scipy/sparse-inv-then-multiply
| Rule | sparse-inv-then-multiply |
|---|---|
| Pack | SciPy |
| Severity | warning |
| Fix | unsafe, with xpd check --fix --unsafe-fixes |
| Library versions | every version |
The inverse of most sparse matrices is dense, so inv(A) @ b builds a huge array. Call spsolve(A, b), which was 260 times faster.
x = spla.inv(A) @ bx = spla.spsolve(A, b)What the rule reports
Section titled “What the rule reports”The rule reports scipy.sparse.linalg.inv(A) @ b and inv(A).dot(b), because the inverse of
most sparse matrices is dense. On a tridiagonal matrix of 1,000 rows, whose inverse holds 1,000,000
values, spsolve ran 260 to 270 times faster.
How to fix it
Section titled “How to fix it”xpd check --fix --unsafe-fixes applies the fixes of this rule. Check the result, because an unsafe fix can change what the code does.
The unsafe fix writes spla.spsolve(A, b) when inv is reached through its module. The last bits change, and a b of shape (n, 1) gives a 1-D result. inv imported by name is fixed only in a file that imports spsolve from scipy.sparse.linalg.
Message
Section titled “Message”xpd prints this message with the issue. A name that starts with $ stands for code from your file.
$L.inv($A) builds the whole inverse, which is dense for most sparse matrices. Write $L.spsolve($A, $B), which ran 260 to 270 times faster with an inverse of 1,000,000 values.Browse the other rules of the SciPy pack.