first commit

.gitignore

@@ -0,0 +1,8 @@
+__pycache__
+*.pyc
+*.egg-info/
+.DS_Store
+.ipynb_checkpoints
+/.envrc
+/.python3
+/dist

mold/Aromatic/__init__.py

@@ -0,0 +1 @@
+from ._descriptor import *

mold/Aromatic/_descriptor.py

@@ -0,0 +1,26 @@
+from .._base import Descriptor
+
+
+class AromaticAtomsCount(Descriptor):
+    descriptor_name = 'nAromAtom'
+
+    @property
+    def descriptor_key(self):
+        return self.make_key()
+
+    def calculate(self, mol):
+        return sum((1 for a in mol.GetAtoms() if a.GetIsAromatic()))
+
+
+class AromaticBondsCount(Descriptor):
+    descriptor_name = 'nAromBond'
+
+    @property
+    def descriptor_key(self):
+        return self.make_key()
+
+    def calculate(self, mol):
+        return sum((1 for b in mol.GetBonds() if b.GetIsAromatic()))
+
+_descriptors = [AromaticAtomsCount, AromaticBondsCount]
+__all__ = [d.__name__ for d in _descriptors]

mold/AtomCount/__init__.py

@@ -0,0 +1 @@
+from ._descriptor import *

mold/AtomCount/_descriptor.py

@@ -0,0 +1,47 @@
+from .._base import Descriptor
+
+
+class AtomCount(Descriptor):
+    descriptor_defaults = [
+        ('Atom',), ('HeavyAtom',),
+        ('H',), ('B',), ('C',), ('N',), ('O',), ('S',), ('P',),
+        ('F',), ('Cl',), ('Br',), ('I',), ('X',),
+    ]
+
+    @property
+    def explicitHydrogens(self):
+        return self.symbol in set(['H', 'Atom'])
+
+    @property
+    def descriptor_name(self):
+        return 'n' + self.symbol
+
+    @property
+    def descriptor_key(self):
+        return self.make_key(self.symbol)
+
+    def __init__(self, symbol):
+        self.symbol = symbol
+
+        if symbol == 'X':
+            self.f = self.calcX
+        elif symbol == 'Atom' or symbol == 'HeavyAtom':
+            self.f = self.calcAll
+        else:
+            self.f = self.calc
+
+    def calcX(self, mol):
+        X = set([9, 17, 35, 53, 85, 117])
+        return sum((a.GetAtomicNum() in X for a in mol.GetAtoms()))
+
+    def calc(self, mol):
+        return sum((a.GetSymbol() == self.symbol for a in mol.GetAtoms()))
+
+    def calcAll(self, mol):
+        return mol.GetNumAtoms()
+
+    def calculate(self, mol):
+        return self.f(mol)
+
+_descriptors = [AtomCount]
+__all__ = [d.__name__ for d in _descriptors]

