def test_smiles(self):
with open('%s/../data/hwi-compounds.csv' % self.path, 'rb') as fh:
reader = csv.DictReader(fh, delimiter="\t")
for row in reader:
if len(row['smiles']) == 0: continue
mol = smilin(row['smiles'])
assert mol != None
def _set_summary_stats(self, path):
"""
Set summary data for each compound (ex. mw,density,smiles).
"""
cols = ['molecular_weight', 'density']
data = {}
with open(path, 'rb') as csvfile:
reader = csv.DictReader(csvfile, delimiter="\t")
for row in reader:
data[row['name'].lower()] = row
for ck in self.cocktails:
for cp in ck.components:
if cp.name not in data:
logger.info("Missing summary data for compound: %s" % cp.name)
continue
row = data[cp.name]
for key in cols:
if key in row and len(row[key]) > 0:
setattr(cp, key, float(row[key]))
else:
setattr(cp, key, None)
if 'smiles' in row and len(row['smiles']) > 0:
cp.smiles = row['smiles']
try:
mol = smilin(cp.smiles)
except:
logger.info("Invalid smiles format, failed to parse smiles for compound: %s" % cp.name)
else:
logger.info("Missing smiles data for compound: %s" % cp.name)
def mol(self):
if self.smiles in _mol_cache:
return _mol_cache[self.smiles]
mol = None
if self.smiles is not None:
try:
mol = smilin(self.smiles)
_mol_cache[self.smiles] = mol
except:
logger.critical("Invalid smiles format, failed to parse smiles for compound: %s" % self.name)
return mol
def test_smilein(self):
smiles_strings = []
with open("{}/smiles.txt".format(self.path), 'rb') as fh:
for line in fh:
smiles_strings.append(line.strip())
for smile in smiles_strings:
mol = smilin(smile)
out = mol.arbsmiles()
can = mol.cansmiles()
for bond in mol.bonds:
print bond.symbol, bond.bondorder, bond.bondtype, bond.fixed
for atom in mol.atoms:
print atom, atom.sumBondOrders()
print smile, out, can
def test_ecfp(self):
"""Test Butyramide (example used in ECFP paper)"""
mol = smilin("CCCC(=O)N")
# ECFP_0
fp = ecfp(mol, radius=0)
assert len(fp) == 5
# ECFP_2
fp = ecfp(mol, radius=1)
assert len(fp) == 11
# ECFP_4
fp = ecfp(mol, radius=2)
assert len(fp) == 14
# ECFP_6
fp = ecfp(mol, radius=3)
assert len(fp) == 14
def molecule_embedding(filename):
dict = {}
df = pd.read_excel(filename)
for id, smile in zip(df.MOLENAME, df.SMILES):
try:
mol = smilin(smile)
try:
fp = ecfp(mol, radius=3)
n_bits = 1024 # Number of bits in fixed-length fingerprint
fingerprint = [0 for _ in range(n_bits)
] #fingerprints as a python list
for nbrhood_hash in fp.keys():
bit = nbrhood_hash % n_bits
fingerprint[bit] = 1
#print(fingerprint)
dict[id] = fingerprint
except:
print('Bad ecfp for', id)
except:
print('Bad mol for', id)
return list(dict.values())
