Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Fetching contributors…

Cannot retrieve contributors at this time

395 lines (314 sloc) 12.286 kB
"""LInked List sparse matrix class
from __future__ import division, print_function, absolute_import
__docformat__ = "restructuredtext en"
__all__ = ['lil_matrix','isspmatrix_lil']
from bisect import bisect_left
import numpy as np
from scipy.lib.six.moves import xrange
from .base import spmatrix, isspmatrix
from .sputils import getdtype, isshape, issequence, isscalarlike, ismatrix, \
IndexMixin, upcast_scalar
from warnings import warn
from .base import SparseEfficiencyWarning
class lil_matrix(spmatrix, IndexMixin):
"""Row-based linked list sparse matrix
This is an efficient structure for constructing sparse
matrices incrementally.
This can be instantiated in several ways:
with a dense matrix or rank-2 ndarray D
with another sparse matrix S (equivalent to S.tolil())
lil_matrix((M, N), [dtype])
to construct an empty matrix with shape (M, N)
dtype is optional, defaulting to dtype='d'.
dtype : dtype
Data type of the matrix
shape : 2-tuple
Shape of the matrix
ndim : int
Number of dimensions (this is always 2)
Number of nonzero elements
LIL format data array of the matrix
LIL format row index array of the matrix
Sparse matrices can be used in arithmetic operations: they support
addition, subtraction, multiplication, division, and matrix power.
Advantages of the LIL format
- supports flexible slicing
- changes to the matrix sparsity structure are efficient
Disadvantages of the LIL format
- arithmetic operations LIL + LIL are slow (consider CSR or CSC)
- slow column slicing (consider CSC)
- slow matrix vector products (consider CSR or CSC)
Intended Usage
- LIL is a convenient format for constructing sparse matrices
- once a matrix has been constructed, convert to CSR or
CSC format for fast arithmetic and matrix vector operations
- consider using the COO format when constructing large matrices
Data Structure
- An array (``self.rows``) of rows, each of which is a sorted
list of column indices of non-zero elements.
- The corresponding nonzero values are stored in similar
fashion in ````.
def __init__(self, arg1, shape=None, dtype=None, copy=False):
self.dtype = getdtype(dtype, arg1, default=float)
# First get the shape
if isspmatrix(arg1):
if isspmatrix_lil(arg1) and copy:
A = arg1.copy()
A = arg1.tolil()
if dtype is not None:
A = A.astype(dtype)
self.shape = A.shape
self.dtype = A.dtype
self.rows = A.rows =
elif isinstance(arg1,tuple):
if isshape(arg1):
if shape is not None:
raise ValueError('invalid use of shape parameter')
M, N = arg1
self.shape = (M,N)
self.rows = np.empty((M,), dtype=object) = np.empty((M,), dtype=object)
for i in range(M):
self.rows[i] = [][i] = []
raise TypeError('unrecognized lil_matrix constructor usage')
# assume A is dense
A = np.asmatrix(arg1)
except TypeError:
raise TypeError('unsupported matrix type')
from .csr import csr_matrix
A = csr_matrix(A, dtype=dtype).tolil()
self.shape = A.shape
self.dtype = A.dtype
self.rows = A.rows =
def set_shape(self,shape):
shape = tuple(shape)
if len(shape) != 2:
raise ValueError("Only two-dimensional sparse arrays "
"are supported.")
shape = int(shape[0]),int(shape[1]) # floats, other weirdness
raise TypeError('invalid shape')
if not (shape[0] >= 0 and shape[1] >= 0):
raise ValueError('invalid shape')
if (self._shape != shape) and (self._shape is not None):
self = self.reshape(shape)
except NotImplementedError:
raise NotImplementedError("Reshaping not implemented for %s." %
self._shape = shape
shape = property(fget=spmatrix.get_shape, fset=set_shape)
def __iadd__(self,other):
self[:,:] = self + other
return self
def __isub__(self,other):
self[:,:] = self - other
return self
def __imul__(self,other):
if isscalarlike(other):
self[:,:] = self * other
return self
raise NotImplementedError
def __itruediv__(self,other):
if isscalarlike(other):
self[:,:] = self / other
return self
raise NotImplementedError
# Whenever the dimensions change, empty lists should be created for each
# row
def getnnz(self):
return sum([len(rowvals) for rowvals in])
nnz = property(fget=getnnz)
def __str__(self):
val = ''
for i, row in enumerate(self.rows):
for pos, j in enumerate(row):
val += " %s\t%s\n" % (str((i, j)), str([i][pos]))
return val[:-1]
def getrowview(self, i):
"""Returns a view of the 'i'th row (without copying).
new = lil_matrix((1, self.shape[1]), dtype=self.dtype)
new.rows[0] = self.rows[i][0] =[i]
return new
def getrow(self, i):
"""Returns a copy of the 'i'th row.
new = lil_matrix((1, self.shape[1]), dtype=self.dtype)
new.rows[0] = self.rows[i][:][0] =[i][:]
return new
def _get1(self, i, j):
if i < 0:
i += self.shape[0]
if i < 0 or i >= self.shape[0]:
raise IndexError('row index out of bounds')
if j < 0:
j += self.shape[1]
if j < 0 or j >= self.shape[1]:
raise IndexError('column index out of bounds')
row = self.rows[i]
data =[i]
pos = bisect_left(row, j)
if pos != len(data) and row[pos] == j:
return self.dtype.type(data[pos])
return self.dtype.type(0)
def __getitem__(self, index):
"""Return the element(s) index=(i, j), where j may be a slice.
