We Keep Coding

Unexpected angle bracket "<" in numpy multiply() of matrices

How does adj.multiply(adj.T > adj) work?
The code runs fine but I don't expect an angle bracket > in a multiply statement. I believe the docs stated to provide two values to multiply(), but it's still working and producing an output matrix by replacing (adj.T > adj) with (True), (False), (adj.T != adj), but not (adj.T = adj). Also, that multiply() method is not attached on the end of a variable, whereas it is used as adj.multiply() here. The source of the method multiply seemed to just convert it to a csr_matrix and run numpy's multiply(), then IIRC converts it back to coo_matrix. The .T of course means "transpose".
# build symmetric adjacency matrix
adj = adj + adj.T.multiply(adj.T > adj) - adj.multiply(adj.T > adj)`
For some context, "adj" is a scipy coo_matrix from graph convolutional network on github, which I'm trying to understand how the input is prepared for.
adj = sp.coo_matrix((np.ones(edges.shape[0]), (edges[:, 0], edges[:, 1])),
shape=(labels.shape[0], labels.shape[0]),
dtype=np.float32)
Attempting to reproduce the code requires running that whole page.
The following is easier to recreate and test:
import numpy as np
import scipy.sparse as sp
asdf = sp.coo_matrix((np.ones(5), (np.ones(5), np.ones(5))), shape=(5,5),
dtype=np.float32)
print(asdf)
print(asdf.toarray())
asdf = asdf + asdf.T.multiply(asdf.T > asdf) - asdf.multiply(asdf.T > asdf)
print("asdf")
print(asdf.toarray())
at row=1,col=1, value was 5, where a asdf.T.multiply(True) statement doubled its value 5, to 10. Passing two variables separated by space or comma doesn't work.
Update:
I placed a number (not a whole matrix) before the ">" angle bracket and it produced this error:
/usr/local/lib/python3.6/dist-packages/scipy/sparse/compressed.py:287: SparseEfficiencyWarning: Comparing a sparse matrix with a scalar greater than zero using < is inefficient, try using >= instead.
warn(bad_scalar_msg, SparseEfficiencyWarning)
Seeing that, I realized that there was a different sparse multiply() method that didn't show in google without explicitly typing "sparse". Its documentation is here, but I don't see how it handles an angle bracket or condition.

Make a record for this problem.
When I try to handle confusion about this line of code:
adj = adj + adj.T.multiply(adj.T > adj) - adj.multiply(adj.T > adj)
I need to understand three parts:
1. adj.T > adj
It builds a bool type matrix with the same shape of matrix adj.
2. multiply()
The function achieves Hadamard product actually, rather than Dot product. The official document gives the description: "Point-wise multiplication by another matrix".
3. the way to build symmetric adjacency matrix
We'd better reserve the larger weight item between (i, j) and (j, i) instead of adding them together simply, as the latter will add the weight of the edge.
There is the code to explain them:
import numpy as np
import scipy.sparse as sp
row = np.array([0, 0, 0, 1, 1, 1, 2, 2, 2])
col = np.array([0, 1, 2, 0, 1, 2, 0, 1, 2])
A = sp.coo_matrix(([0, 3, 2, 0, 0, 0, 7, 4, 0], (row, col)),
shape=(3, 3), dtype=np.float32)
print(A.toarray())
print((A.T > A).toarray())
print(A.T.multiply(A.T > A).toarray())
A = A + A.T.multiply(A.T > A) - A.multiply(A.T > A)
print("The symmetric adjacency matrix:", A.toarray(), sep='\n')

