rdkit.DataStructs.cDataStructs module

Module containing an assortment of functionality for basic data structures.

At the moment the data structures defined are:

Bit Vector classes (for storing signatures, fingerprints and the like:

• ExplicitBitVect: class for relatively small (10s of thousands of bits) or

  dense bit vectors.

• SparseBitVect: class for large, sparse bit vectors

DiscreteValueVect: class for storing vectors of integers SparseIntVect: class for storing sparse vectors of integers

class rdkit.DataStructs.cDataStructs.DiscreteValueType(*values)

Bases: Enum

EIGHTBITVALUE = 3

FOURBITVALUE = 2

ONEBITVALUE = 0

SIXTEENBITVALUE = 4

TWOBITVALUE = 1

class rdkit.DataStructs.cDataStructs.DiscreteValueVect(self)

class rdkit.DataStructs.cDataStructs.DiscreteValueVect(self, valType: rdkit.DataStructs.cDataStructs.DiscreteValueType, length: int)

class rdkit.DataStructs.cDataStructs.DiscreteValueVect(self, pkl: bytes)

class rdkit.DataStructs.cDataStructs.DiscreteValueVect(self, pkl: str)

Bases: object

A container class for storing unsigned integer values within a particular range.

The length of the vector and type of its elements (determines the maximum value that can be stored) are both set at construction time.

As you would expect, _DiscreteValueVects_ support a set of binary operations so you can do things like:

dvv3 = dvv1 & dvv2 the result contains the smallest value in each entry dvv3 = dvv1 | dvv2 the result contains the largest value in each entry dvv1 += dvv2 values are truncated when necessary dvv3 = dvv1 + dvv2 values are truncated when necessary dvv1 -= dvv3 would-be negative values are set to zero dvv3 = dvv1 - dvv2 would-be negative values are set to zero

Elements can be set and read using indexing (i.e. bv[i] = 4 or val=bv[i])

GetTotalVal(self) → int

Get the sum of the values in the vector, basically L1 norm

GetValueType(self) → rdkit.DataStructs.cDataStructs.DiscreteValueType

Get the type of value stored in the vector

class rdkit.DataStructs.cDataStructs.ExplicitBitVect(self)

class rdkit.DataStructs.cDataStructs.ExplicitBitVect(self, size: int)

class rdkit.DataStructs.cDataStructs.ExplicitBitVect(self, pkl: bytes)

class rdkit.DataStructs.cDataStructs.ExplicitBitVect(self, pkl: str)

class rdkit.DataStructs.cDataStructs.ExplicitBitVect(self, size: int, bitsSet: bool)

Bases: object

A class to store explicit bit vectors.

This class is most useful for situations where the size of the vector is relatively small (tens of thousands or smaller).

For larger vectors, use the _SparseBitVect_ class instead.

As you would expect, _ExplicitBitVects_ support a set of binary operations so you can do things like:

bv3 = bv1 & bv2 (bitwise and) bv3 = bv1 | bv2 (bitwise or) bv3 = bv1 ^ bv2 (bitwise xor) bv3 = ~bv1 (bitwise negation)

Bits can be set and read using either the Set/UnsetBit() and GetBit() methods or by indexing (i.e. bv[i] = 1 or if bv[i]).

FromBase64(self, inD: str) → None

Initializes the vector from a base64 encoded binary string.

GetBit(self, which: int) → bool

Returns the value of a bit.

GetNumBits(self) → int

Returns the number of bits in the vector (the vector’s size).

GetNumOffBits(self) → int

Returns the number of off bits.

GetNumOnBits(self) → int

Returns the number of on bits.

GetOnBits(self) → list[int]

Returns a tuple containing IDs of the on bits.

SetBit(self, which: int) → bool

Turns on a particular bit. Returns the original state of the bit.

SetBitsFromList(self, onBitList: collections.abc.Iterable) → None

Turns on a set of bits. The argument should be a tuple or list of bit ids.

ToBase64(self) → str

Converts the vector to a base64 string (the base64 encoded version of the results of ToString()).

ToBinary(self) → bytes

Returns an internal binary representation of the vector.

ToBitString()

ToList(self) → list

Return the Bitvector as a python list (faster than list(vect))

UnSetBit(self, which: int) → bool

Turns off a particular bit. Returns the original state of the bit.