This always returns a copy for consistency, since slices into
Python lists return copies.
# Utilities found in IndexMixin
i, j = self._unpack_index(index)
if isscalarlike(i) and isscalarlike(j):
return self._get1(int(i), int(j))
i, j = self._index_to_arrays(i, j)
if i.size == 0:
return lil_matrix((0,0), dtype=self.dtype)
return self.__class__([[self._get1(iii, jjj) for iii, jjj in
zip(ii, jj)] for ii, jj in
zip(i.tolist(), j.tolist())])
def _insertat2(self, row, data, j, x):
""" helper for __setitem__: insert a value in the given row/data at
column j. """
if j < 0: # handle negative column indices
j += self.shape[1]
if j < 0 or j >= self.shape[1]:
raise IndexError('column index out of bounds')
if not np.isscalar(x):
raise ValueError('setting an array element with a sequence')
x = self.dtype.type(x)
except (ValueError, TypeError):
raise TypeError('Unable to convert value (%s) to dtype [%s]' % (x,
pos = bisect_left(row, j)
if x != 0:
if pos == len(row):
elif row[pos] != j:
row.insert(pos, j)
data.insert(pos, x)
data[pos] = x
if pos < len(row) and row[pos] == j:
del row[pos]
del data[pos]
def __setitem__(self, index, x):
i, j = self._unpack_index(index)
# shortcut for common case of full matrix assign:
if (isspmatrix(x) and isinstance(i, slice) and i == slice(None) and
isinstance(j, slice) and j == slice(None)
and x.shape == self.shape):
x = lil_matrix(x, dtype=self.dtype)
self.rows = x.rows =
i, j = self._index_to_arrays(i, j)
if isspmatrix(x):
x = x.toarray()
# Make x and i into the same shape
x = np.asarray(x, dtype=self.dtype)
x, _ = np.broadcast_arrays(x, i)
if x.shape != i.shape:
raise ValueError("shape mismatch in assignment")
# Set values
for ii, jj, xx in zip(i.ravel(), j.ravel(), x.ravel()):
self._insertat2(self.rows[int(ii)],[int(ii)], int(jj), xx)
def _mul_scalar(self, other):
if other == 0:
# Multiply by zero: return the zero matrix
new = lil_matrix(self.shape, dtype=self.dtype)
res_dtype = upcast_scalar(self.dtype, other)
new = self.copy()
new = new.astype(res_dtype)
# Multiply this scalar by every element.[:] = [[val*other for val in rowvals] for
rowvals in]
return new
def __truediv__(self, other): # self / other
if isscalarlike(other):
new = self.copy()
# Divide every element by this scalar = [[val/other for val in rowvals] for
rowvals in]
return new
return self.tocsr() / other
def copy(self):
from copy import deepcopy
new = lil_matrix(self.shape, dtype=self.dtype) = deepcopy(
new.rows = deepcopy(self.rows)
return new
def reshape(self,shape):
new = lil_matrix(shape, dtype=self.dtype)
j_max = self.shape[1]
for i,row in enumerate(self.rows):
for col,j in enumerate(row):
new_r,new_c = np.unravel_index(i*j_max + j,shape)
new[new_r,new_c] = self[i,j]
return new
def toarray(self, order=None, out=None):
"""See the docstring for `spmatrix.toarray`."""
d = self._process_toarray_args(order, out)
for i, row in enumerate(self.rows):
for pos, j in enumerate(row):
d[i, j] =[i][pos]
return d
def transpose(self):
return self.tocsr().transpose().tolil()
def tolil(self, copy=False):
if copy:
return self.copy()
return self
def tocsr(self):
""" Return Compressed Sparse Row format arrays for this matrix.
indptr = np.asarray([len(x) for x in self.rows], dtype=np.intc)
indptr = np.concatenate((np.array([0], dtype=np.intc), np.cumsum(indptr)))
nnz = indptr[-1]
indices = []
for x in self.rows:
indices = np.asarray(indices, dtype=np.intc)
data = []
for x in
data = np.asarray(data, dtype=self.dtype)
from .csr import csr_matrix
return csr_matrix((data, indices, indptr), shape=self.shape)
def tocsc(self):
""" Return Compressed Sparse Column format arrays for this matrix.
return self.tocsr().tocsc()
def isspmatrix_lil(x):
return isinstance(x, lil_matrix)
Jump to Line
Something went wrong with that request. Please try again.