With the following code, I consistently find the multiply() will multiply by a scalar or matrix value, but with a condition, it performs the operator previous to it if the condition is true, without performing a multiplication.
For example, the following:
asdf = asdf + asdf.T.multiply(True)
will only perform the + operation because of the True boolean being passed. In this case there is no multiplication. Same results, if instead of True we have 3>2, which is true, (it doesn't multiply by the 3 or 2).
row = np.array([0,1,2,3,4])
col = np.array([0,1,2,3,4])
asdf = sp.coo_matrix(([1,2,3,4,5], (row, col)), shape=(5,5), dtype=np.float32)
print(asdf)
print(asdf.toarray())
asdf = asdf + asdf.T.multiply(2 < 3) - asdf.multiply(7 != 8)
print("asdf")
print(asdf.toarray())

Einsum for high dimensions

Considering the 3 arrays below:
np.random.seed(0)
X = np.random.randint(10, size=(4,5))
W = np.random.randint(10, size=(3,4))
y = np.random.randint(3, size=(5,1))
i want to add and sum each column of the matrix X to the row of W ,given by y as index. So ,for example, if the first element in y is 3 , i'll add the first column of X to the fourth row of W(index 3 in python) and sum it. i'll do it over and over until all columns of X are added to the specific row of W and summed.
i could do it in different ways:
1- using for loop:
for i,j in enumerate(y):
W[j]+=X[:,i]
2- using the add.at function
np.add.at(W,(y.ravel()),X.T)
3- but i can't understand how to do it using einsum.
i was given a solution ,but really can't understand it.
N = y.max()+1
W[:N] += np.einsum('ijk,lk->il',(np.arange(N)[:,None,None] == y.ravel()),X)
Anyone could explain me this structure?
1 - (np.arange(N)[:,None,None] == y.ravel(),X). i imagine this part refers to summing the column of X to the specific row of W ,according to y. But where s W ? and why do we have to transform W in 4 dimensions in this case?
2- 'ijk,lk->il' - i didnt understand this either.
i -refers to the rows,
j - columns,
k- each element,
l - what does 'l' refers too?.
if anyone can understand this and explain to me , i would really appreciate.
Thanks in advance.

Let's simplify the problem by dropping one dimension and using values that are easy to verify manually:
W = np.zeros(3, np.int)
y = np.array([0, 1, 1, 2, 2])
X = np.array([1, 2, 3, 4, 5])
Values in the vector W get added values from X by looking up with y:
for i, j in enumerate(y):
W[j] += X[i]
W is calculated as [1, 5, 9], (check quickly by hand).
Now, how could this code be vectorized? We can't do a simple W[y] += X[y] as y has duplicate values in it and the different sums would overwrite each other at indices 1 and 2.
What could be done is to broadcast the values into a new dimension of len(y) and then sum up over this newly created dimension.
N = W.shape[0]
select = (np.arange(N) == y[:, None]).astype(np.int)
Taking the index range of W ([0, 1, 2]), and setting the values where they match y to 1 in a new dimension, otherwise 0. select contains this array:
array([[1, 0, 0],
[0, 1, 0],
[0, 1, 0],
[0, 0, 1],
[0, 0, 1]])
It has len(y) == len(X) rows and len(W) columns and shows for every y/row, what index of W it contributes to.
Let's multiply X with this array, mult = select * X[:, None]:
array([[1, 0, 0],
[0, 2, 0],
[0, 3, 0],
[0, 0, 4],
[0, 0, 5]])
We have effectively spread out X into a new dimension, and sorted it in a way we can get it into shape W by summing over the newly created dimension. The sum over the rows is the vector we want to add to W:
sum_Xy = np.sum(mult, axis=0) # [1, 5, 9]
W += sum_Xy
The computation of select and mult can be combined with np.einsum:
# `select` has shape (len(y)==len(X), len(W)), or `yw`
# `X` has shape len(X)==len(y), or `y`
# we want something `len(W)`, or `w`, and to reduce the other dimension
sum_Xy = np.einsum("yw,y->w", select, X)
And that's it for the one-dimensional example. For the two-dimensional problem posed in the question it is exactly the same approach: introduce an additional dimension, broadcast the y indices, and then reduce the additional dimension with einsum.
If you internalize how every step works for the one-dimensional example, I'm sure you can work out how the code is doing it in two dimensions, as it is just a matter of getting the indices right (W rows, X columns).