UnSetBitsFromList(self, offBitList: collections.abc.Iterable) → None

Turns off a set of bits. The argument should be a tuple or list of bit ids.

class rdkit.DataStructs.cDataStructs.FPBReader(self, filename: str, lazy: bool = False)

Bases: object

A class for reading and searching FPB files from Andrew Dalke’s chemfp. Note that this functionality is still experimental and the API may change in future releases.

docstring

GetBytes(self, which: int) → bytes

returns a particular fingerprint as bytes

GetContainingNeighbors(self, bv: bytes) → tuple

returns indices of neighbors that contain this fingerprint (where all bits from this fingerprint are also set)

GetFP(self, idx: int) → rdkit.DataStructs.cDataStructs.ExplicitBitVect

returns a particular fingerprint as an ExplicitBitVect

GetId(self, idx: int) → str

returns the id of a particular fingerprint

GetNumBits(self) → int

returns the number of bits in a fingerprint

GetTanimoto(self, which: int, bytes: bytes) → float

return the tanimoto similarity of a particular fingerprint to the bytes provided

GetTanimotoNeighbors(self, bv: bytes, threshold: float = 0.7) → tuple

returns tanimoto similarities to and indices of all neighbors above the specified threshold

GetTversky(self, which: int, bytes: bytes, ca: float, cb: float) → float

return the Tverksy similarity of a particular fingerprint to the bytes provided

GetTverskyNeighbors(self, bv: bytes, ca: float, cb: float, threshold: float = 0.7) → tuple

returns Tversky similarities to and indices of all neighbors above the specified threshold

Init(self) → None

Read the fingerprints from the file. This can take a while.

class rdkit.DataStructs.cDataStructs.IntSparseIntVect(self)

class rdkit.DataStructs.cDataStructs.IntSparseIntVect(self, pkl: bytes)

class rdkit.DataStructs.cDataStructs.IntSparseIntVect(self, pkl: str)

class rdkit.DataStructs.cDataStructs.IntSparseIntVect(self, arg: int, /)

Bases: object

A container class for storing integer values within a particular range.

The length of the vector is set at construction time.

As you would expect, _SparseIntVects_ support a set of binary operations so you can do things like:

Arithmetic: siv1 += siv2 siv3 = siv1 + siv2 siv1 -= siv3 siv3 = siv1 - siv2 “Fuzzy” binary operations: siv3 = siv1 & siv2 the result contains the smallest value in each entry siv3 = siv1 | siv2 the result contains the largest value in each entry

Elements can be set and read using indexing (i.e. siv[i] = 4 or val=siv[i])

Overloaded function.

1. __init__(self) -> None

Constructor

2. __init__(self, pkl: bytes) -> None

3. __init__(self, pkl: str) -> None

4. __init__(self, arg: int, /) -> None

Constructor

GetLength(self) → int

Returns the length of the vector

GetNonzeroElements(self) → dict

returns a dictionary of the nonzero elements

GetTotalVal(self, useAbs: bool = False) → int

Get the sum of the values in the vector, basically L1 norm

ToBinary(self) → bytes

returns a binary (pickle) representation of the vector

ToList(self) → list

Return the SparseIntVect as a python list

UpdateFromSequence(self, seq: collections.abc.Iterable) → None

update the vector based on the values in the list or tuple

class rdkit.DataStructs.cDataStructs.LongSparseIntVect(self)

class rdkit.DataStructs.cDataStructs.LongSparseIntVect(self, pkl: bytes)

class rdkit.DataStructs.cDataStructs.LongSparseIntVect(self, pkl: str)

class rdkit.DataStructs.cDataStructs.LongSparseIntVect(self, arg: int, /)

Bases: object

A container class for storing integer values within a particular range.

The length of the vector is set at construction time.

As you would expect, _SparseIntVects_ support a set of binary operations so you can do things like:

Arithmetic: siv1 += siv2 siv3 = siv1 + siv2 siv1 -= siv3 siv3 = siv1 - siv2 “Fuzzy” binary operations: siv3 = siv1 & siv2 the result contains the smallest value in each entry siv3 = siv1 | siv2 the result contains the largest value in each entry

Elements can be set and read using indexing (i.e. siv[i] = 4 or val=siv[i])

Overloaded function.