mold/Autocorrelation/__init__.py

@@ -0,0 +1 @@
+from ._descriptor import *

mold/Autocorrelation/_descriptor.py

@@ -0,0 +1,205 @@
+from .._base import Descriptor
+from rdkit import Chem
+from .. import _atomic_property
+import numpy as np
+
+
+class AutocorrelationBase(Descriptor):
+    explicitHydrogens = True
+
+    @property
+    def gasteigerCharges(self):
+        return getattr(self, 'attribute', 0) == _atomic_property.get_charge_explicitHs
+
+
+class DistanceMatrix(AutocorrelationBase):
+    @property
+    def descriptor_key(self):
+        return self.make_key()
+
+    def calculate(self, mol):
+        return Chem.GetDistanceMatrix(mol)
+
+
+class AVec(AutocorrelationBase):
+    @property
+    def descriptor_key(self):
+        return self.make_key(self.attribute)
+
+    def __init__(self, attribute):
+        self.attribute = attribute
+
+    def calculate(self, mol):
+        return np.array([self.attribute(a) for a in mol.GetAtoms()])
+
+
+class CAVec(AutocorrelationBase):
+    @property
+    def descriptor_key(self):
+        return self.make_key(self.attribute)
+
+    @property
+    def dependencies(self):
+        return dict(avec=AVec.make_key(self.attribute))
+
+    def __init__(self, attribute):
+        self.attribute = attribute
+
+    def calculate(self, mol, avec):
+        return avec - avec.mean()
+
+
+class GMat(AutocorrelationBase):
+    @property
+    def descriptor_key(self):
+        return self.make_key(self.distance)
+
+    @property
+    def dependencies(self):
+        return dict(dmat=DistanceMatrix.make_key())
+
+    def __init__(self, distance):
+        self.distance = distance
+
+    def calculate(self, mol, dmat):
+        return dmat == self.distance
+
+
+class GSum(AutocorrelationBase):
+    @property
+    def descriptor_key(self):
+        return self.make_key(self.distance)
+
+    @property
+    def dependencies(self):
+        return dict(gmat=GMat.make_key(self.distance))
+
+    def __init__(self, distance):
+        self.distance = distance
+
+    def calculate(self, mol, gmat):
+        s = gmat.sum()
+        if s == 0:
+            return np.nan
+        else:
+            return s
+
+
+class Autocorrelation(AutocorrelationBase):
+    @property
+    def descriptor_name(self):
+        return '{}{}{}'.format(self.__class__.__name__, self.distance, self.attr_name)
+
+    @property
+    def descriptor_key(self):
+        return self.__class__.make_key(self.attribute, self.distance)
+
+    def __init__(self, distance, attribute):
+        if attribute == 'c':
+            self.attr_name = 'c'
+            self.attribute = _atomic_property.get_charge_explicitHs
+
+        else:
+            self.attr_name, self.attribute = _atomic_property.getter(attribute)
+
+        self.distance = distance
+
+    @property
+    def _avec(self):
+        return AVec.make_key(self.attribute)
+
+    @property
+    def _cavec(self):
+        return CAVec.make_key(self.attribute)
+
+    @property
+    def _gmat(self):
+        return GMat.make_key(self.distance)
+
+    @property
+    def _gsum(self):
+        return GSum.make_key(self.distance)
+
+    @property
+    def _ATS(self):
+        return ATS.make_key(self.distance, self.attribute)
+
+    @property
+    def _ATSC(self):
+        return ATSC.make_key(self.distance, self.attribute)
+
+    @property
+    def _AATSC(self):
+        return AATSC.make_key(self.distance, self.attribute)
+
+MAX_DISTANCE = 8
+
+
+class ATS(Autocorrelation):
+    descriptor_defaults = [(d, a) for a in 'mvepis' for d in range(MAX_DISTANCE + 1)]
+
+    @property
+    def dependencies(self):
+        return dict(avec=self._avec, gmat=self._gmat)
+
+    def calculate(self, mol, avec, gmat):
+        r = float(avec.dot(gmat).dot(avec))
+        return r if self.distance == 0 else r / 2.0
+
+
+class AATS(ATS):
+    @property
+    def dependencies(self):
+        return dict(ATS=self._ATS, gsum=self._gsum)
+
+    def calculate(self, mol, ATS, gsum):
+        r = ATS / gsum
+        return r if self.distance == 0 else r * 2.0
+
+
+class ATSC(Autocorrelation):
+    descriptor_defaults = [(d, a) for a in 'cmvepis' for d in range(MAX_DISTANCE + 1)]
+
+    @property
+    def dependencies(self):
+        return dict(cavec=self._cavec, gmat=self._gmat)
+
+    def calculate(self, mol, cavec, gmat):
+        r = float(cavec.dot(gmat).dot(cavec))
+        return r if self.distance == 0 else r / 2.0
+
+
+class AATSC(ATSC):
+    @property
+    def dependencies(self):
+        return dict(ATSC=self._ATSC, gsum=self._gsum)
+
+    def calculate(self, mol, ATSC, gsum):
+        r = ATSC / gsum
+        return r if self.distance == 0 else r * 2.0
+
+
+class MATS(Autocorrelation):
+    descriptor_defaults = [(d, a) for a in 'cmvepis' for d in range(1, MAX_DISTANCE + 1)]
+
+    @property
+    def dependencies(self):
+        return dict(avec=self._avec, AATSC=self._AATSC, cavec=self._cavec)
+
+    def calculate(self, mol, avec, AATSC, cavec):
+        return len(avec) * AATSC / (cavec ** 2).sum()
+
+
+class GATS(MATS):
+    @property
+    def dependencies(self):
+        return dict(avec=self._avec, gmat=self._gmat, gsum=self._gsum, cavec=self._cavec)
+
+    def calculate(self, mol, avec, gmat, gsum, cavec):
+        W = np.tile(avec, (len(avec), 1))
+        n = (gmat * (W - W.T) ** 2).sum() / (2 * gsum)
+        d = (cavec ** 2).sum() / (len(avec) - 1)
+        return n / d
+
+_descriptors = [ATS, AATS, ATSC, AATSC, MATS, GATS]
+__all__ = [d.__name__ for d in _descriptors]