Matrix QR factorization algorithms

I've been trying to visualize QR decomposition in a step by step fashion, but I'm not getting expected results. I'm new to numpy so it'd be nice if any expert eye could spot what I might be missing:
import numpy as np
from scipy import linalg
A = np.array([[12, -51, 4],
[6, 167, -68],
[-4, 24, -41]])
#Givens
v = np.array([12, 6])
vnorm = np.linalg.norm(v)
W_12 = np.array([[v[0]/vnorm, v[1]/vnorm, 0],
[-v[1]/vnorm, v[0]/vnorm, 0],
[0, 0, 1]])
W_12 * A #this should return a matrix such that [1,0] = 0
#gram-schmidt
A[:,0]
v = np.linalg.norm(A[:,0]) * np.array([1, 0, 0])
u = (A[:,0] - v)
u = u / np.linalg.norm(u)
W1 = np.eye(3) - 2 * np.outer(u, u.transpose())
W1 * A #this matrix's first column should look like [a, 0, 0]
any help clarifying the fact that this intermediate results don't show the properties that they are supposed to will be greatly received

NumPy is designed to work with homogeneous multi-dimensional arrays, it is not specifically a linear algebra package. So by design, the * operator is element-wise multiplication, not the matrix product.
If you want to get the matrix product, there are a few ways:
You can create np.matrix objects, rather than np.ndarray objects, for which the * operator is the matrix product.
You can also use the # operator, as in W_12 # A, which is the matrix product.
Or you can use np.dot(W_12, A) or W_12.dot(A), which computes the dot product.
Any one of these, using the data you give, returns the following for Givens rotation:
>>> np.dot(W_12 A)[1, 0]
-2.2204460492503131e-16
And this for the Gram-Schmidt step:
>>> (W1.dot(A))[:, 0]
array([ 1.40000000e+01, -4.44089210e-16, 4.44089210e-16])

build matrix from blocks

I have an object which is described by two quantities, A and B (in real case they can be more than two). Objects are correlated depending on the value of A and B. In particular I know the correlation matrix for A and for B. Just as example:
a = np.array([[1, 1, 0, 0],
[1, 1, 0, 0],
[0, 0, 1, 1],
[0, 0, 1, 1]])
b = np.array([[1, 1, 0],
[1, 1, 1],
[0, 1, 1]])
na = a.shape[0]
nb = b.shape[0]
correlation for A:
so if an element has A == 0.5 and the other equal to A == 1.5 they are fully correlated (red). Otherwise if an element has A == 0.5 and the second item has A == 3.5 they are uncorrelated (blue).
Similarly for B:
Now I want multiply the two correlation matrixes, but I want to obtain as final matrix a matrix with two axis, where the new axes are a folded version of the original axes:
def get_folded_bin(ia, ib):
return ia * nb + ib
here what I am doing:
result = np.swapaxes(np.tensordot(a, b, axes=0), 1, 2).reshape(na* nb, na * nb)
visually:
and in particular this must hold:
for ia1 in xrange(na):
for ia2 in xrange(na):
for ib1 in xrange(nb):
for ib2 in xrange(nb):
assert(a[ia1, ia2] * b[ib1, ib2] == result[get_folded_bin(ia1, ib1), get_folded_bin(ia2, ib2)])
actually my problem is to do it with more quantities (A, B, C, ...) in a general way. Maybe there is also a simpler function within numpy to do that.

np.einsum lets you simplify the tensordot expression a bit:
result = np.einsum('ij,kl->ikjl',a,b).reshape(-1, na * nb)
I don't think there's a way of eliminating the reshape.
It may also be easier to generalize to more arrays, though I wouldn't get carried away with too many iteration variables in one einsum expression.