1. __init__(self) -> None

Constructor

2. __init__(self, pkl: bytes) -> None

3. __init__(self, pkl: str) -> None

4. __init__(self, arg: int, /) -> None

Constructor

GetLength(self) → int

Returns the length of the vector

GetNonzeroElements(self) → dict

returns a dictionary of the nonzero elements

GetTotalVal(self, useAbs: bool = False) → int

Get the sum of the values in the vector, basically L1 norm

ToBinary(self) → bytes

returns a binary (pickle) representation of the vector

ToList(self) → list

Return the SparseIntVect as a python list

UpdateFromSequence(self, seq: collections.abc.Iterable) → None

update the vector based on the values in the list or tuple

class rdkit.DataStructs.cDataStructs.MultiFPBReader(self, initOnSearch: bool = False)

Bases: object

A class for reading and searching multiple FPB files from Andrew Dalke’s chemfp. Note that this functionality is still experimental and the API may change in future releases.

docstring

AddReader(self, rdr: rdkit.DataStructs.cDataStructs.FPBReader) → int

adds an FPBReader to our set of readers

GetContainingNeighbors(self, bv: bytes, numThreads: int = 1) → tuple

returns indices of neighbors that contain this fingerprint (where all bits from this fingerprint are also set)

GetNumBits(self) → int

returns the number of bits in a fingerprint

GetReader(self, which: int) → rdkit.DataStructs.cDataStructs.FPBReader

returns one of our readers

GetTanimotoNeighbors(self, bv: bytes, threshold: float = 0.7, numThreads: int = 1) → tuple

returns tanimoto similarities to and indices of all neighbors above the specified threshold

GetTverskyNeighbors(self, bv: bytes, ca: float, cb: float, threshold: float = 0.7, numThreads: int = 1) → tuple

returns Tversky similarities to and indices of all neighbors above the specified threshold

Init(self) → None

Call Init() on each of our children. This can take a while.

class rdkit.DataStructs.cDataStructs.RealValueVect(self)

class rdkit.DataStructs.cDataStructs.RealValueVect(self, length: int)

class rdkit.DataStructs.cDataStructs.RealValueVect(self, pkl: bytes)

class rdkit.DataStructs.cDataStructs.RealValueVect(self, pkl: str)

Bases: object

A container class for storing real values.

The length of the vector is set at construction time.

As you would expect, _RealValueVects_ support a set of binary operations so you can do things like:

rvv3 = rvv1 & rvv2 the result contains the smallest value in each entry rvv3 = rvv1 | rvv2 the result contains the largest value in each entry rvv1 += rvv2 rvv3 = rvv1 + rvv2 rvv1 -= rvv3 rvv3 = rvv1 - rvv2

Elements can be set and read using indexing (i.e. bv[i] = 4 or val=bv[i])

GetTotalVal(self) → float

Get the sum of the values in the vector, basically L1 norm

class rdkit.DataStructs.cDataStructs.SparseBitVect(self)

class rdkit.DataStructs.cDataStructs.SparseBitVect(self, size: int)

class rdkit.DataStructs.cDataStructs.SparseBitVect(self, pkl: bytes)

class rdkit.DataStructs.cDataStructs.SparseBitVect(self, pkl: str)

Bases: object

A class to store sparse bit vectors.

This class is most useful for situations where the size of the vector is large and relatively few bits are set

For smaller or denser vectors, the _ExplicitBitVect_ class is much faster.

As you would expect, _SparseBitVects_ support a set of binary operations so you can do things like:

bv3 = bv1 & bv2 (bitwise and) bv3 = bv1 | bv2 (bitwise or) bv3 = bv1 ^ bv2 (bitwise xor) bv3 = ~bv1 (bitwise negation) NOTE: this operation is likely

to be VERY slow and inefficient.

Bits can be set and read using either the Set/UnsetBit() and GetBit() methods or by indexing (i.e. bv[i] = 1 or if bv[i]).

FromBase64(self, inD: str) → None

Initializes the vector from a base64 encoded binary string.

GetBit(self, which: int) → bool

Returns the value of a bit.

GetNumBits(self) → int

Returns the number of bits in the vector (the vector’s size).

