def test_duplicate_fprint_names_detected(self):
from e3fp.fingerprint.fprint import CountFingerprint
from e3fp.fingerprint.db import FingerprintDatabase
array = (np.random.uniform(0, 1, size=(10, 1024)) > 0.9).astype(
np.double
)
fprints = [
CountFingerprint.from_vector(array[i, :]) for i in range(10)
]
for i, fp in enumerate(fprints):
fp.name = str(i)
db = FingerprintDatabase(fp_type=CountFingerprint)
db.add_fingerprints(fprints)
fprints2 = [
CountFingerprint.from_vector(array[i, :]) for i in range(10)
]
for i, fp in enumerate(fprints2):
fp.name = str(i)
db.add_fingerprints(fprints2)
self.assertIs(db.fp_type, CountFingerprint)
self.assertTrue(np.issubdtype(db.array.dtype, np.uint16))
self.assertEqual(db.fp_num, 20)
self.assertEqual(db.bits, 1024)
self.assertEqual(len(db.fp_names_to_indices), 10)
for k, v in db.fp_names_to_indices.items():
k = int(k)
self.assertEqual(k, v[0])
self.assertEqual(k, v[1] - 10)
np.testing.assert_almost_equal(
array[k % 10, :], db.array[k, :].todense().getA().flatten()
)
def test_update_props(self):
from e3fp.fingerprint.fprint import CountFingerprint
from e3fp.fingerprint.db import FingerprintDatabase
array = (np.random.uniform(0, 1, size=(10, 1024)) > 0.9).astype(
np.double
)
fprints = [
CountFingerprint.from_vector(array[i, :]) for i in range(10)
]
for i, fp in enumerate(fprints):
fp.name = str(i)
fp.set_prop("index", i)
db = FingerprintDatabase(fp_type=CountFingerprint)
db.add_fingerprints(fprints)
fprints2 = [
CountFingerprint.from_vector(array[i, :]) for i in range(10)
]
for i, fp in enumerate(fprints2):
fp.name = str(i + len(fprints))
fp.set_prop("index", i)
db.add_fingerprints(fprints2)
indices = db.get_prop("index")
self.assertEqual(indices.shape[0], 20)
self.assertListEqual(
indices.tolist(), list(range(10)) + list(range(10))
)
def test_countfprint_from_counts(self):
from e3fp.fingerprint.fprint import CountFingerprint
in_counts = {3: 1, 1: 4, 5: 1}
bits = 32
fprint = CountFingerprint.from_counts(in_counts, bits=bits)
out_counts = fprint.counts
self.assertEqual(in_counts, out_counts)
def fingerprint(self):
"""
Compute the fingerprint for a compound
:returns: The fingerprint sparse vector dict
"""
try:
return getattr(self, '_fp')
except AttributeError:
pass
self._fp = CountFingerprint(counts={})
if self.mol() is not None:
fp = Fingerprint.from_rdkit(self.mol())
self._fp = CountFingerprint.from_fingerprint(fp)
return self._fp