mold/BondCount/__init__.py

@@ -0,0 +1 @@
+from ._descriptor import *

mold/BondCount/_descriptor.py

@@ -0,0 +1,64 @@
+from .._base import Descriptor
+
+
+class BondCount(Descriptor):
+    descriptor_defaults = [
+        ('',), ('2',),
+        ('S',), ('S2',), ('S3',),
+        ('D',), ('D2',),
+        ('T',), ('Q',), ('M',),
+    ]
+
+    @property
+    def descriptor_name(self):
+        return 'nBonds{}'.format(self.bond_type)
+
+    @property
+    def explicitHydrogens(self):
+        return self.bond_type in set(['2', 'S', 'S2'])
+
+    @property
+    def kekulize(self):
+        return self.bond_type != 'M'
+
+    @property
+    def descriptor_key(self):
+        return self.make_key(self.bond_type)
+
+    def __init__(self, bond_type):
+        self.bond_type = bond_type
+
+    def nBonds(self, n, mol):
+        return sum((1 for b in mol.GetBonds()
+                    if b.GetBondTypeAsDouble() == n))
+
+    def nBondsNA(self, n, mol):
+        return sum((1 for b in mol.GetBonds()
+                    if b.GetBondTypeAsDouble() == n
+                    and not b.GetIsAromatic()))
+
+    def nBondsM(self, mol):
+        return sum((1 for b in mol.GetBonds()
+                    if b.GetBondTypeAsDouble() > 1.0))
+
+    def calculate(self, mol):
+        bt = self.bond_type
+        if bt in ' 2':
+            return mol.GetNumBonds()
+        elif bt == 'S':
+            return self.nBonds(1, mol)
+        elif bt in ['S2', 'S3']:
+            return self.nBondsNA(1, mol)
+        elif bt == 'D':
+            return self.nBonds(2, mol)
+        elif bt == 'D2':
+            return self.nBondsNA(2, mol)
+        elif bt == 'T':
+            return self.nBonds(3, mol)
+        elif bt == 'Q':
+            return self.nBonds(4, mol)
+        elif bt == 'M':
+            return self.nBondsM(mol)
+
+_descriptors = [BondCount]
+__all__ = [d.__name__ for d in _descriptors]

mold/BondCount/_skelton/__init__.py

@@ -0,0 +1 @@
+from ._descriptor import *

mold/BondCount/_skelton/_descriptor.py

@@ -0,0 +1,14 @@
+from .._base import Descriptor
+
+__all__ = []
+
+
+class NAME(Descriptor):
+    @property
+    def descriptor_key(self):
+        pass
+
+    def calculate(self, mol):
+        pass
+
+_descriptors = []

mold/CarbonTypes/__init__.py

@@ -0,0 +1 @@
+from ._descriptor import *