_code': atc_code,
'created_at': datetime.utcnow().isoformat()
})
# Save generated compounds
with open(fname, 'w') as f:
json.dump(results, f)
logFid.close()
header = fid.readline()
smilesDict = {}
print "Might need to change in accordance with file format "
print "**********Please check file format*********"
idx = 0
for line in fid:
idx = idx + 1
lineList = string.split(line, "\t")
smiles = lineList[2]
name = lineList[3]
className = string.strip(lineList[4])
pIC50 = string.strip(lineList[5])
logFid = open(logFile, "a")
#logFid.write("........... Processing "+smiles+" .............\n")
info = validate_smiles(smiles)
if info:
logFid.write(str(info) + " " + smiles + " " + name + "\n")
mol = Chem.MolFromSmiles(smiles)
if not mol:
logFid.write("Could not be transformed to rdkit object " + smiles +
" " + name + "\n")
stdzdSmiles = None
else:
stdzdSmiles = standardize(mol, name, texFile, verbose)
if stdzdSmiles:
smilesDict[name] = stdzdSmiles
outFid = open(outFileName, "a")
outFid.write(
str(stdzdSmiles) + "\t" + name + "\t" + className + "\t" +
pIC50 + "\n")
def process_data(self, csv_data, txt_file, data_size, smile_size,
group_size):
with open(txt_file, "w") as my_output_file:
with open(csv_data, "r") as my_input_file:
[
my_output_file.write("$$$".join(row) + '\n')
for row in csv.reader(my_input_file)
]
my_output_file.close()
with open('sl.txt', 'r') as myfile:
data = myfile.read()
data = data.split("\n")
D = []
drug_names = []
drug_smiles = []
drug_group = []
c = 0
for i in data:
d = i.split("$$$")
D.append(d)
print(d)
if len(d) > 5 and c > 0 and c <= data_size:
drug_names.append(d[12])
drug_group.append(d[2])
drug_smiles.append(d[13])
c += 1
drug_mol = []
for i in drug_smiles:
if validate_smiles(i) != []:
drug_smiles.pop(drug_smiles.index(i))
drug_smiles = list(
filter(lambda a: len(a) <= smile_size and len(a) > 0, drug_smiles))
l = []
for j in drug_smiles:
drug_mol.append(Chem.MolFromSmiles(j))
l.append(len(j))
fps = [FingerprintMols.FingerprintMol(x) for x in drug_mol]
n = group_size
count = 0
Prob = []
Data_Final = []
Data_Final1 = []
data_prov = []
for i in fps:
probs = [DataStructs.FingerprintSimilarity(i, j) for j in fps]
indexes = heapq.nlargest(n, range(len(probs)), probs.__getitem__)
Data1 = [drug_smiles[k] for k in indexes]
Data = [self.smiles_encoder(drug_smiles[k]) for k in indexes]
Prob.append([probs[k] for k in indexes])
count = 0
for i in Data1:
count += 1
if (i not in data_prov and count == 1):
data_prov.append(i)
Data_Final.append(Data)
Data_Final1.append(Data1)
return Data_Final
drug_names.append(d[12])
drug_group.append(d[2])
drug_smiles.append(d[13])
c+=1
# for drug in data:
# print(drug_names[0:3])
# print(drug_smiles[0:3])
# print(drug_group[0:3])
#
drug_mol=[]
#aplicando validate_smiles(drug_smiles[1406]) verifica-se que a molecula e invalida
#remover todas as invalidas
for i in drug_smiles:
if validate_smiles(i) != []:
drug_smiles.pop(drug_smiles.index(i))
drug_smiles = list(filter(lambda a: len(a) <= 100 and len(a) > 0, drug_smiles))
# print(validate_smiles(drug_smiles[1407]))
# print(len(drug_smiles))
l=[]
for j in drug_smiles:
drug_mol.append(Chem.MolFromSmiles(j))
l.append(len(j))
# print(l)
fps = [FingerprintMols.FingerprintMol(x) for x in drug_mol]