I think finally I have found a solution:
np.kron(a,b)
and then I can compose with
np.kron(np.kron(a,b), c)

Roll rows of a matrix independently

I have a matrix (2d numpy ndarray, to be precise):
A = np.array([[4, 0, 0],
[1, 2, 3],
[0, 0, 5]])
And I want to roll each row of A independently, according to roll values in another array:
r = np.array([2, 0, -1])
That is, I want to do this:
print np.array([np.roll(row, x) for row,x in zip(A, r)])
[[0 0 4]
[1 2 3]
[0 5 0]]
Is there a way to do this efficiently? Perhaps using fancy indexing tricks?

Sure you can do it using advanced indexing, whether it is the fastest way probably depends on your array size (if your rows are large it may not be):
rows, column_indices = np.ogrid[:A.shape[0], :A.shape[1]]
# Use always a negative shift, so that column_indices are valid.
# (could also use module operation)
r[r < 0] += A.shape[1]
column_indices = column_indices - r[:, np.newaxis]
result = A[rows, column_indices]

numpy.lib.stride_tricks.as_strided stricks (abbrev pun intended) again!
Speaking of fancy indexing tricks, there's the infamous - np.lib.stride_tricks.as_strided. The idea/trick would be to get a sliced portion starting from the first column until the second last one and concatenate at the end. This ensures that we can stride in the forward direction as needed to leverage np.lib.stride_tricks.as_strided and thus avoid the need of actually rolling back. That's the whole idea!
Now, in terms of actual implementation we would use scikit-image's view_as_windows to elegantly use np.lib.stride_tricks.as_strided under the hoods. Thus, the final implementation would be -
from skimage.util.shape import view_as_windows as viewW
def strided_indexing_roll(a, r):
# Concatenate with sliced to cover all rolls
a_ext = np.concatenate((a,a[:,:-1]),axis=1)
# Get sliding windows; use advanced-indexing to select appropriate ones
n = a.shape[1]
return viewW(a_ext,(1,n))[np.arange(len(r)), (n-r)%n,0]
Here's a sample run -
In [327]: A = np.array([[4, 0, 0],
...: [1, 2, 3],
...: [0, 0, 5]])
In [328]: r = np.array([2, 0, -1])
In [329]: strided_indexing_roll(A, r)
Out[329]:
array([[0, 0, 4],
[1, 2, 3],
[0, 5, 0]])
Benchmarking
# #seberg's solution
def advindexing_roll(A, r):
rows, column_indices = np.ogrid[:A.shape[0], :A.shape[1]]
r[r < 0] += A.shape[1]
column_indices = column_indices - r[:,np.newaxis]
return A[rows, column_indices]
Let's do some benchmarking on an array with large number of rows and columns -
In [324]: np.random.seed(0)
...: a = np.random.rand(10000,1000)
...: r = np.random.randint(-1000,1000,(10000))
# #seberg's solution
In [325]: %timeit advindexing_roll(a, r)
10 loops, best of 3: 71.3 ms per loop
# Solution from this post
In [326]: %timeit strided_indexing_roll(a, r)
10 loops, best of 3: 44 ms per loop

In case you want more general solution (dealing with any shape and with any axis), I modified #seberg's solution:
def indep_roll(arr, shifts, axis=1):
"""Apply an independent roll for each dimensions of a single axis.
Parameters
----------
arr : np.ndarray
Array of any shape.
shifts : np.ndarray
How many shifting to use for each dimension. Shape: `(arr.shape[axis],)`.
axis : int
Axis along which elements are shifted.
"""
arr = np.swapaxes(arr,axis,-1)
all_idcs = np.ogrid[[slice(0,n) for n in arr.shape]]
# Convert to a positive shift
shifts[shifts < 0] += arr.shape[-1]
all_idcs[-1] = all_idcs[-1] - shifts[:, np.newaxis]
result = arr[tuple(all_idcs)]
arr = np.swapaxes(result,-1,axis)
return arr