GetNumOffBits(self) → int

Returns the number of off bits.

GetNumOnBits(self) → int

Returns the number of on bits.

GetOnBits(self) → list[int]

Returns a tuple containing IDs of the on bits.

SetBit(self, which: int) → bool

Turns on a particular bit. Returns the original state of the bit.

SetBitsFromList(self, onBitList: collections.abc.Iterable) → None

Turns on a set of bits. The argument should be a tuple or list of bit ids.

ToBase64(self) → str

Converts the vector to a base64 string (the base64 encoded version of the results of ToString()).

ToBinary(self) → bytes

Returns an internal binary representation of the vector.

ToBitString()

ToList(self) → list

Return the BitVector as a python list.

UnSetBit(self, which: int) → bool

Turns off a particular bit. Returns the original state of the bit.

UnSetBitsFromList(self, offBitList: collections.abc.Iterable) → None

Turns off a set of bits. The argument should be a tuple or list of bit ids.

class rdkit.DataStructs.cDataStructs.UIntSparseIntVect(self)

class rdkit.DataStructs.cDataStructs.UIntSparseIntVect(self, pkl: bytes)

class rdkit.DataStructs.cDataStructs.UIntSparseIntVect(self, pkl: str)

class rdkit.DataStructs.cDataStructs.UIntSparseIntVect(self, arg: int, /)

Bases: object

A container class for storing integer values within a particular range.

The length of the vector is set at construction time.

As you would expect, _SparseIntVects_ support a set of binary operations so you can do things like:

Arithmetic: siv1 += siv2 siv3 = siv1 + siv2 siv1 -= siv3 siv3 = siv1 - siv2 “Fuzzy” binary operations: siv3 = siv1 & siv2 the result contains the smallest value in each entry siv3 = siv1 | siv2 the result contains the largest value in each entry

Elements can be set and read using indexing (i.e. siv[i] = 4 or val=siv[i])

Overloaded function.

1. __init__(self) -> None

Constructor

2. __init__(self, pkl: bytes) -> None

3. __init__(self, pkl: str) -> None

4. __init__(self, arg: int, /) -> None

Constructor

GetLength(self) → int

Returns the length of the vector

GetNonzeroElements(self) → dict

returns a dictionary of the nonzero elements

GetTotalVal(self, useAbs: bool = False) → int

Get the sum of the values in the vector, basically L1 norm

ToBinary(self) → bytes

returns a binary (pickle) representation of the vector

ToList(self) → list

Return the SparseIntVect as a python list

UpdateFromSequence(self, seq: collections.abc.Iterable) → None

update the vector based on the values in the list or tuple

class rdkit.DataStructs.cDataStructs.ULongSparseIntVect(self)

class rdkit.DataStructs.cDataStructs.ULongSparseIntVect(self, pkl: bytes)

class rdkit.DataStructs.cDataStructs.ULongSparseIntVect(self, pkl: str)

class rdkit.DataStructs.cDataStructs.ULongSparseIntVect(self, arg: int, /)

Bases: object

A container class for storing integer values within a particular range.

The length of the vector is set at construction time.

As you would expect, _SparseIntVects_ support a set of binary operations so you can do things like:

Arithmetic: siv1 += siv2 siv3 = siv1 + siv2 siv1 -= siv3 siv3 = siv1 - siv2 “Fuzzy” binary operations: siv3 = siv1 & siv2 the result contains the smallest value in each entry siv3 = siv1 | siv2 the result contains the largest value in each entry

Elements can be set and read using indexing (i.e. siv[i] = 4 or val=siv[i])

Overloaded function.

1. __init__(self) -> None

Constructor

2. __init__(self, pkl: bytes) -> None

3. __init__(self, pkl: str) -> None

4. __init__(self, arg: int, /) -> None

Constructor

GetLength(self) → int

Returns the length of the vector

GetNonzeroElements(self) → dict

returns a dictionary of the nonzero elements

GetTotalVal(self, useAbs: bool = False) → int

Get the sum of the values in the vector, basically L1 norm

ToBinary(self) → bytes

returns a binary (pickle) representation of the vector

ToList(self) → list

Return the SparseIntVect as a python list

UpdateFromSequence(self, seq: collections.abc.Iterable) → None

update the vector based on the values in the list or tuple