I implement a pure numpy.lib.stride_tricks.as_strided solution as follows
from numpy.lib.stride_tricks import as_strided
def custom_roll(arr, r_tup):
m = np.asarray(r_tup)
arr_roll = arr[:, [*range(arr.shape[1]),*range(arr.shape[1]-1)]].copy() #need `copy`
strd_0, strd_1 = arr_roll.strides
n = arr.shape[1]
result = as_strided(arr_roll, (*arr.shape, n), (strd_0 ,strd_1, strd_1))
return result[np.arange(arr.shape[0]), (n-m)%n]
A = np.array([[4, 0, 0],
[1, 2, 3],
[0, 0, 5]])
r = np.array([2, 0, -1])
out = custom_roll(A, r)
Out[789]:
array([[0, 0, 4],
[1, 2, 3],
[0, 5, 0]])

By using a fast fourrier transform we can apply a transformation in the frequency domain and then use the inverse fast fourrier transform to obtain the row shift.
So this is a pure numpy solution that take only one line:
import numpy as np
from numpy.fft import fft, ifft
# The row shift function using the fast fourrier transform
# rshift(A,r) where A is a 2D array, r the row shift vector
def rshift(A,r):
return np.real(ifft(fft(A,axis=1)*np.exp(2*1j*np.pi/A.shape[1]*r[:,None]*np.r_[0:A.shape[1]][None,:]),axis=1).round())
This will apply a left shift, but we can simply negate the exponential exponant to turn the function into a right shift function:
ifft(fft(...)*np.exp(-2*1j...)
It can be used like that:
# Example:
A = np.array([[1,2,3,4],
[1,2,3,4],
[1,2,3,4]])
r = np.array([1,-1,3])
print(rshift(A,r))

Building on divakar's excellent answer, you can apply this logic to 3D array easily (which was the problematic that brought me here in the first place). Here's an example - basically flatten your data, roll it & reshape it after::
def applyroll_30(cube, threshold=25, offset=500):
flattened_cube = cube.copy().reshape(cube.shape[0]*cube.shape[1], cube.shape[2])
roll_matrix = calc_roll_matrix_flattened(flattened_cube, threshold, offset)
rolled_cube = strided_indexing_roll(flattened_cube, roll_matrix, cube_shape=cube.shape)
rolled_cube = triggered_cube.reshape(cube.shape[0], cube.shape[1], cube.shape[2])
return rolled_cube
def calc_roll_matrix_flattened(cube_flattened, threshold, offset):
""" Calculates the number of position along time axis we need to shift
elements in order to trig the data.
We return a 1D numpy array of shape (X*Y, time) elements
"""
# armax(...) finds the position in the cube (3d) where we are above threshold
roll_matrix = np.argmax(cube_flattened > threshold, axis=1) + offset
# ensure we don't have index out of bound
roll_matrix[roll_matrix>cube_flattened.shape[1]] = cube_flattened.shape[1]
return roll_matrix
def strided_indexing_roll(cube_flattened, roll_matrix_flattened, cube_shape):
# Concatenate with sliced to cover all rolls
# otherwise we shift in the wrong direction for my application
roll_matrix_flattened = -1 * roll_matrix_flattened
a_ext = np.concatenate((cube_flattened, cube_flattened[:, :-1]), axis=1)
# Get sliding windows; use advanced-indexing to select appropriate ones
n = cube_flattened.shape[1]
result = viewW(a_ext,(1,n))[np.arange(len(roll_matrix_flattened)), (n - roll_matrix_flattened) % n, 0]
result = result.reshape(cube_shape)
return result
Divakar's answer doesn't do justice to how much more efficient this is on large cube of data. I've timed it on a 400x400x2000 data formatted as int8. An equivalent for-loop does ~5.5seconds, Seberg's answer ~3.0seconds and strided_indexing.... ~0.5second.